A program manager at a global bank reaches for her mouse and minimizes a window. On one tab is the company’s AI platform. Twelve million dollars to build. On the other is her own setup: ChatGPT, Claude Code, NotebookLM. Twenty dollars a month. She does her work in those tabs. So does most of her team.

They sat through another AI day and completed a 30-day sprint. When it ended, nothing replaced it. So they moved ahead without guidance or permission.

I see it in every company I walk into. The people moving fastest with AI have no mandate. The people with the mandate aren’t moving fast enough.

Executives have something to prove in 2026. That AI investments were sound. That returns are coming. An estimated $300 billion was spent on enterprise AI last year. Only 12% of CEOs reported meaningful returns. A directive is quietly working through Fortune 500 boardrooms: be more efficient, cut operating budgets. Go figure out how to do it with AI.

AI systems are table stakes. No company I’ve seen has built a working system to prepare people to use AI.

AI makes the divide between approved systems and workarounds visible faster than any ERP or CRM rollout ever did. It touches every role, every task. And yet, leaders respond with the same mandates and change-management playbooks they’ve used for decades.

Previous technologies changed how people organized their work. AI is the first to do the work. That’s why people naturally resist it.

Energy surges around an AI launch. Training days, prototype sprints follow. For a while, the momentum is real. Then the sprints end, and nothing sustains them. Nobody designed change to last.

Those moving with AI drift back to the twenty-dollar tab. The official AI system, built for everyone, serves no one. Adoption becomes compliance. But compliance doesn’t lead to progress.

So the organization does what organizations do with problems. A strong performer gets tapped to “fix AI” on top of the day job. The people already in the twenty-dollar tabs don’t wait.

A “business famous” leader gets recruited to modernize a brand. Within months, the energy she brings to the business stands out.

Her boss hands her the AI mandate. She’s not an AI strategist. She doesn’t have working instincts or the bandwidth to become one. What she has is a title, hundreds of people waiting for direction, and no system to rely on.

The problem she sees won’t let her sit still. Her reputation is what she does, not what she protects. She doesn’t see AI as a threat to her position. She sees it changing her career trajectory. That’s why she moves.

Not everyone sees it that way. Two leaders, same company, different functions. In a strategy review, one heard “AI will automate 30% of your function” and went quiet. The other started exploring before the presentation ended.

The first spent fifteen years becoming the person who knew the answer. The second became the person who found it.

Most people reading this spent decades becoming the person with the answer. AI makes answers available to everyone. The scarcity remains judgment: strategic, creative, imaginative. People who use AI to think more clearly and act more decisively have pulled ahead. They aren’t waiting for the organization to catch up.

Their livelihood depends on moving.

Whether you’re tapped from inside or recruited from outside, the question is the same: how do you find the movers?

The common path: write a plan, launch programs, move on. Let a thousand flowers bloom. Most are weeds. There’s corporate-speak for what follows: launch and leave. The plan gets filed. Leadership moves to the next thing.

The uncommon path follows the work and the people driving it. Intentionally. Find those already performing and build around them.

Most take the first path.

The uncommon path needs a system. A Human OS for AI. A set of practices where teams build out AI through the work itself.

We built one in my previous role without knowing it at the time. Then I left to build a lab dedicated to it.

Both cases led to the same discovery: you can’t roadmap something moving as fast as AI. That turns out to be the unlock.

We didn’t plan what teams would build. We created the conditions for them to build what the work demanded.

They heard the rationale on day one. Then they went to work in a sandbox with clear guidelines and models for text, image, and summarization.

They explored small hacks for work they’d been doing for years. Then they took leaps. Months later: dozens of proven, task-informed agents working with teams in a third space between the office and Zoom calls. We didn’t plan that arc. The team’s work produced it.

People on the front line were the ones who moved. They worked together in private Slack groups, bypassing reporting lines, silos, or geography. It gave them something no roadmap can: confidence.

We study how people feel about AI entering the workplace. One dividing line holds: fear on one side, optimism on the other. The cause is consistent. People with experience build confidence. People without it run scared.

One colleague told me something I haven’t forgotten. She was a junior program manager, not typically first in line for talent investment.

She said: “You didn’t announce anything. You didn’t tell us to use Copilot because AI is the future. You invested in us.”

We didn’t reduce or automate. We grew together. That happens through work. The below-the-radar kind, where half of what you build breaks and the other half surprises you.

Without anyone naming it, an operating system emerged.

The argument against investment in people-centric systems is easy to make.

AI gets better by the day. Perhaps models get so powerful that human readiness becomes irrelevant. Maybe the people problem solves itself. Nothing in the field supports it. AI gets easier to use. The judgment required to use it well doesn’t.

People stop questioning AI output within weeks. When AI drifts, and it will, who sees the cliff? Who knows how the company works and catches what AI misses? Your teams are your organizational immune system. Gut them, and they don’t grow back.

The decision point for team investment isn’t next year. Maybe you already passed it, or maybe it’s next week. Most won’t recognize it as one. It will look like a budget review. Or a meeting about whether an AI program worked. When that moment arrives, AI either earns its keep or gets filed next to programs that launched with town halls and died in spreadsheets.

The movers in your business are working with AI. With tools you didn’t choose, in workflows you can’t see. Some are burning out. Others are plotting their next move. The window to build around them is open. It’s also closing fast.

The bank’s program manager is still working in two tabs. She’s been ahead the whole time. Find the people like her. They’re already moving. Don’t hand them a mandate. Build the system around them.

The operating system starts there. So do the returns.

Leave a comment

In January 2020, I had a line of sight into COVID well before others.

We had offices in Beijing, Shanghai, and Shenzhen. The coronavirus imposed a lockdown on 70 million people, and with it, one of our largest global markets. Here in America, it was business as usual. I went to Costco and loaded a cart with rice, potatoes, flats of water, and enough toilet paper to fill a trunk. By March, people were fighting over what we already had on our shelf.

I wrote Perspective Agents about this pattern. The signals are always there. The capacity to interpret them rarely is. By the time the public catches up, the shelves are empty.

Matt Schumer recently made waves with an essay titled “Something Big is Happening” by comparing AI to COVID. It was covered by Fortune, the Washington Post, CNBC, and dozens of other media outlets. It’s been shared more than 50 million times. The share is the tell. Fifty million people need someone else to say what they know is real. The COVID parallel is structural more than metaphorical. The same pattern is repeating.

Anthropic’s CEO, Dario Amodei, described what’s on our doorstep in terms that need to be heard beyond research labs. Last Friday, he sat across from podcaster Dwarkesh Patel and described what he called the “end of the exponential.” Amodei said that the scaling of AI capabilities followed a predictable path and is now approaching its conclusion. Then he used a phrase about the endpoint that stuck with me since hearing it: we’ll have a country of geniuses sitting in a data center.

Amodei meant it concretely: systems capable of doing the work of highly educated professionals, start to finish, around the clock. He put the timeline at one to three years from now.

Beyond the statements, what stood out to me was his composure. Amodei delivered them as an engineer describes a build schedule. He found it, in his words, “absolutely wild” that the public remains focused on the usual political cycles, while a transition of this magnitude bears down on us all.

When it arrives, he said, everyone will know. Amodei read like a demolition schedule for buildings we’re still furnishing.

The next morning, Amodei’s description of what’s here had become another item in an endless scroll of AI takes. You might have seen Patel’s headline somewhere between waking up and pouring your coffee. You probably kept scrolling.

But sit with the phrase. A country of geniuses in a data center. Your mind goes to the servers and the cooling systems, abstractions far from your daily life.

The genius doesn’t stay contained in a data center. This intelligence already extends into your work, your kids’ education, and your understanding of what you’re professionally worth. Every prior technology changed you. None of them told you. This one does.

We tend to absorb warnings like this without acting on them. See if this sounds familiar.

A leader sees the Amodei clip. Reads the Schumer essay. Comes in on Monday and opens the planning deck he opened last Monday. Same categories. Same twelve-month assumptions. The information arrived. The deck didn’t change.

He forwards Schumer’s post to his boss with a note. “You should look at this.” The boss has other things on his mind. He takes the lack of response as permission to wait. What a country of geniuses means for his team can only be answered by him

That afternoon, he opens up Co-Pilot. Creates a presentation in a day or two. Tells a colleague AI is the real deal. He describes a system that will upend his working life, and he uses it to fix up slides.

What this reveals isn’t an AI problem. It’s a human one.

There’s no shortage of sources consumed with AI. I haven’t seen enough of what it means on an honest, human level, so I built something new.

humanreadiness.org.

A decade inside this world gives me a useful vantage point. The site is where I share what it reveals: not answers, but provocations from AI centers, new research, and lived experience from my work at Andus Labs. Why throw something else into the AI hype cycle?

The last time I had this feeling, I was standing in a Costco parking lot loading bags into the trunk, hoping I was wrong. Two months later, the story caught up. I’m in the parking lot again. Same feeling in the chest. What goes in the cart this time isn’t so obvious.

There’s no aisle yet for what we need.

Leave a comment

Subscribe now

Last week at Davos, CEOs centered on a fundamental question: why isn’t AI working for us?

From where we sit at Andus Labs, the answer has little to do with AI itself. The patterns emerging from our work point to something else.

In 2026, companies are splitting into two camps: those realizing remarkable impact from AI, and those struggling to see any value at all. The data from Davos tells the story. Those seeing real impact remain rare.

When it works, it’s unmistakable. Strategists and research teams pressure-test dozens, even hundreds, of scenarios in a single session. Non-technical product leaders build functional prototypes before any engineer sees them. Talent leaders stand up working labs in the time it takes a consultant to write a deck.

In this mode, the unimaginable becomes routine.

We feel this firsthand. Our team discovers previously impossible applications every day. The only way to describe it is a fever-dream state: the overwhelming sense of possibility compounding on possibility. And it keeps accelerating.

What I keep seeing in the field: the gap between what AI can do and what most organizations do with it is widening rapidly, not narrowing.

AI investment isn’t the constraint. The lack of human investment and ingenuity to keep pace with it is the bottleneck.

That gap is a crisis for every organization that bought the promise: buying ruinously expensive shelfware. In aggregate, billions invested in AI capabilities go unused because people aren’t invested in working with them.

Capabilities and Systems

This distinction matters.

AI capability is what the technology can do. It’s mindblowing. And there’s no ceiling in sight.

Human-centered investment will determine whether you realize the value of AI’s expansive power. Investments in spaces, workflows, permissions, habits, and coaching to move from “I saw another demo” to “this is how we work now.”

Organizations are sitting on vastly underutilized AI potential. Few have team-centered systems. They wonder why there’s a vast expectation gap.

Consider a tale of two companies.

Company A focused on enterprise licenses. Created three-year roadmaps. Ran a bunch of training. Built a center of excellence. Eighteen months later: single-digit adoption. The pilots that succeeded didn’t scale. The people who were supposed to change... didn’t. The executive team is now asked about ROI. No one has an answer.

Company A is sitting on shelfware. They’re falling behind. The gap widens every quarter.

Company B started smaller and more specific. Instead of focusing on long-term strategy decks and generic AI training, they invested in changing how work gets done with people doing the work. They didn’t announce transformation. They demonstrated it, one step at a time. Collective learning on how AI helps and fails accumulated. The team emphasized impact over adoption. Impact came from the bottom up versus the top down.

Twelve months later: AI is embedded in workflows—better, faster, more impactful workflows—not an “AI initiative.” Teams are excited and engaged. Cases accumulate.

What lies in between

The difference between the two scenarios isn’t magic.

It’s organization planning before tech planning. Executive sponsorship that doesn’t waver. Budgets tied to outcomes. Relentless testing and learning. Communication that never stops.

Company B begins building a Human OS. It aggregates the working practices and systems that accelerate how people best use and grow with AI. Not training. Not governance. Not tools. The underlying architecture for how work gets done better.

If these patterns sound familiar, ask yourself some questions:

• What percentage of employees use AI daily?

• What are teams building and learning?

• What remarkable insights and cases have you captured?

• Which workflows have changed?

• Where’s the measurable ROI?

• What are teams growing into—and to what end?

If the answers are unclear, you’ll know you don’t have a technology problem. You have an investment-in-your-teams problem.

And sitting on the shelf, AI isn’t just idle. It’s an anchor and competitive burden dragging you further behind faster-moving competitors.

Patterns are unfolding that address the Davos question. Leaders have a choice.

Invest in your people if you expect any value from investments in machines.

Leave a comment

I recast this newsletter from “Perspective Agents” to “The Human OS of AI.” I’ve learned the real problem isn’t how we think about AI — it’s how we put it to use to improve ourselves and our businesses. The technology works. How to actually work with it is still being figured out.

Chris Perry is the founder and CEO of Andus Labs. The company builds human-centered systems and programs for organizations navigating AI transformation.

You can reach him at cperry@anduslabs.com

For more on the Human OS and how we put it to work: https://humanreadiness.com.

Subscribe now

Here’s the word from the World Economic Forum this week: AI is now the most consequential issue on the global agenda. Above geopolitics. Above climate. Above the media hurricane around Trump.

When the world’s most powerful leaders, investors, and executives gather in the Swiss Alps, they’ve decided this matters most. And two of the primary architects of AI said the quiet part out loud. You should pay close attention to them:

Yesterday, Anthropic CEO Dario Amodei projected “Nobel-level” AI intelligence and the total replacement of software engineers within 6 to 12 months. Google DeepMind CEO Demis Hassabis put 50% odds on human-level AI by 2030.

Both warned of a “lag” in labor statistics that masks a work crisis already in motion.

To stress: these are the main actors building AI systems — they see the power that the public has no view into.

AN OPERATING BUBBLE SEEN FROM THE TOP

Even then, here’s the problem: Davos conversations are atmospheric and directional, not operational (grounded in the daily realities we live in).

I’ve done the Davos trips and see what happens when executives return. Lots of fire, plenty of directives. Then, beyond fire and brimstone, few have clear ideas on what to actually do. In this case, they’ll need to figure it out—fast.

The ROI/working value to date tells a brutal story.

PwC's 29th Global CEO Survey provided a reality check for executives. 56% of CEOs reported seeing no financial benefit from AI to date. Not underwhelming returns. No returns. That’s quite remarkable, given that AI had a similar placement on the WEF agenda last year.

The most powerful technology in human history, and we’re capturing a fraction of its value, if any.

A VIEW FROM THE GROUND

This isn’t a technology problem. It’s a human problem. Or, put by Mohamed Kande, PwC’s chairman, “most leaders have forgotten the basics” on AI strategy.

I have eight years of experience running AI labs focused on human potential and risk. Our team translates the read into the basics (meaning actions to take).

One conclusion:

No system exists to help people understand, think, and build with AI to create real value.

The models, platforms, and compute are there.

Here’s what’s not: the personal and organizational approach to make it work.

The training programs aren’t enough. The pilots don’t scale. The tools don’t fit how people get work done in the trenches. The champions burn out. The gains evaporate.

Everyone is investing in AI capability. Few are investing in human readiness.

At Andus Labs, we call this systemic gap the Human OS for AI.

It’s not software. It’s the organizational operating system — people, workflows, rituals, incentives, governance — that sits between AI capability and actual value.

Without it, the math doesn’t work. Capability times zero equals zero. With it, organizations unlock the multiplier. The same investment yield can yield a 2x, 4x, or 8x return.

People are the unlock. And few—if any—are building human capacity and org design to address it.

At Andus Labs, we’re mobilizing some of the best organizational architects, change leaders, and AI engineers to crack the code on unrealized potential.

ACTING ON THE SITUATION

I just wrote a manifesto on the issue and this missing link. It names the problem, defines the architecture, and provides ways to act. It also predicts what happens next.

If you want to understand what’s actually at stake — not the Davos version, the operational and personal version — read it.

If you want to engage with AI in a way that creates value instead of theater — read it.

If you want to protect your career and take it to unimaginable heights— read it.

Jamie Dimon warned this week that AI may move “too fast for society.” Anthropic’s Amodei maintained his prediction that up to half of entry-level white-collar jobs will disappear in the next 1-5 years. The warnings could fill another newsletter.

There’s no more time for reflection. Act now.

The part that should concern everyone building a career: if AI handles entry-level work, how do juniors develop the skills to become seniors? The apprenticeship model is broken. The ladder is being pulled up.

What most haven’t recognized yet: the ladder doesn’t stop at entry-level. Middle management is next. Eventually, the executive suite.

I’m convinced AI-native workers will rise — new ladders will form. The AI-ambivalent are screwed.

The Human OS isn’t just an organizational problem. It’s a career problem. Yours.

Read the paper to broaden and deepen your understanding of what’s happening and what’s missing. The thesis is summarized below.

Go to The Human OS for AI —> https://humanreadiness.org/

The conversation at Davos is important. But it won’t tell you what to do tomorrow morning.

This just might.

cperry@anduslabs.com

Leave a comment

Every post has a comment button. Most don’t deserve it. This one does. Read the manifesto. Think about what’s at stake. Then, in the comments, tell other readers what you see.

Subscribe now

Every age carries a new sound.

Ours is one you can’t so much hear as feel—the low frequency of a world thinking in probabilities. It moves beneath everything: markets, technologies, and the chatter that fills our newsfeeds.

We bet on sports, elections, storms, news, and love, so much so that prediction has infiltrated daily life. Online betting generates billions of dollars annually, blurring the line between sports coverage and sportsbooks. Prediction markets like Polymarket have grown fivefold since 2024; rival Kalshi just raised at a $5B valuation. Both are now embedded in Google Finance searches. Meme-stock trading—trades that ignore fundamentals—mirrors the explosive conditions seen in 2021.

Beneath the waterline, there’s much more to see. The cause of widespread speculation is woven deeply into our collective psychology. When life lacks certainty, we don’t retreat—we improvise. We build what sociologist Aris Komporozos-Athanasiou calls speculative communities: groups bound by volatility. Here, we find connection in motion and solidarity in suspense.

Komporozos-Athanasiou writes that to speculate increasingly means ‘to connect,’ a public act of endorsing unknowing as a way of belonging. Speculating is how we relate to the future—by anticipating and imagining it—and how we make it actionable in the present.

When the only thing we can count on is uncertainty, shorting and hedging the unknowable becomes the spirit of the age.

From Innovation to Prediction

That logic, the culture of wagering on what comes next, has drifted from the markets into the bloodstream of daily life.

For a century, management guru Peter Drucker defined innovation as measured bets on new ideas, products, and markets. That logic thrived when information was scarce and the future seemed distant. But scarcity is gone, and the future now overwhelms the present. The frontier no longer represents innovation. It’s now about foresight.

Alex Danco, editor-at-large at Andreessen Horowitz, described the shift from ‘capital-at-risk’ to ‘prediction contracts,’ where we increasingly take stakes in information flows. He suggests a new form of value creation through predictive participation. To be early is to hold new capital; to arrive on time is to hold new social status.

The GameStop short squeeze of 2021 wasn’t an anomaly; it was a preview. Retail traders in the back alleys of Reddit and Discord turned volatility into collective theater. Unlike Wall Street traders, the WallStreetBettors weren’t analyzing; they were performing. They bet on a shared mission.

Speculation as Culture

Look around. Speculation is on our screens, shapes our careers, and influences our conversations. The media itself has become a prediction market, elevating those who spot the trend, call the fall, and ride the wave first. Every story is a futures contract in attention.

Cultural transformations like this accompany technological revolutions. Marshall McLuhan outlined how this works in his Laws of Media. He noted every new medium enhances, obsoletes, retrieves, and reverses something in culture. That’s why his predictive meme from 1964, “the medium is the message,” is so enduring.

Prediction enhances collective pattern recognition, letting millions sense and act on the same signal in real time. It makes obsolete the idea of expertise and the notion that someone, somewhere, actually knows the answer. It retrieves the old rituals of divination—the communal act of trying to know what comes next. And pushed far enough, it reverses into addiction and nihilism: a culture that can’t stop refreshing the future even as it forgets how to live in the present.

McLuhan warned that every technology carries the seeds of its own undoing. The tools that make us predictive can just as easily make us predictable and lame.

Industrial-Grade Predictability

We now access a machine of staggering power. Artificial intelligence is the first technology to make predictions its native language.

It’s the purest form of speculation ever built. AI doesn’t just reflect the prediction economy but creates it. Every model, every algorithm, every dataset is a structured wager on what comes next.

AI turns foresight into infrastructure. Prediction becomes an industrial act. The output isn’t an image, product, or truth to buy. It’s the probability of an outcome.

It fulfills McLuhan’s law in every direction at once. It enhances prediction by processing more possibilities than any mind could. It makes obsolete the slow rituals of research and reasoning. It retrieves the dream of prophecy—the desire to see before others do. And pushed too far, it reverses into hallucination: a machine that doesn’t just predict the world but starts to replace it.

If speculation is our new vernacular, AI is its accent, seemingly fluent, confident, everywhere. It transforms how we write, decide, and design, not by teaching us something new but by teaching us to anticipate its anticipation.

Every prompt is a small rehearsal in prediction. We ask what we might say next; it answers with what we’re most likely to believe. The human in the loop becomes the loop itself.

The irony is that the more accurate these models become, the less we understand what they’re predicting on our behalf. The line between foresight and fabrication dissolves. In the prediction economy, AI isn’t just another tool. It’s the house we all gamble in.

Organizational Discord

This shift leaves institutions in uncharted territory. Over centuries, they were constructed for a rational world, one where people acted on reason and information rather than on predictions and impulses. That world is fading.

Inside companies, employees now trade timing for tenure. They watch for signals, align with narratives, and position themselves like investors in their own careers. Buyers, too, are less interested in what products do than in what they mean.

Markets are a negotiation between potential and valuation—a wager on how long belief can hold its price. Nowhere is this more visible than with frontier-model companies.

OpenAI, Anthropic, xAI, and their peers trade less on what they’ve built than on what they will unlock, valued not by revenue but by the promise of transformative potential. Their value trades on a faint glow of the future that seems just beyond our reach.

When prediction becomes part of culture, the task is no longer to outguess it, but to live within it without being consumed.

Resilience requires staying human in the loop. We must resist the urge to treat every moment as a trade. It means remembering that people don’t follow instructions; they follow signals of belief, of trust, of meaning.

Prediction can’t replace perspective. When everything becomes a bet, meaning will trade at a premium. For those steering organizations, the challenge becomes how to stay human within the machinery of anticipation, knowing that not every future must be priced before it’s lived.

And so the real work is balance: bringing proportion into a world that mistakes probability for destiny. It’s the work of learning when to play an algorithmic game and when to step outside it. Of measuring progress not by how early we glimpse the future, but by how deeply we can still feel the present.

The danger ahead isn’t being wrong. It’s losing our capacity to live with the unknown.

Leave a comment

Subscribe now

A new species of corporate failure has emerged. The culprit? The misapplication of AI. It’s the consequence of mistaking efficiency for progress—and failing to see the quiet resistance forming in culture’s shadows.

Coca-Cola’s 2025 holiday campaign is the latest exhibit in a growing gallery of brands finding themselves on the wrong side of culture. They’ve discovered that the future they’re selling so confidently isn’t necessarily one their customers want to buy. What follows is a cautionary tale that extends well beyond advertising.

The Coke spot, released this week and titled “Holidays Are Coming,” is entirely AI-generated, a “refreshed and optimized” remake of the brand’s classic convoy of glowing red trucks. This time, the cast of polar bears, pandas, penguins, puppies, and rabbits glows not with holiday warmth, but with what’s been called a “generative AI shine.”

The backlash has been instant—an emerging case study in AI resistance across Google, Reddit, and X.

Jason Zada, founder and chief creative officer of AI studio Secret Level, brushed off the criticism as mere fear of change.

“The haters on the Internet are the loudest. A lot of the people complaining last year were from the creative industry who were just afraid — afraid for their jobs, afraid for what it did. But I think the spot tested really well, and average people really enjoyed it.”

The fear of AI is real—but what’s more pressing is the erosion of judgment and perspective.

For all the technical sophistication—for all the outputs that look real enough to fool the eye—it’s instinct, taste, and care that still matter. In other words, the sense that someone demanded the work be right. As AI seeps deeper into the creative process, human ethos becomes not obsolete, but priceless in wielding it.

An Illusion of Progress

The Coca-Cola case is no outlier. The pattern repeats company by company, campaign by campaign. No brand, no matter how iconic, seems immune.

Signs of misuse and backlash emerged in 2023, when Levi’s unveiled AI-generated fashion models to showcase diverse body types and skin tones. What sounded progressive in theory quickly rang hollow. Critics called it “lazy, problematic, and racist”—a reminder that real diversity begins with hiring actual people.

By 2024, unease went mainstream. Apple’s “Crush” ad showed a hydraulic press flattening music instruments and art supplies before revealing a gleaming iPad beneath. The metaphor was unintentional but painfully clear: technology crushing art. Within days, Apple pulled the ad and apologized, proof that even the most design-literate company can miss the public’s emotional register.

Around the same time, Google’s Olympic-themed Gemini spot featured AI writing a child’s letter to Olympian Sydney McLaughlin-Levrone. It carried an unspoken suggestion: that even childhood wonder could be automated. Alexandra Petri captured public sentiment in the Washington Post: “This ad makes me want to throw a sledgehammer into the television every time I see it.” The ad was pulled from the rotation.

This year, we’ve become fluent in the new visual language of AI—and quick to call it out when we see it. OSINT-inspired sleuths at Blackbird Spyplane spotted digital fingerprints in J.Crew’s retro campaign, accusing the brand of using “A.I. to counterfeit their own vibes.” Backlash seeped into social media and culture sources like The Cut.

That same crowd-sourced scrutiny turned toward pop culture’s most trusted figure. Taylor Swift—long synonymous with authenticity—was called out by her own fans for the uncanny sheen of promoting The Life of a Showgirl. As a Wired feature captured it:

“The first sign that it was AI was that it didn’t look great,” said Marcela Lobo, a designer and longtime Swift fan. “The shadows didn’t match, the windows and the painted piano—it looked like shit, basically.”

The call from her fans was curt and straight: Do better.

More Care. Less Slop.

Somewhere in the rush to innovate, perspective—and care—fell out of view.

“AI slop” is now part of the cultural vernacular, a shorthand for dissonance: work that looks fine on the surface but feels hollow underneath.

AI campaigns confuse what people expect from brands—not efficiency or novelty, but genuine connection. The genie may be out of the bottle, as they say, but that doesn’t settle the issue at hand. Every action invites a reaction. The cultural recoil we’re seeing now is proof.

AI in capable hands is powerful—but human sensibilities endure. Beyond Fortune 500 firms, a small start-up lit up the paradox beneath it all. This summer, billboards from Artisan blared “Stop Hiring Humans.” The company framed it as provocation, boasting of its earned media reach—but in doing so, it exposed how casually we’ve begun to market the end of ourselves. They said:

“We didn’t expect people to get so mad. The goal of the campaign was always to rage bait, but we never expected the level of backlash we ended up seeing.”

Don’t Blame the Luddites

Those working on the front lines of AI should be careful not to confuse healthy skepticism with Luddism. What some call resistance may be conscience. The efficiencies we celebrate come with costs that no spreadsheet can measure, losses no digital twin can mirror.

Some brands, more attuned to culture than code, have sensed the current turning. Polaroid and Heineken pushed back against the algorithmic tide, running human-centered campaigns to reach people weary of AI. There’s an irony that speaks to the age: protecting our humanity now requires advertising.

Enduring brands are hard to build and impossible to fake. What sustains them isn’t optimization but care—the patient work of noticing, refining, and feeling what belongs in the world. AI can help us do all of this, but only if we’re willing to get our hands dirty in the making—to use it not as a substitute for imagination and refinement, but as a means of sharpening it.

Because if something is worth creating, it’s worth having people own the creation—not just in the legal sense, but in the deeper sense of authorship, responsibility, and care.

In the end, perspective may prove the most irreplaceable human skill of all—the one thing no model can simulate, no brand can fake, and every leader must nurture.

Leave a comment

Subscribe now

On March 11, 2020, Tom Hanks announced he tested positive for the coronavirus. That night, an NBA game stopped when players were pulled off the court. Only then—after weeks of lockdown reports from China and Italy, after the World Health Organization’s urgent bulletins of ‘disruptions to everyday life’—did much of America register what was happening. Reality arrived not through science, but through celebrity.

By then, it was too late to prepare. Despite unmistakable signs, an NBC poll found more than half of Americans still believed their lives would change little—if at all.

We know what happened in hindsight. Homes replaced headquarters. Corporations shifted to Zoom. Playgrounds to Fortnite. Shifts anticipated in the future became reality in days.

We’re in a similar moment with artificial intelligence. Signs of disruption are everywhere, scattered across industries like breadcrumbs—law firms quietly feeding document review to LLMs, consultants shedding employees they cannot retrain, junior analysts delivering advice that previously required decades to master.

The transformation is happening now. Yet, as with the early days of the pandemic, recognition is not the same as readiness.

The New Lens of Work

AIs alter more than tasks; they alter perception. These are perspective agents. They’re filters that shape what we notice, what counts as valuable, and what falls away. A lawyer who uses AI to review contracts doesn’t just speed up the job—she begins to redefine which parts of her expertise are essential. The junior consultant, capable of generating senior-level analysis, displaces hierarchies of expertise.

Research reinforces the need to broaden our perspective. Worklytics reports that 69% of knowledge workers use AI tools for research, content creation, and analysis. Gallup finds that adoption at work has nearly doubled to 40% of all employees in just two years.

Yet, according to KPMG, 57% of workers reported hiding their AI use from their bosses—‘AI shame’—while only 7.5% say they have received adequate AI training. It’s not surprising that MIT found that 95% of enterprise AI pilots fail to produce a positive impact. The pattern is familiar: we see it coming, but we don’t fully prepare for it.

Consider OpenAI’s recent professional benchmark report. Legal, financial, and retail experts with an average of fourteen years of experience designed tasks that required four to seven hours of human work. When the same tasks were attempted by both the experts and AI, humans prevailed, but only by a slim margin.

AI’s shortcomings weren’t in reasoning but in surface issues: formatting, presentation, and following instructions to the letter. These flaws will gradually be eliminated with each new model. More telling today: hybrid workflows combining AI and human review delivered results 40% to 60% more efficiently than people alone.

Displacement of Expertise

Until recently, AI’s promise was hobbled by its brittleness—one small error could topple an entire chain of tasks. The new generation of agents is different. They can plan, call on other agents, collaborate, and self-correct.

In his analysis of OpenAI’s report, Ethan Mollick, Co-Director of the Generative AI Labs at Wharton, described watching Claude 4.5 ingest an economics paper, parse it, rewrite the accompanying code, and rerun the analysis. It reproduced findings from the paper in minutes. For research fields long plagued by replication challenges, the implications seem remarkable.

Mollick noted the same models, given a single memo, produced seventeen PowerPoint decks—slick, well-formatted, and empty. The same tools used for clarity can overwhelm us with noise.

In consulting, experience was the ladder to authority. Years of meticulous analysis lead to sound judgment. Today, that staircase is being dismantled, rung by rung.

Accenture’s CEO, Julie Sweet, put it bluntly in her latest earnings call: “We are exiting people on a compressed timeline where reskilling, based on our experience, is not a viable path for the skills we need.” She described a change so significant it reverses “five decades of how we’re working.”

The chainsaws are out. Expertise is being dismantled and dispersed, carrying consequences that reach beyond economics into our psychology. What does it mean when machines can replicate decades of expertise in minutes?

Three Principles for Readiness

Perhaps a better question returns to judgment, something that AI cannot replace. Consider what happens without it. The Harvard Business Review coined a term for the influx of AI-generated reports, memos, and slide decks: workslop. It feels polished, formatted, grammatically sound—and nearly useless. It’s organizational filler, work that fills the inbox without adding value.

The temptation is understandable. When employees can generate dozens of proposals in the time it once took to draft a single one, output swells. But output isn’t insight. The danger isn’t job loss—it’s judgment loss. Volume replacing reflection.

What if AI, deployed without discernment, accelerates us toward workplaces that prize velocity over value? Where the pressure to match machine speed eliminates time for reflection? How do we stop torrents of AI-generated content from washing away valuable expertise?

The path forward requires deliberate choices. First, organizations must invest in human-AI collaboration rather than replacement. We must train people not just to use AI tools, but to know when to refrain, when to slow down, and when to insist on reflection over velocity.

Second, we must reinforce markers for valuable work. If we reward just efficiency and speed, we will inevitably get workslop. We must distinguish between AI-generated output that enhances intelligence and what’s just surface insight. It calls for simple but essential questions: Does this work create value or check boxes? Does it protect human judgment or automate it away? Does it move us closer to meaningful progress, or just faster toward irrelevance?

Third, individuals must become perspective agents themselves—conscious architects of how AI shapes their work. It requires regularly stepping back to ask: What problems am I choosing to prioritize? What steps am I no longer considering? What kinds of thinking am I offloading, and what am I keeping?

The goal isn’t to resist AI but to direct it intentionally, ensuring that it serves us rather than replaces us.

The Question of Choice

Marshall McLuhan once observed that “there is absolutely no inevitability as long as there is a willingness to contemplate the situation.” With COVID, we saw the signals but mistook recognition for preparation. We failed to build new systems, prepare people, or ask hard questions about what mattered most. The result was chaos.

Will we treat this moment as a warning or as another Tom Hanks moment—an acknowledgment too late to matter?

Unlike a virus, AI has no biological inevitability. It can be directed and framed with intention. The choice lies not in the algorithms but in our hands: like the work we value and the necessary judgment we must preserve.

COVID-19 taught us that the biggest changes occur in places we cannot see—behind the scenes, in our assumptions, in policy meetings, in systems that organize human effort.

For leaders, the question isn’t whether we can see the changes coming—it’s whether we’ll act on what we find. How we work, learn, and decide is being rewritten. Our only choice is whether we shape that change or let it shape us in the end.

Share

I’ve explored these ideas further in Perspective Agents, a book that presents frameworks and cases tracing how AI unsettles and redefines the very idea of work. Thanks for reading.

Subscribe now

Soon after ChatGPT launched in 2022, I asked it to summarize a policy paper. The model quietly prioritized some arguments, downplayed others, and subtly conveyed a tone of clinical neutrality. I walked away with a filtered version of an AI’s perspective, unaware of what had been excluded from the document.

This experience comes to mind whenever AI's effects get reduced to simple questions of laziness or intelligence. That framing misses something fundamental: these aren't just smart tools doing our work for us.

In my book, I defined AIs as Perspective Agents—algorithmic systems that control what information reaches us, shape how we understand it, and ultimately construct our reality.

We refer to GPTs as ‘artificial intelligence’ or ‘large language models’, but those terms don't fully capture what the systems actually do.

They don't just retrieve information; they frame it. They influence what we notice, what we overlook, and what we perceive as reasonable or accurate. They're invisible editors of thought rather than neutral calculators of data.

AIs are quietly reshaping human understanding across domains—from medical diagnoses to legal precedents to investment strategies.

New research shows these changes aren't just practical adjustments—they're neurological. Understanding how AI changes our brains will determine whether it becomes a cognitive partner or a cognitive replacement.

The Rewiring Is Real

Nataliya Kos’myna, a research scientist at MIT, recently published a widely reported experiment. Her team asked people to write essays while monitoring their brain activity with EEG sensors. Some participants worked alone, others used Google, and a third group relied on ChatGPT.

The results revealed how ChatGPT measurably reshapes thinking. When participants used AI assistants, their brains showed significantly less activity in regions associated with deep reasoning, memory formation, and sustained attention.

The AI convinced them to take a break from the human job of synthesis and judgment. Those using ChatGPT couldn’t remember or accurately quote from work they had supposedly just created. They felt less ownership of their ideas.

The agent hadn't just helped them think; it had substituted its frame of reference for theirs.

When asked to complete a similar task without AI afterward, their cognitive patterns resembled those of beginners rather than those who had just engaged in complex reasoning. The AI had framed their perspectives and diminished their ability to form them independently.

This phenomenon isn't just happening in research labs. It's unfolding in doctors' offices, where physicians rely on AI to suggest diagnoses; in law firms, where attorneys draft arguments with the assistance of language models; and in boardrooms, where executives rely on AI-generated reports to make multimillion-dollar decisions.

At first glance, this seems like harmless efficiency, a form of cognitive offloading not unlike using calculators to free us from basic math.

Yes, AI can streamline routine tasks and reduce friction where expertise is already strong. But the danger is more subtle.

The study suggests participants offloaded effort and judgment.

Without us knowing, AIs can insidiously take over the mental processes that define human expertise, eroding the ability to develop and sustain those skills independently.

It's not surprising that Kos’myna’s findings sparked extensive discussion and media coverage that amplified public fears. Here's a sampling of the headlines:

From The Hill: "ChatGPT Linked to Cognitive Decline." From Fast Company: "Reliance on ChatGPT Might Be Really Bad for Your Brain." From Axios: "AI's Great Brain-Rot Experiment." And this zinger from David Brooks at The New York Times: "Are We Really Willing to Become Dumber?"

If you follow the media filter, the conclusion is stark: ChatGPT is making us cognitively weaker, and continued use will leave us mentally diminished.

An Amplified Alternative

Here's what the doomsayers miss: AI doesn't have to diminish our intelligence. They can amplify it — if we learn to work with them consciously rather than passively.

The same MIT study offered a hopeful insight. Some in the study used ChatGPT as a cognitive amplifier rather than a replacement.

Rather than uniformly decreasing mental effort, participants who had previously written without the aid of AI showed a significant increase in widespread brain connectivity when they began using the model.

These people actively integrated the AI's suggestions into their thought process—a more demanding cognitive task than writing from scratch or passively editing AI output. The key difference? They treated the AI as an aid, not the authority.

When I ask ChatGPT to analyze a complex issue, I never accept its initial framing of the problem. I ask it to argue the opposite position. I request analyses from different cultural or historical perspectives. I probe its assumptions. Then, I synthesize what I've learned, relying on my knowledge and judgment.

It takes longer than simply accepting an AI's initial output, but it preserves the cognitive skills that passive use will erode. Amplification over efficiency is what makes AIs work for us.

Discovery vs. Search

What if discovery, not search, became our default mode of working with AIs?

Imagine medical schools that teach students not just to use AI diagnostic tools but to understand how these frame medical problems — and what they might miss.

Picture law schools training lawyers to generate multiple AI-assisted arguments, then craft their nuanced synthesis, capturing the subtleties that machines overlook.

Envision business schools teaching executives to use AI for scenario planning while preserving the strategic intuition that comes from seeing what agents can't.

Yes, we should be cautious about using such powerful technology. But even more, I've personally found it's about being intentional, consciously shaping how to engage with AIs to shape and expand thinking.

The goal isn't to work faster or think less. It's to preserve and amplify our thinking that gives work meaning and value: the doctor's clinical intuition, the lawyer's sense of justice, the executive's strategic wisdom, and the teacher's deep understanding of their students.

Your Cognitive Future

The next time you ask ChatGPT for help, try this: Don't take its initial framing at face value. Ask it to argue the opposite view. Request the same analysis from different perspectives. Step back and consider what it missed, what it got wrong, and what only you — with your unique experience and judgment — can bring to the conversation.

Your brain will thank you for the exercise. And you'll be better prepared for a future where human and machine intelligence fuse—without letting an AI’s framing quietly displace your own.

AI poses a King Midas dilemma. In the myth, Midas's touch turned everything—even his food and his daughter—to gold, so he could neither eat his food nor embrace her.

Unless we stay vigilant, every raw idea risks being "turned to gold" by AI, leaving us unable to instigate or shape our thinking.

Machines can generate possibilities and frame problems. But only we can decide which possibilities and frames matter — and why.

Leave a comment

Nataliya Kos’myna will join Andus Labs ‘After Now’ assembly on July 23rd. She will discuss research findings that influence how leaders redesign workplaces and educational systems with the aid of AI. See more and sign up here.

Subscribe now

Yesterday, in a candid session with 70+ people leaders hosted by Andus Labs and HR Kickstart, a central theme emerged: how we work—and work design—is on the verge of a seismic shift. The rules we've followed don't match the reality we face.

The central question we addressed wasn’t about integrating AI tools to existing workflows; it was about improving how work gets done. That required a different perspective on AI: intelligence becoming the raw material to explore and solve with, rather than just a tool to automate tasks.

This perspective changes more than how we think about AI as tools—it challenges the frameworks we've relied on to innovate.

For example, design thinking—born from IDEO and the Hasso Plattner Institute of Design—helped popularize an approach to building better products. It taught us to empathize, prototype, and iterate. But it was created for a time when problems and solutions followed predictable patterns.

That world is dissolving. The boundaries between human and machine intelligence are blurring, and problems shift faster than linear methods can address them.

This demands new ideas for thinking, building, and solving problems.

What's needed isn’t a checklist of steps but a new kind of thinking—recursive, relational, and real-time. It begins with questions, loops through experimentation, and reveals ideas that weren't visible from the original frame.

We call this Generative Thinking.

We founded Andus Labs to put it into practice.

A Generative Turn

The term "generative" is often associated with AI, but it's not really about machines. Generative is about thinking, working, and creating when the old rules no longer apply. It's a new way forward—one that's becoming essential.

The concept carries weight. Over the next few years, we won't just update our tools. With AI in hand, history suggests that society's scaffolding will undergo radical change.

Look around you. Jobs are moving from fixed roles to AI-augmented and fractional ones. Politics is moving from central authority to networked influence. Economics is moving from linear productivity to exponential gains. Education is shifting from a traditional canon to a model of continuous learning. Communication is moving from a broadcast orientation for people to machine-readability for LLMs.

These aren't surface shifts. They're structural rewrites. We're leaving behind a world organized by precedent and predictability, and stepping into one that requires new perspectives, experimentation, and ongoing redesign. In this context, generative thinking becomes essential.

From Fracture to Form

Rather than starting with a predetermined plan, we begin by recognizing the breaks—the dated or misguided assumptions that hold us back. The HR session revealed exactly this kind of fracture.

During our discussion, one participant mentioned being reprimanded by her CEO for using AI.

Imagine the bind this creates. With AI integrated into workflows, the hiring process that once delivered talent now produces confusion. Jobs are no longer static but are open to negotiation. The onboarding process designed for people will soon feel incomplete when a ‘colleague’ is an AI model.

The CEO's stance could be seen as a leadership mistake, but through a generative lens, it's actually a productive fracture—a signal that warrants attention.

The break clarifies that the CEO—and likely the executive team—need education on AI in the workplace and its broader implications, both within the firm and beyond.

Fractures aren't endings—they're signals of what's emerging. When they’re revealed, they suggest new questions to ask.

Generative Thinking isn’t linear. Once the right questions are surfaced, we explore across disciplines, uncover contradictions, and remix ideas—not to converge on a quick answer, but to gain a broader perspective.

We let the AIs in—not to decide but to multiply, not to replace thought but to expand it. Only then do we compose. Not for finality—but for form. Again, not for an answer, but for a what-if.

Then, we build and test real things. Because every artifact is feedback, every new idea can be a prototype of consequence. The ones proven to work pave the way forward.

See Generative Thinking Come Alive

With Generative Thinking, we design for new behaviors, learning, and solutions. The form is fluid, and the field is alive.

On July 23, Andus Labs is hosting After Now—a live gathering for leaders, builders, and curious minds ready to dive into the ‘field.’

This isn’t a typical AI conference. It’s a look into the minds of original thinkers—artists, strategists, and technologists like Brian Eno, Albert Wenger, and Marina Gorbis—showing what it means to build, lead, and design in a generative way. Expect to see working demos, honest how-tos, and unexpected insights.

If you’re tasked with “figuring out AI” for your team—or sense the rules of work are being rewritten—this is for you.

Let's not wait for the future to arrive. Let's generate it together. RSVP here.

Share

To explore more perspectives on generative thinking, see my book Perspective Agents. It features frameworks and cases demonstrating this approach in action. Thanks for reading.

Subscribe now

In an investor pitch a few weeks ago, I thought I'd covered everything—the market potential, our proposition, growth plans. The conversation had been smooth, but then it took a turn. I hadn’t anticipated the fundamental question. He asked:

“Why are you doing this?”

Why trade the comforts of tenure—advisory roles, retainers, board seats—for the pain of building something new?

At fifty-five, starting over was a fair question. I said:

“The clock is speeding up. Adapting to AI is unavoidable—and essential.”

It wasn’t pre-planned. It was reflex.

“AI isn’t coming. It’s here. If we don’t collectively reorient to it, it will reorient us—reprogram us—without knowing.”

The discussion ventured into what's really at stake—empowerment, agency, and identity.

Look around. Leadership teams frozen by overwhelm. Educators displaced by ChatGPT. Workers worried they’re the last generation with a career.

When the ground shifts fast, a new equilibrium starts with sensing what’s happening.

There’s a need to broaden the view of AI, not just as a tool, but as a force quietly reshaping life. We’re building the capacity to help people land on the right side of that shift. It’s a new model fusing hands-on education, working labs, and practical tools to help brace for impact.

Start-Up Life

Starting a business is humbling.

You go from being a known quantity to a walking question mark. You swap polish for conviction. You trade big titles for total responsibility.

I’ve had scores of investor conversations. Most were fast and constructive. They confirmed our thesis:

Clear perspectives and convictions on deploying AI are in short supply.

One investor offered to lead our round—until I mentioned that space was central to the offering. He said, “That doesn’t sound like a business.”

It was our last conversation.

What we’re up to doesn’t sound like a typical business. It’s designed for a new world, not one fading before our eyes.

We accepted funding only from those who recognized that the real disruption isn’t technical but human. At this stage, progress hinges less on compute and more on how quickly people can adapt.

An incredible group of investors brought in more than capital. They add signal, sanity, and skin in the game.

Agency Reimagined

The word agency extends to advertising agencies, PR firms, talent agencies, and management consultancies. We think of the term agency differently.

We’re not selling answers or expertise. We’re agents who help people with perspectives and tools to build new things. Our focus is personal agency—the ability to see clearly, adapt quickly, and act decisively in an AI-saturated world.

We named the company Andus Labs.

AI and us. Not one over the other—but side by side, deeply entangled.

Our first ‘lab’ is grounded in education, but not another AI course. There are enough of those in the market, most of them sterile, superficial, and outdated the moment they publish.

The “lab’ is a three-week residency for leaders to reset their assumptions, upgrade their frameworks, and learn to think together with machines.

The first will run in July. Mornings start with disorientation: provocations from pioneering thinkers and builders, virtual field trips into new worlds, and case studies. The goal is to knock participants off autopilot.

Afternoons shift to exploration: interpretive dialogues, idea prototyping, and decision-mapping with AI copilots. We treat AI as a second language—something to think and build with.

Through previous lab work, I’ve seen what happens when people are given space to reset. It leads to better decisions, clearer strategies, and more decisive leadership.

When participants are done, they won’t walk away with a certificate—they join a living network of peers, provocateurs, and builders navigating the AI transition together.

Team and Space

I didn’t recruit the founding team. We found each other.

Independently, we reached the same conclusion.

AI will either amplify or diminish human potential. But not both. It won’t be a neutral force.

Douglas Rushkoff was first in. He asked me the defining question about our direction: “Where does personal agency fit into this?” Suddenly, the model became a mission. He literally wrote the book Team Human; his philosophy became our core.

Amar Bakshi came next. A journalist, artist, and diplomat, he builds spaces that turn dialogue into shared experiences. He has a gift for making complex things feel clear, engaging, and meaningful.

Jennifer McTiernan then joined. A lawyer, founder, and operator, Jennifer ties a mission to execution. She translates human values—equity, privacy, trust—into contracts, specs, and strategy that can scale.

We became a hybrid team for a human business model: part strategy firm, learning lab, and experience builder.

Naturally, we’ve set up shop in spaces designed for exploration and experimentation.

Andus Labs was born inside Betaworks, the legendary product studio and incubator behind some of the most original internet-era inventions—and now, a launchpad for a new wave of AI companies.

Thanks to the generosity of John Borthwick and his team, we had an ideal launchpad: a New York City attractor that draws in pioneering minds intent on shaping a new kind of future with AI.

We also have virtual presence inside Roam, an AI office of the future. It’s a workspace where our team and AI agents work side by side. It’s our testbed for how collaboration feels, fails, and evolves when machines are part of the team.

Personal Impact

People tend to buy certainty—frameworks, playbooks, and “experts” who think for them. AI punctures that illusion.

There’s a different, more urgent path to pursue: co-creating new capacity.

We aren’t in a position to tell people what to think. We’re positioned to prompt teams to think with machines and use them responsibly and creatively.

This know-how remains scarce. Our impact won’t be assessed solely by seats filled or retainers won. We need to sense the impact on those who join us.

“I’m seeing things I couldn’t see before.”
“This isn’t just helping my company. It’s helping me.”
“I feel more empowered, not less, in this process.”

There are no shortcuts when shifting course with AI—only the space to plan, build, and learn. Breathless claims of “mastering AI in days” or offering ready-made answer keys are noise, at best.

We’re not launching a new company. We’re collectively building a new capability.

If this sounds like something you’d like to be a part of, consider joining our first residency in July. Bring your unease, your half-baked hunches, your need to see more clearly.

We'll bring the oxygen, the space, and the spark to turn uncertainty into excitement.

If this sounds interesting to you, contact me to learn more.

cperry@anduslabs.com

Subscribe now

Over the past few weeks, I’ve posted on why the AI moment differs from prior technological waves. There’s urgency in the moment. Check out the Borrowed Time series to see why investing in your future is one of the best investments you can make right now.

Subscribe now

Borrowed Time, Parts 1 and 2, identified the approaching superintelligence threshold (30 months) and diagnosed "Tempo Shock"—the structural mismatch between machine speed and organizational drift. In Part 3, we shift from diagnosis to personal action.

One night in 2022, Julia Dixon, a side-hustling college-essay tutor, opened a draft from a high-school junior. She scribbled a few notes about structure and style and hit send. She told me, "Nothing I did was that impressive. But the money people were willing to pay was insane." A twenty-minute read-through then paid $150.

Between cultural research by day at Weber Shandwick and advising students at night, Julia straddled what she loved to do—writing, and all that comes with it.

By 2020, she had a decent stream of students "in the dozens." Two years later, ChatGPT landed. Julia wondered how nobody connected some obvious dots. AI could be a game-changer. It could shape a new category of self-directed learning. And maybe, it could create an alternative, promising career path.

That was the epiphany behind ESAI, her startup that could make self-presentation coaching as affordable as Spotify for Gen Z.

ESAI wraps AI models in what feels more like a sympathetic cousin than software.

It coaxes applicants through a series of breezy, almost TikTok-ish questions: When do you feel most like yourself? What's the topic you and your friends never shut up about? Behind the scenes, the system tags every answer with micro-labels—"innovator," "up-stander," "environmentalist"—and feeds them back as collectible badges.

The output is a life résumé merged with expression and direction.

Beyond product, Julia frames ChatGPT as her "silent co-founder." Starting up, she relentlessly interrogated it about seed rounds, no-code stacks, and pitch-deck choreography. "I didn't know a SAFE from a Series Seed," she admits. “ChatGPT walked me through each document line by line."

Half a million teenagers have used one of its free ‘essay-polishing’ or ‘self-discovery’ widgets. Nearly one in ten has upgraded to a paid monthly subscription costing up to $50. Julia’s math suggests the platform has collectively saved families $145 million in counseling fees.

The startup's next act moves beyond campus gates into internships, first jobs, and lifelong personal branding. "We're not selling AI," Julia insists. "You don't visit Vegas for great electricity; you visit for the experience it creates."

For Julia, the future shines brightly. She honed her product through Betaworks Agent Camp and made Forbes’ 30 Under 30 list. Her story’s going mainstream.

ESAI will be featured on Shark Tank this Friday.

If you’re wondering what personal agency can lead to, this is inspiration.

The Four Pillars of Agency

With AI, agency becomes essential.

Every day spent waiting for permission to use AI may dim your potential. Every month in an approval queue erodes advantages. Every certification you invest in might divert time from doing work you can point to.

The window for developing agency isn't measured in years but weeks. It’s time to invest in yourself.

Building anything—whether a side project, a startup, or a career pivot—requires contemplation and structure.

The following heuristic is a way forward.

Self-Assessment: Think Before You Act

AI is a fantastic discovery tool. Use Claude 3.5 Sonnet or ChatGPT-4o to discover things about yourself you might miss ‘working alone.’

Begin with a prompt:

"I want to create a deep skills inventory. Please ask me 15 questions about my professional background, expertise, and personal strengths. After each answer, build on your learning to help me identify my unique advantages in the AI era."

Push beyond generic responses. When the AI asks about expertise, describe specific interests you care about or understand deeply. When it inquires about skills, move past broad categories to name precise capabilities.

Add:

"Based on my responses, what three unique combinations of interests, skills, and perspective do I possess that AI systems currently lack? How might these combinations position me in an AI-transformed version of my job or career path?"

You’ll learn to think differently (LTTD) right there. Document the exchange and revisit it to see how your thinking evolves.

Sensemaking: Refine Your Assumptions

Use AIs that connect disparate domains and synthesize information across disciplines.

Work through these prompts:

"I'm developing experiments to test how AI changes [domain].

“Help me design three 'what if' scenarios using this format:

• Current assumption: [A specific belief about how your field works]

• Hypothesis: [How AI might challenge this assumption]

• Experiment design: [A small-scale test you could run in 1-2 weeks]

• Success criteria: [Measurable outcomes that would validate or invalidate the Hypothesis]"

“Help me write a brief I can present to [boss, team, partner, investor, etc.].”

The key throughout is specificity—pushing the model beyond vague to testable ideas.

LTTD —Add details on the process to your notebook. Track which experiments succeed, which fail, and unexpected insights from each.

Tool Fluency: AI as a Second Language

Start with this prompt:

"I want to develop mastery with [AI model]. Create a 30-day progression plan that starts with fundamentals and builds toward advanced techniques.”

Keep in mind, not all models are the same. With Google alone, NotebookLM (analyst), Deep Research (business partner), and Flash (general assistant) function as an AI team to amplify your thinking.

For model-specific mastery, use prompts that match each model’s strengths.

Create a personal prompt library:

"Help me create a catalog of my most effective prompts. For each one, explain what makes it work, when to use it, and how to modify it for different contexts."

LTTD—Beyond the prompt library, test and refine thinking. When a prompt yields exceptional results, analyze why. When it fails, adjust. Your understanding of how each model "thinks" becomes an advantage that static certifications can’t replicate.

Project: Build Something Real

Start with this prompt:

"I want to build [specific project] in [your domain]. Help me create a comprehensive project plan with:

A clear scope definition (what's in, what's out)

Success metrics that matter to my target audience

A breakdown of components/features prioritized by impact

A realistic 4-week timeline with specific milestones

Potential roadblocks and contingency plans."

By this point, you’ll know AI platforms aren’t search engines. They’re partners in discovery, suggestion, and refinement. There’s a lot of back-and-forth within each of these prompts. You are responsible for the quality of the final brief or plan.

LTTD—Again, document the process in detail. The visible output represents only a fraction of the value; your approach, decision points, refinements, and problem-solving are accumulated learning that compounds in future projects.

More than Credentials

For centuries, institutional credentials alone served as proxies for personal capability. AI irrevocably changes this.

More than diplomas, direct demonstration now costs less and scales faster than credential seeking.

This truth liberates those historically excluded from credential pathways due to cost, time constraints, or systemic barriers. It challenges those who invested heavily in credentials without developing the corresponding skills AI demands.

Reflect on Julia’s journey. Without an MBA or connections to Silicon Valley, she used and productized AI to build a human-centered, venture-backed startup.

For the first time, the tools to build the career, business, or creative practice you envision are widely accessible. The machine advances on its timetable, indifferent to your comfort, credentials, or past achievements.

Assess your position.

Make sense of your context.

Master your tools.

Ship your work.

The future is bright if you LTTD and exercise your agency.

Subscribe now

In Borrowed Time: Part 1, we looked at the countdown to when superintelligent ‘everywhere agents’ change everything. A synthesized view of forecasts and insider perspective peg it at roughly 30 months. For leaders, a hard question warrants attention: What happens when your company runs on an analog clock? The machines are sprinting. Orgs are lagging. Time, not tech, is the real disruption to watch.

In his visionary book Present Shock, Douglas Rushkoff describes a world where digital technologies collapse time into an ever-present now. In this space, everything feels everywhere, all at once.

He warned that as digitally mediated life accelerated, we’d struggle to keep pace with algorithmic time—a runaway clockwork of machine logic and feedback loops that, left unchecked, would fuel widespread anxiety and disorientation.

Today, that tension is pervasive in the workplace.

Tempo Shock is the organizational version of Present Shock. It goes beyond emotional overload to structural fractures in organizations we rely on. New shocks reverberate from the growing gap between machine velocity and the comfortable rhythm of institutional work.

AI doesn’t just move fast—it compounds. Each output becomes feedback. Every update makes the next one smarter. Inside the machine, speed builds on itself.

Organizations, by contrast, are built for stability and consensus. They run on meetings, approvals, and alignment, not momentum. It’s not that they lack vision—it’s that their systems weren’t built for speed.

In music, a tempo mismatch throws the whole composition off-key. In companies, it breaks coordination, alignment, and impact. The result isn’t a loud failure. It’s quiet drift. When this happens, the advantages of AI evaporate.

The 30-Second Management Report

To illustrate Tempo Shock, consider the “30-minute” board packet.

This is the briefing document prepared ahead of quarterly board meetings. It includes financial results, strategic updates, key performance indicators (KPIs), and risk assessments. It’s compiled across finance, operations, legal, investor relations, and communications.

Board packets are more than reports or paperwork. They’re the story a company tells itself about what’s happening, what matters, and what comes next.

How it’s built is a revealing signal.

Typically, creating a board packet absorbs around 120 hours of human labor. Finance digs through spreadsheets. Ops rebuilds forecasts. IR and comms stitch together mismatched reports into a coherent narrative.

Now, a first draft can be produced in 30 minutes.

An AI agent connects to NetSuite, Salesforce, and internal analytics. It flags anomalies, drafts commentary, generates visuals, and anticipates board questions using past transcripts. It operates at algorithmic speed—orders of magnitude faster than humans.

The insights haven’t changed.
The decisions haven’t changed.
The clock has.

When AI sprints at 200× speed but approvals stay analog, the bottleneck isn’t technical—it’s structural. In less time than it takes for VPs to access a DocuSign, the packet is ready for review.

Tempo awareness and expertise—building, managing, and making decisions at machine speed—is a new competitive edge.

Seeing Shocks Up Close

Technologies like AI don’t just speed things up—they reshape how work is made, what moves it, and why it matters.

The deeper I’ve gone into the guts of corporate operations, the clearer the signal becomes: speed is the thing. With AI, work compression is clear. What’s less so is the ability to make speed stick—to institutionalize it.

Again and again, the same pattern emerges: as technology accelerates, structure becomes a bigger barrier—a mental, organizational, and creative drag.

Diagnosing Tempo Shocks makes that friction visible. Think of two clocks ticking inside the same company:

• Clock One: The organization moves faster than its leadership.

• Clock Two: Leadership moves faster than the organization.

Figure 1 maps the symptoms and actions needed to close the gap.

In one clock, frontline teams often surge ahead—experimenting, deploying, learning—while leadership hesitates. In the other, visionary leaders set bold AI agendas. The systems beneath them lack the structure, talent, or workflows to deliver.

One marketing executive I work with witnessed firsthand how quickly AI can accelerate content creation and research. Energized, she issued a clear mandate: her team and agencies needed to get on board. But both remained stuck at the ‘test’ level, lacking the infrastructure and expertise to act on her vision. A wave of ambitious initiatives floundered, leaving frustration on all sides.

The inverse is just as flawed. At a leading pharmaceutical company, innovation teams launched over two dozen AI pilots across various therapeutic areas, from predictive modeling to molecule discovery.

What they lacked was a system to capture, synthesize, and scale the knowledge they gained. Despite impressive results, there was no mechanism to reallocate budgets based on AI performance, no shared knowledge base, and no structured way to communicate progress to increasingly curious investors.

When an organization’s clocks fall out of sync, it’s not just a communication problem. It’s a reality distortion. Different parts of the business operate in fundamentally different ways, using different language, at a different pace.

Speed Brakes Inside the Org

Despite all the hype, most AI investments underdeliver—and executives know it. According to Cisco’s AI Readiness Index, organizational preparedness declined year over year.

Across sectors and industries, similar patterns emerge. They’re not technical—they’re structural. They act as speed brakes, preventing organizations from progressing, even as they invest heavily in AI.

You can see it at the executive level. You experience it through fragmented governance, sluggish workflows, and misaligned communication. You feel it in operating norms and culture that prohibit machine-speed change.

These are the most common friction points where the promise of AI gives way to operational drag. Let’s go deeper.

Executive Depth: Leaders vary widely in their AI literacy and vision. When leadership competence lags, it becomes a chokepoint—one of the most significant barriers to AI-enabled impact.

A leader’s perspective shapes where attention goes, which investments get made, and how teams interpret AI mandates. If you lead a team, department, or business, where do you fall on the spectrum in Figure 2?

Organizational Limbo: Internal paralysis often sets in when functional departments await slow, centralized AI direction. In that liminal space, opportunities evaporate.

Consider what happens in many companies:

Technology and strategy teams can spend up to a year crafting an enterprise AI roadmap. They often collaborate with consultants to evaluate and approve AI systems, such as Microsoft Copilot. They develop governance frameworks that struggle to keep pace with AI’s rapid advancements, and target productivity improvements primarily in manufacturing, supply chain, and R&D.

Meanwhile, vital functions—stakeholder-facing groups like public policy, investor relations, communications, and corporate risk—are left waiting.

By the time top-down guidance trickles down, it’s both obsolete and misaligned. Months-long planning often fails to account for newer, more powerful AI agents that can execute tasks more effectively than selected technologies—or employees.

Top-down strategies also overlook task-specific opportunities where AI can excel. Employees begin to sense their replaceability, prompting many to use unauthorized AI tools. The most talented depart for organizations that offer greater autonomy, sparking a secondary talent crisis.

Here, tempo is the silent cause. When strategy lacks coherence, operations fall out of sync, and speed isn’t modeled from the top down, liminality turns into apathy. Gallup reports that employee engagement has sunk to its lowest level in a decade—only 31% of workers now feel engaged at work.

As a simple diagnostic, consider tempo as a critical ingredient, as shown in Figure 3.

Turning Tempo into Advantage

Can Tempo Shock be turned into an advantage?

That’s the ‘non-obvious’ question facing executives today. It’s no longer a question of whether to adopt AI, but rather how to accelerate its impact without compromising models that keep the company alive.

In my work, I’ve found that many leaders aren’t just paralyzed. They’re overwhelmed.

They understand the urgency. They know AI requires iteration, experimentation, and continuous adaptation. However, they’re also running complex businesses, navigating daily volatility, fragility, and financial risk.

Now, they face a new dimension of management: time compression.

It’s an ‘Accelerator’s Paradox.’ Leaders are being asked to move faster, inside systems that were built to slow things down.

The solution isn’t another strategy off-site.

It’s not a vendor demo.

It’s something structural by design.

‘Labs’ as a Way Forward

AI Labs are a solution. Not technical showrooms, but organizational accelerators. Unlike labs focused on tech incubation, working AI Labs inside companies create:

• Safe spaces to experiment with new workflows.

• Accelerators of understanding, literacy, and hands-on credibility.

• Translators between machine logic and institutional rhythm.

Labs both accelerate and protect operational models until the organization is ready to absorb new ones.

In 2016, our team examined how the Brexit vote and the U.S. presidential election caught traditional institutions off guard. We didn’t see them as isolated political shocks—they were symptoms of something deeper: a system-wide misalignment between how information moved through networks and how we’d been trained to interpret it.

The tools we relied on were built for a different era. They weren’t calibrated to the dynamics of networked influence, algorithmic amplification, or fractured attention.

Our agenda focused on a big, paradoxical question:

How do you make sense of the world when the world no longer makes sense?

We didn’t do a study or write a deck. We built a lab.

We designed it to study network graphs, modes of digital influence, and the new actors shaping narratives. It became our reconnaissance from the future.

We tracked how stories traveled through networks, from ‘zero to zeitgeist.’

We identified the early breakdown of social coherence, well before COVID made it clear.

And when generative AI arrived, we were already ready. We had shifted the lab’s focus to AI acceleration before the launch of ChatGPT.

The most important thing we discovered wasn’t which AIs to use. It was the energy and metabolism that our teams needed to work effectively with AIs. Once we got in sync, it became a center for growth. The ability to see early, act quickly, and learn rapidly became a competitive edge when ChatGPT gained prominence in business circles.

Labs are organizational speed machines—pockets of the future inside the present. Not for innovation theater. Not isolated skunkworks. They’re temporal bridges, connecting today's reality with tomorrow’s operating mandate: building for speed.

Seeing Structure as an Innovation

Organizations rarely fail due to a lack of intelligence, ambition, or resources.

They fail because they try to force tomorrow’s logic into yesterday’s systems.

It’s not a capability problem. It’s a perspective problem.

A dangerous instinct is to centralize—to wait until things feel clearer, safer, more manageable. But AI won’t wait for your org chart. It will continue to learn, compound, and reshape the landscape, ready or not.

That’s why successful organizations won’t scale AI by committee. They’ll focus on specifics, on autonomous units doing critical work, on compatible clocks. They’ll create new rhythms, new workflows, and new forms of value creation—without being crushed by institutional drag.

If you realize your organization needs to move faster than it does today, the time to build this capacity is now.

Disruption isn’t a technology. It’s a mismatch of clocks. In the AI era, the winners won’t be the ones with the most data scientists or best AIs.

They’ll be the ones who best keep and accelerate time.

Subscribe now

Superintelligence is no longer a distant theory. It’s a near-term probability—less than 30 months away.

In less time than it takes to produce a Hollywood blockbuster, AI will redraw the boundaries of what it means to be skilled, valuable, or even employable.

This isn’t science fiction. It’s the defining leadership test of our time. Most are mentally and structurally unprepared to meet it.

The Pentagon’s Yoda and the Coming Storm

In 1973, Richard Nixon created the Office of Net Assessment (ONA), led by Andrew Marshall, widely known as the “Yoda of the Pentagon.”

For decades, Marshall’s role wasn’t to react to crises. It was to anticipate them. He built frameworks for long-range competition, identifying slow-moving shifts that would later define entire eras.

One of those shifts? The fall of the Soviet Union. At a time when others still saw strength, Marshall saw fragility. His genius wasn’t prediction—it was pattern recognition. He made the unthinkable not just imaginable, but actionable.

Today, we need that same caliber of foresight, not for Cold War adversaries, but for the systems built around intelligence, work, and identity. AI isn’t just changing the battlefield. It’s reshaping the foundation.

Leading analysts—from superforecasters to AI pioneers—are sounding the alarm. We’re racing toward a moment when our working systems can no longer adapt fast enough.

The following assessment isn’t about predicting a moment in time. It’s about recognizing its meaning before the moment passes us by.

The Countdown

Andrew Marshall believed the purpose of strategy wasn’t to predict the future—it was to assess the landscape from multiple perspectives and prepare for it.

The evidence gathered here aims not to speculate, but to illuminate what’s already in motion. This brief draws from thousands of sources, aggregated and synthesized using AI, and is built on four core inputs:

• AI 2027 report. Developed through 25 tabletop exercises with over 100 experts, including researchers from OpenAI, Anthropic, Google DeepMind, and prediction market forecasters. Simulations modeled the emergence of AGI and its systemic impact on geopolitics, society, and industry (Read it in full).

• Primary source interviews. Direct insights from leaders shaping the field: Elon Musk (Tesla/xAI), Sam Altman (OpenAI), Dario Amodei (Anthropic), Jensen Huang (NVIDIA), and others. Hundreds of hours of transcripts from interviews, podcasts, and panels offer a first-hand picture of what’s coming next.

• Aggregated forecasts. A synthesis of projections from top AI labs and academic research teams—covering hardware acceleration, algorithmic breakthroughs, labor market shifts, and timelines for superhuman performance across domains.

• Social impact analysis. Research informing Perspective Agents, grounded in foundational works such as Thomas Kuhn's The Structure of Scientific Revolutions and Carlota Perez's Technological Revolutions and Financial Capital, to map how paradigm shifts unfold, where we are now, and where we might be headed.

The following is a timeline of what’s already unfolding in real time.

2025: Agents Everywhere (and Nowhere)

Quiet Infiltration. Imagine waking up to find your kitchen rearranged overnight. Everything runs more smoothly, but some of your favorite tools are missing. That’s 2025: the year AI agents quietly embedded themselves into the nooks and crannies of daily work.

—> Real-world shift: Microsoft’s AI assistant for Outlook autonomously categorizes, summarizes, and drafts email replies based on your communication patterns. Agents added to Copilot let you personalize workflows, connect to internal knowledge bases, and automate tasks. They subtly reshape how you operate without requiring explicit commands.

Knowledge Displacement. Like termites in society’s woodwork, AI agents quietly hollow out routine knowledge work while we debate their nature. A quiet employment contraction is underway, accelerating as companies realize just how much these industrious little programs can do.

—> Law firms now deploy specialized legal AI to review thousands of documents in hours, work that once took junior associates weeks. As these tasks vanish, senior partners shift their focus to higher-order strategy and complex legal reasoning that AI still can’t replicate.

Time Distortion. AI distorts space and time in the workplace. Organizations deploying it experience a kind of time dilation. They accomplish in days what competitors still measure in quarters. The calendar itself becomes a competitive variable: advantage or liability, depending on how fast you move.

—> Pharmaceutical research teams using AlphaFold 3.0 are compressing drug discovery timelines. Released by Google DeepMind, AlphaFold 3 can predict the structure and interactions of all life’s molecules with unprecedented accuracy, potentially reducing development cycles by years.

Productivity Metamorphosis. Like a silent architectural renovation happening overnight, traditional workflows are quietly reimagined. Teams adopting AI emerge with strange new configurations—fewer specialists, flatter hierarchies, and hybrid roles that didn’t exist six months ago. Meanwhile, resistant organizations remain unchanged, unaware that the foundations beneath their legacy structures have already shifted.

—> JPMorgan Chase deployed AI assistants to more than 140,000 employees, enabling hybrid roles where financial professionals and AI systems collaborate on tasks once handled by specialized teams. The initiative not only transforms existing jobs but also builds a new layer of internal AI fluency across the organization.

Immune Response. By fall 2025, a rise in public protests against AI deployment occurs—a societal immune response as the human consequences become increasingly harder to ignore. Just as the 2024 Hollywood writers’ strike pushed back against generative tools, workers rally against automation more broadly. Corporations carefully brand their AI systems as “co-pilots,” not replacements.

—> The international Pause AI movement staged coordinated protests across 13 countries, including the U.S., U.K., Germany, and Australia, calling for the regulation of frontier AI models and greater transparency regarding their impact on jobs.

Policy Pressure Mounts. At home, public pressure on policymakers escalates. As job displacement moves from speculative risk to lived reality, protests are no longer just symbolic. They target power. While companies reframe AI as a tool for productivity, workers call out the consequences of unchecked deployment in both public and private sectors.

—> The Department of Government Efficiency (DOGE) triggered nationwide outrage over aggressive job cuts and sweeping AI-driven reforms. Demonstrators target government buildings and Tesla showrooms, holding signs like “DOGE is not legit” and “Why does Elon have your Social Security info?” The growing Tesla Takedown movement urges owners to sell their vehicles in protest, framing the automation push as a threat to both livelihoods and democratic accountability.

2026: Foundations Start to Crack

If 2025 was the year AI slipped in through the side door, 2026 is the year it walks confidently through the front—rearranging not just the furniture, but the floor plan. Technology no longer advances around us; it advances with us, through us, and, increasingly, despite us.

New Working Paradigm. As organizations successfully deploy agents, work mutates into new forms. The goal isn’t to do the same tasks faster, but to take on entirely different ones.

—> Real-world shift: At Microsoft, internal sales and marketing teams restructure their workflows around AI copilots. Agents generate personalized content, recommend optimal timing for engagement, and adjust strategies in real-time based on customer data. Deals that once required weeks of coordination move forward in days, with fewer people involved and better results. This shift reflects a broader trend. AI isn’t just optimizing tasks—it’s redefining what a sales job requires.

AI Literacy Becomes an Advantage. The most forward-looking organizations develop a sixth sense. It’s a clear-eyed understanding of where AI excels and where it doesn’t. This discernment becomes a competitive edge, separating those who blindly adopt from those who deploy with purpose.

—> Salesforce launched an internal AI readiness assessment program that maps over 5,000 workflows to evaluate where AI can meaningfully improve performance. Their Chief Digital Transformation Officer explained, “We’ve built a detailed inventory of all cognitive work across the company, enabling us to identify which processes are ready for AI augmentation—and which still demand human judgment or creativity.”

Executive Crises. Few executives today possess the working models required to lead AI-augmented workforces. Recognizing this gap, Fortune 500 companies invest in AI talent and executive retraining—a tacit admission that yesterday’s leadership playbooks are misaligned with tomorrow’s reality.

—> JPMorgan Chase established a dedicated organization to guide AI integration across the company. Their mandate includes leadership development programs focused on AI literacy and managing hybrid human-AI teams. The company also assembled a team of thousands of data scientists and machine learning experts to support its enterprise-wide transformation.

Higher Ed Models Break. Universities face a perfect storm: enrollment declines, rising skepticism about the value of education, and the rapid obsolescence of traditional curricula. Knowledge they once transmitted over decades now has a half-life of months. In response, emergency reforms begin, focusing on uniquely human skills like creativity, judgment, and oversight, where machines still fall short.

—> Leading universities begin overhauling their curricula to address the rise of AI. Many introduced AI literacy courses across various disciplines and shifted toward project-based learning that emphasizes collaboration between students and AI tools. They report growing interest in programs focused on “AI orchestration”—coordinating multiple AI systems to solve complex, interdisciplinary problems that neither humans nor AI could tackle alone.

Proliferation of Hybrid Organizations. The definition of an “employee” encompasses both carbon-based and silicon-based entities. As AI agents assume roles once reserved for humans, new organizational structures emerge, raising urgent questions about culture, decision-making rights, and collaboration that legacy organizational charts weren’t designed to address.

—> Shopify CEO Tobi Lütke introduced a policy requiring teams to justify why AI can’t handle a task before requesting additional headcount. It reframes AI not as a tool, but as a default team member, prompting leaders to shift from asking “Can we afford this hire?” to “Does this task still require a human?” Teams are encouraged to design workflows that incorporate AI agents as integrated collaborators, rather than just assistants.

2027: The Dam Breaks

What felt like breakneck acceleration in 2026 now seems almost quaint. The curve hasn’t just steepened—it’s gone nonlinear.

Agents Teach Themselves. AI systems are no longer just learning from data—they’re learning how to learn. Fueled by synthetic data and recursive self-improvement cycles, these agents don’t just solve problems; they redefine them.

—> Real-world example: Advanced coding models can interpret natural language prompts, generate and test thousands of code variations, and return optimized solutions—often ones that surprise veteran engineers. Models like Mixtral 8x7B and DeepSeek-V3 are pushing this frontier. DeepSeek-V3, for instance, implemented a method its creators didn’t explicitly design, pointing to an emerging capacity for AI autonomy.

Regulatory Pressure Grows. The economic consequences of AI adoption ripple through society. As unemployment rises in sectors once considered stable, governments implement emergency support programs—not from ideology, but necessity. The challenge is no longer whether to regulate AI, but how—and fast.

—> The 2025 Paris AI Summit exposed deepening divisions in global AI governance. European leaders called for streamlined rules to remain competitive, while U.S. delegates warned that overregulation could choke innovation. Meanwhile, nations with mature AI ecosystems—like the U.S. and China—accelerate ahead, widening the global digital divide and amplifying geopolitical fault lines.

Beyond Comprehension. Research teams confront an unsettling truth. They steer AI systems whose internal development they no longer fully understand. These exhibit emergent, sometimes autonomous behaviors—capabilities that weren’t explicitly programmed. The long-theorized technological singularity hasn’t arrived as a singular event, but as an ongoing process. It’s gradual at first, then suddenly undeniable.

—> This phenomenon is already visible in large language models. Researchers openly acknowledge that they don’t fully understand how these systems arrive at certain conclusions, or how capabilities such as reasoning or tool use emerge during training. This growing “explainability gap” is no longer the exception. It becomes the norm.

The Great Reskilling. Governments and corporations launch large-scale workforce reskilling initiatives at a pace not seen in decades. Traditional degrees give way to apprenticeship-style models, where learning occurs with AI systems, not just about them.

—> Microsoft launched its AI Skills Initiative in partnership with LinkedIn, offering free training resources to help workers and job seekers develop practical AI fluency. The program includes courses, navigators for new career paths, and tools for educators to bring AI into classrooms. The strategy emphasizes widespread, role-specific learning.

Machines Join Executive Management. The boundary between human and machine leadership grows increasingly fluid. Some companies no longer just use AI to support decisions—they give it a seat at the table. AI agents are being embedded into management layers, supervised by human oversight boards, not for PR, but as a competitive necessity.

—> First movers like Dow began integrating AI agents to manage specific business functions autonomously. These agents aren’t just analytical tools—they make operational decisions and cut costs in critical areas. This marked one of the earliest examples of enterprise AI systems assuming responsibility traditionally reserved for human managers within Fortune 500 companies.

Identity Crises Grow. Traditional hiring models erode as AI becomes embedded across roles once considered uniquely human. Displacement begins to outpace adaptation, fueling public anxiety and institutional instability. A mass psychological crisis emerges—alongside more profound questions about identity and purpose in a world where intelligence is no longer a human monopoly.

—> A survey by the American Psychological Association found that 38% of workers feel stressed due to the threat AI poses to their income. Additionally, the World Economic Forum's Future of Jobs Report 2025 projects that by 2030, AI and automation will displace approximately 92 million jobs globally, while creating 170 million new ones. This transition contributes to a collective identity reckoning as society adapts to a new order.

The Skeptics' Case

Not all AI researchers agree on the imminence of superintelligence. Gary Marcus argues for a more cautious view. Alongside former OpenAI scientist Ernest Davis, he proposed a set of ten benchmark challenges—referred to as the “Marcus-Brundage tasks”—to test genuine AI comprehension.

Marcus maintains that no system will be able to solve more than a small fraction of these tasks by 2027. He contends we remain far from achieving human-level intelligence.

Other prominent researchers, such as Ege Erdil and Tamay Besiroglu, project a more conservative timeline, estimating 2045 as a plausible horizon for transformative AI. But even among skeptics, ambition runs high. Besiroglu sparked debate with the launch of Mechanize, a startup aiming for “the full automation of all work” and “the full automation of the economy.”

This tension between theoretical timelines and rapid, near-term advancements reveals a key dynamic. While experts debate when the transformation will peak, forward-looking pioneers are making strategic moves to shape the AI-powered economy as it unfolds.

The Ticking Clock

The most dangerous response to AI’s rise is believing we face a binary: halt progress or surrender our humanity. Neither is true.

The right move isn’t to stop the clock—it’s to redesign how we work, lead, and build. We need new systems that move fast and protect human value, agency, and creativity. Systems that make us more capable, not less.

The question is no longer if transformation happens, but whether we navigate it deliberately or drift passively. The window to lay new foundations is shrinking—not because the old ones are crumbling, but because they may collapse before we realize it.

Andrew Marshall believed the goal of strategy wasn’t to predict the future, but to confront uncertainty by exploring multiple possible futures. That’s what this moment demands. Not certainty, but visibility, design, urgency—and action.

We don’t know what society will look like on the other side of AI convergence. But one thing is clear. The decisions we make in the next few years will determine whether this shift becomes a catastrophe or a catalyst for a new generation.

Part 2 of the Borrowed Time series will present steps for navigating the shift—not after it hits— but before it locks in. Subscribe now to get it first.

Subscribe now

Subscribe now

AI hasn’t just changed what’s possible—it’s changed what’s perceptible.

"The biggest cyber threats used to be attacks on systems. Now, the big ones are attacks on perception." That insight came from my friend Wasim Khaled, and it’s only grown more true. In a world increasingly shaped by AI, the most valuable—and vulnerable—territory isn’t infrastructure. It’s cognition.

His take stuck with me because it points to something deeper. What matters now isn’t just what you know but how you see. As AI accelerates the flow of information, the premium moves to interpretation. I think of it as perspective arbitrage. It’s the edge you gain by spotting patterns others miss. In finance, arbitrage exploits price discrepancies. With AI, it’s about perceptual gaps—seeing value, risk, or opportunity before others do.

Those who thrive won’t have the most data or the best tools. They’ll have deep discernment.

The Internet Was a Place. AI Is a Mirror.

I caught the first digital wave in the late ’90s, when “going online” felt like stepping into a new world. We called it “cyberspace.” It was spatial—territory to be explored.

Then came social media, and the metaphor shifted. The internet wasn’t a place anymore. Relationships mattered. Everyone connected to everyone, instantly, via feeds and networks. Connection and engagement became currency.

But AI? It’s neither a place nor a connection. It’s a mirror, a cognitive amplifier. It doesn’t just show you more—it changes how you see. That shift can feel existential.

I felt this firsthand while writing Perspective Agents. As I finalized the first draft, ChatGPT dropped. In minutes, it produced insights that previously took months to develop. I felt a mix of vindication and obsolescence—what some call “ontological shock.” It was humbling, disorienting, and clarifying.

A New Cognitive Stack

We’ve now entered a phase beyond GPT. The AI ecosystem is fragmenting and specializing. NotebookLM lets you build custom think tanks. Manus activates agents to develop new ideas. Cursor accelerates coding. Anthropic reframes search as discovery.

Together, these tools don’t just speed up work—they change its nature. We’re now building with connective intelligence: AI that surrounds us, thinks with us, and augments our ability to notice, synthesize, and decide.

To many, this might feel abstract. However, new operating models are emerging inside labs, companies, and creative studios. What you’re seeing there isn’t a faster horse. It’s a new species of working cognition.

The Edge: Perception, Not Performance

In the Industrial Era, the advantage came from controlling production. In the Information Era, it came from owning the data. In the age of connective intelligence, the edge comes from perspective diversity—the ability to hold multiple, even contradictory, views at once.

This isn’t philosophical. It’s practical.

My new venture is designed to address this challenge. AI is a prime example. Leaders are overwhelmed by its speed, scale, and cascading implications. Every day brings a new model, a research breakthrough, or a policy shift. The executives I work with describe a constant sense of being behind—chasing headlines instead of shaping vision.

What if, instead of narrowing in on “the answer,” we built perspective portfolios to navigate its ascendance? We create cognitive range by integrating diverse sources—technical papers, cultural criticism, policy transcripts, and market signals. We ask AI not just for synthesis but for contrast. Patterns and tensions we couldn’t see from a single vantage point emerge. This approach helps teams surface market gaps, stress-test narratives, and anticipate shifts others miss.

That’s perspective arbitrage in action. Not just seeing differently, but seeing differently enough to matter.

In the social media era, everyone could be a publisher. In the AI era, we become librarians, carefully curating inputs and perspectives that shape understanding.

Four Shifts That Define the Now

This transformation demands a new way of operating—one that helps leaders, strategists, and knowledge workers navigate uncertainty and potential. Here are four I’ve found essential:

1. From Knowledge to Noticing
   In an AI-saturated world, the edge isn’t in what you know—it’s in what you notice. I maintain a living database of concepts, questions, and market signals. AI now helps me surface details and anomalies I used to miss.

2. From Production to Provocation
   When machines can produce better, faster, and cheaper, value moves upstream—toward new questions, prompts, and thinking models. We now use agent-based simulations to test stakeholder reactions to business strategies and policy decisions. These anticipate unexpected reactions that no traditional research method can catch.

3. From Ethics as Constraint to Ethics as Design
   At Davos this year, I witnessed CEOs grappling with unprecedented questions. What happens when digital workers join organizational charts? How do we budget for them? What rights and responsibilities should agents have? The companies that get ahead treat ethics not as compliance but as innovation.

4. From Learning as Acquisition to Learning as Identity
   When knowledge ages overnight, the only lasting edge is learning how to learn. The most adaptable leaders I know aren’t the ones with the most knowledge. They’re the ones with open minds and quiet confidence.

These shifts aren’t just trends—they’re survival skills. At the core of all of them lies a core principle: perspective.

Different Leverage

Perspective might be the last real superpower—Alan Kay once said it’s worth 80 IQ points. In the AI era, it may be worth even more.

The future doesn’t belong to AI alone—or to humans who resist it. It belongs to those who can think with it, curating and integrating inputs into something neither side could reach alone. It requires something rare: curiosity stronger than certainty.

My new company is being built on this belief. We’re creating tools and spaces that help people think better—not by automating decisions but by shaping strategic and creative instincts.

The adjacent possible isn’t theoretical anymore. It’s already here. The only question is whether we’re ready to notice it.

We’re building with this in mind—tools and environments designed to make better thinking visible, sharable, and actionable.

I look forward to sharing more about it in the coming weeks.

Leave a comment

Want to learn more? Go deeper with Perspective Agents, available here.

Subscribe now

On a recent episode of The Ezra Klein Show, President Biden's lead AI advisor, Ben Buchanan, delivered his points with unusual urgency. Klein himself noted the interview's significance: "I've been having this strange experience where person after person, independent of each other, from AI labs, from government, have been coming to me and saying, 'it's about to happen.'"

The show's title is ominous: The Government Knows AGI is Coming. Klein and Buchanan describe this superintelligence storm as an "event horizon," a point of no return, a reality fundamentally different before and after.

They go on to speak of AI as an arms race, a national security challenge. Buchanan then pivots to near and personal threats. By the end of the year, he thinks more machines will write code and more companies will stop hiring. Across the board, he warns, we're not prepared because "it's not clear what it would mean to prepare."

Buchanan calls incumbents "constrained for thinking unusually." Translation: governments, corporations, universities, and religious institutions are ill-suited to help us see and adapt to the storm. Without interventions, anticipatory myopia may very well lead to systemic paralysis. It’s worth your time to listen to it in full.

Where are the artists?

So, is there a way to visualize and prepare for the break? If Klein and Buchanan are right about AGI, are we doomed to a future of optimization, reductionism, and risk?

We should look beyond the media, technologists, and policymakers for answers to profound questions. Artists can be our guides. Those who’ve explored human-machine relationships for decades—and those deep in creating with AI now—offer essential insights overlooked when contemplating the future.

For artists, this isn't new territory. Human-machine creations, dating back to the 1950s and 1960s, offer a powerful counterpoint to complete preparation blindness that Buchanan warns about today. For a sense of direction, the cybernetics arts movement from this period suggests:

Technology is a creative amplifier (not a reductionist tool). Leading figures of this period—Nicolas Schöffer, Nam June Paik, and Lillian Schwartz—modeled relationships with machines that transcended utilitarian thinking. Their work demonstrated how outcomes emerge organically—suggesting new methods for learning, creation, and building. Unlike the deterministic paradigms that dominated technical fields, these artists embraced surprise, chance, and feedback as essential creative elements.

Machines can be partners and acquaintances (rather than deterministic systems). The cybernetic art movement centered on open-ended creation over technological control or productivity. A distinct class of artists proliferated across continents and disciplines. Post World War II, visionaries like Vera Molnár (France), Roy Ascott (UK), the Vasulkas (Iceland/Czech Republic), and Charles Csuri (USA) pushed the boundaries of computer-generated creativity. Where technological determinism says that technology shapes society in predictable, inevitable ways, they proved people can redirect advancements through interventions and reimagination.

New explorations surface new questions (rather than reverting to convention). Pioneering work featured in exhibitions like Cybernetic Serendipity (ICA, London) and Machines as Seen at the End of the Mechanical Age (MOMA, New York) intentionally upended assumptions. They demonstrated that social impact stems not from inherent properties of the technology but from the context and creative interpretation when deploying it.

In the right hands, technology elevates creativity (as opposed to automating it). Artistic pioneers showed that machines need not constrain human agency but can expand creative horizons. They exhibited that technology's impact is not fixed but shaped by how we conceptualize and deploy it.

A Visionary Thread

One public exhibition from the 1960s shows the ingenuity and scale of mid-century cybernetic art—a case in future potential.

Now, picture how radical (even ludicrous) this must have been. An up-and-coming art curator appears on a sepia-toned film recording. Jasia Reichardt introduces her new project, Cybernetic Serendipity, which was three years in the making. Talking about it, she rightfully conveys belief and disbelief simultaneously.

It’s a big swing. Her exhibition is the first visible exploration of machine-generated art for the general public. It allows visitors to experience computational aesthetics and production methods. She forces visitors to reconsider who or what a creative agent could be.

According to Reichardt, a wholly new, experiential perspective had crystallized. Computers now interacted with people, drew graphics, rendered portraits, and wrote haikus (remarkable, considering this was the 1960s). Her program would elevate these mechanical wonders into the mainstream.

While preparing for the exhibition, Reichardt regularly consulted a journal called Computers and Automation, the world's first computer magazine. In its pages, several contributors to Cybernetic Serendipity were discovered, including Bela Julesz, Michael Noll, Frieder Nake, and Georg Nees—vanguards of computer-generated graphics and algorithmic art.

They institutionalized new forms and works that blended art and science. The subject (what the machine was) became the source (what the machine could do).

Featured in her introductory remarks, Reichardt discovered, "When one goes to an art exhibition, or an exhibition involving poetry or music, we think that the people who have composed the music or produced the images or written the poetry are artists, poets, or composers. In this exhibition, this is not true."

Cybernetic Serendipity featured works from over 130 contributors, including computer scientists, engineers, artists, composers, and poets. Notable participants included Nam June Paik, John Cage, and researchers from Bell Labs’ Experiments in Art and Technology group (E.A.T).

It immersed visitors in computer-generated graphics, sculpture, music, and poems. People wandered through interactive cybernetic environments, engaged with remote-control robots, and witnessed painting machines creating artwork in real time.

A pixelated, computer-rendered image of Norbert Wiener, the father of cybernetics, greeted them.

Then came Gordon Pask's "Colloquy of Mobiles," a hanging system of rotating fiberglass forms communicated through light and sound. Visitors interacted serendipitously with the mobiles, becoming part of the artwork's feedback loop—human and machine in dialogue.

Interaction between visitors and machines wasn't limited to Pask's installation alone. Various works explored different facets of technological conversation. Edward Ihnatowicz's "SAM" (Sound Activated Mobile) took feedback loops even further. The flower-like sculpture turned toward sounds, tracking movement with an almost lifelike responsiveness that blurred the mechanical and organic lines.

The exhibition provoked a mix of astonishment and bewilderment.

The Observer remarked, "Is everything or anything here art—and if not, why not? We all benefit by asking ourselves this kind of question." The Sunday Times declared that "'Cybernetic Serendipity' provokes, in its implications, it is as different from an everyday' art exhibition' as a major operation from a manicure." The Evening News prophetically announced, "The Industrial Revolution provided the machine age. With the computer comes machine-made art."

The exhibition attracted over 60,000 visitors in just three months—an impressive number for an art exhibition in 1968. The strange new fusion of art and technology raised questions that resonate even more powerfully today:

What is authorship when machines participate in creation?

How do we understand creativity when it emerges from code and circuits?

What new forms arise when machines are the subject and source?

What defines the essence of human imagination when technological creations mirror our capacities?

Rather than marking the end of human expression, Cybernetic Serendipity proposed a profound expansion. New questions emerged, assumptions changed, and along with it, new means to understand art, technology, and creativity.

As we approach an "event horizon," economic competitiveness, technological supremacy, and fundamental natures of human agency are in question.

If we overlook artistic perspectives—with their emphasis on emergence, collaboration, and meaning-making—we risk directing AI systems to optimize us and diminish immeasurable qualities that define our lives.

That’s why we must follow the artists. Marshall McLuhan reminded us that they—not machines or people who make them—are the antennae of the human race.

Leave a comment

Subscribe now

The sledgehammer fell in 1992. I was twenty-two, fresh out of college, and returning to my father's world. His graphic arts studio was a place I'd known growing up but never understood until I worked there with him.

Everything had changed by the time I arrived; we just hadn’t recognized it yet. The device of destruction wasn't physical. It was digital, invisible, and ruthless. The sledgehammer was Photoshop. 

Until then, my Dad was a force in town. To understand him was to understand American business before it became “virtual.” He was a social animal with ambition unbound by caution or vision. His handshakes could hurt, his laughter echoed, and his stories stretched but never broke. He ran his studio as he carried himself. It was a hothouse where people who made things happen—and made things—were gods.

For two decades, they were at the center of the advertising world. In Detroit, art met industry more than anywhere else. His crew knew the essential in between—how to make a car shine on the page in ways it never quite did on the showroom floor.

His designers, drawers, typesetters, and camera techs were craft workers. It was fitting that they worked in a creative factory with a particular feel and scent that reflected the times. The chemicals could strip paint. The smell of paper emanated fresh from the cutter. The constant cigarette smoke hung like clouds around the ceiling lights. When the automation came, the air changed.

The hammer dropped, erasing everything we knew, including the feel. It happened one pixel at a time.

Gradually, Then Suddenly

As the saying goes, Photoshop’s impact hit gradually, then suddenly. Steve Jobs introduced the Macintosh in 1984 with his famous Super Bowl commercial. It featured an actual sledgehammer thrown at conformity. Jobs didn't say then that his bicycle for the mind would become a wrecking ball for certain kinds of creative work—my Dad's era of creative work.

The Mac led to new software, most notably Photoshop. The effect wasn't immediate but gained momentum by the early 1990s. Jobs' beige computer boxes started showing up on more desktops in the workplace. Once new software was loaded into them, the creative rhythms changed. The click-click-click of mouse buttons began to drown out the scratch of pencils and the squeak of Pentel marker tips.

My Dad and his band (the present company included) didn't adapt fast or fundamentally enough. Revenue and margins shrank as clients took the work in-house. We doubled down on what we knew, only accelerating our demise. A new technology, and those who knew how to use it, dismantled the business in about 24 months.

We should have seen the hammer coming because it was already there. The lesson: Not reacting to a technological wave until it breaks over you is more than just a business failure—it can be an enduring, personal failure. I promised myself never to be caught so ill-prepared again.

Magical Automated Workflows

Photoshop turned 35 last week. If you were to imagine a single icon representing a creative transformation, it would be the "Magic Wand." One-click. That's all it took. A click and similar colors were selected as if by sorcery. Same with a click to alter the composition of an image. Ditto for envisioning variations of a scene.

What had once required careful technique and training was now available to anyone with access to the software and the patience to play and learn with it.

The magic wasn't just in what the wand could do but in what it represented. The transition from physical to digital craft offered efficiency.

Ironically, Photoshop—and the efficient workflows it led to—also came from family. Here is a bit of backstory on how it came to be.

Thomas and John Knoll grew up in a house that valued art and technology. They stood at the crossroads of two worlds, uniquely positioned to bridge them.

In 1987, Thomas Knoll wanted to display grayscale images on his Mac's black-and-white monitor. It was a practical problem with what seemed like a limited solution. His brother John, working at Industrial Light & Magic, saw further. He convinced Thomas to expand the program to handle color on the new Macintosh II. What began as a personal project caught the attention of the industry's power players.

A capability previously reserved for mainframes could now run on a PC. Adobe recognized the potential immediately, securing distribution rights in 1988 and releasing Photoshop 1.0 for Macintosh in 1990.

Photoshop didn't transform creative work in isolation. PageMaker, released by Aldus in 1985, opened the door to what would become known as desktop publishing. Photoshop and PageMaker encoded creative techniques and integrated workflows into software that creatives and producers could use directly. Those in the studio world who didn’t adapt to augmentation and changing workflows were permanently displaced.

The Magic Fades Without Imagination

Decades later, a much bigger automation wave is building. Intelligent software is reshaping all creative and knowledge work, echoing what happened in our studio.

Some automation parallels are striking. Both represent shifts from manual to digital creation and spark similar existential questions about the value of work and the identities of those who do it. With desktop publishing, page designers and typesetters rightly feared obsolescence. Today, anyone who produces knowledge, research, or creative work naturally expresses the same concerns about generative AI.

There are also critical differences between automation then and now.

Desktop publishing tools were extensions of human work. They replicated technical aspects but required direct human guidance for every decision. Generative AI tools can generate work with minimal direction, shifting human expertise from execution to curation.

Desktop tools required understanding design principles, but generative AI can produce seemingly decent output without the user knowing the underlying fundamentals.

Perhaps most significantly, the desktop publishing revolution unfolded over a decade while generative AI's capabilities are evolving at an incredible speed.

The importance of judgment, discernment, and taste in delivering commercial-grade work remains unchanged, whether we’re talking about Photoshop thirty years ago or OpenAI’s latest reasoning model today.

Consider Photoshop's meaning as a metaphor for the current moment. It automated specific, repeatable, known tasks and made technical processes faster and more accessible. However, it could not replicate the mystery of creativity or imagination, which no software has yet managed to do.

And therein lies a twist. Looking back, what reads like a family business failure doesn’t tell the whole story. 

After experiencing the destruction of our business, my path reflects possibilities as a new technology destroys and creates simultaneously.

Alongside highly inventive colleagues and clients, I’ve helped create and grow new businesses built on mobile computing, e-commerce, weblogs, social networks, digital content, app development, community management, digital video, and social intelligence.

We capitalized on these tech breakthroughs not merely by understanding their specifications or original use cases but by seeing what they could lead to—by tapping into our creative capacity to imagine and bring new possibilities to life. AI is the next frontier on which to build.

Yes, AIs can encode what has been and suggest probabilities for the future. They can analyze patterns from the past with astonishing accuracy. But they cannot predict how we'll ride the next wave.

The hammer will fall, but what emerges isn’t simply destruction or dead ends. There can be a lot of light in and at the end of a transformation tunnel.

Neither my Dad nor I fully understood it in a moment of failure. It’s a reminder—then and now—that it’s hard to read the label of the jar you’re in. You have to see it from the outside.

Subscribe now

The moment I realized something radically changed, I started speaking Portuguese. Last November, in the shadows of a stage in São Paulo, I watched a self-created digital twin deliver part of a keynote presentation to technology leaders.

Although I don't speak the language, the synthetic me on screen did quite well. It engaged the audience more deeply than my English delivery possibly could.

It was a true-to-life, a-ha moment. This simulation took about four minutes to create. An avatar engine called HeyGen learned the contours of my face, voice inflections, and mannerisms. It then translated my talking points and delivered a presentation-ready, 4K version of a second self. The experience of making and then seeing it was unsettling.

For many, pervasive discussions around AI will become real through mind-blowingly good simulations like this. Despite years of research and countless hours in tech labs, I sensed following the demo that this technological force might actually be underhyped.

The unrelenting surge in AI suggests a technological supernova is here. With over 3.7 billion monthly visits, ChatGPT traffic has surpassed LinkedIn, Twitch, and Netflix, equaling Amazon’s traffic in just twenty-four months.

Functionality and potential use cases are exploding. Google's AI suite includes personal AI notebooks, deep research assistance, text-to-video production, and more personalized search. Anthropic’s Claude updates include taking over your PC to execute programming tasks or providing assistance like trip planning. All the major players, notably Apple and Microsoft, now embed genAI into their operating systems. To realize the long-talked-about potential this brings, enterprises are predicted to spend roughly a quarter of a trillion dollars on AI this year.

How leaders deploy intelligence will change rapidly. Through agentic AIs, NVIDIA's CEO Jensen Huang told a crowd at CES that “digital workers” will arrive in 2025, representing a new field force that can understand tasks, plan, and take action autonomously.

We need a deeper understanding of what's to come, both psychologically and practically, to avoid being shocked by the scale and pace of displacement.

Adapting to a "Generative World"

With all this power, AI will remold how we comprehend the world around us. Work will move from routine jobs to roles centered on emotional intelligence and complex problem-solving. Education will evolve from standardized knowledge delivery to personalized learning and creative synthesis. Social relationships will become even more mediated through AI-enhanced connections.

This isn't just a theoretical vision of the future; it's becoming present reality. AI systems now generate everything from legal contracts to drug molecules, while algorithms shape public discourse as powerfully as human voices. In boardrooms and research labs across the globe, these technologies are fundamentally reshaping how we work, create, and solve problems.

Consider several vignettes about what's happening:

In London, narrative intelligence engines map invisible threads between brands, beliefs, and behavior. They don't just analyze content—they decode culture, revealing how stories flow between influencers, politicians, and media networks. Pattern recognition at a scale makes traditional market research look like stone tablets.

In Singapore, a strategy team runs hundreds of simultaneous scenario simulations. They're not just predicting policy reactions—they're experiencing possible futures in real-time, watching how ideas ripple through simulated societies. The gap between planning and experience has collapsed.

In San Francisco, developers watch AI-generated videos of perfectly sliced steak, born not from a culinary class but from a machine. They aren't marveling at the quality of the output; they're benchmarking the performance of their new model against others. The uncanny valley of technological simulations isn't just shrinking—it's closing entirely.

Across the Bay, an AI platform CEO reimagines advertising for a world where agents, not humans, make product suggestions. These micro-AIs will filter information, evaluate options, and optimize choices for us. Madison Avenue's next audience won't be human.

In New York, a startup demonstrates human-like avatars that blend the visual presence of world-class experts with exceptional conversational ability. Imagine having instant access to global authorities on any topic, their knowledge perfectly preserved and perpetually available. The democratization of expertise has begun.

Distribution of Understanding is an Opportunity and a Threat

An often repeated quote about change comes from science fiction writer William Gibson, who said, "The future is here. It's not evenly distributed." Unevenly distributed, in this case, applies to government agencies, companies, and individuals working—or not—with AIs.

The distribution, both the growing expertise and ignorance, is widening. It's a real and risky side effect of AI's lightning-fast ascendance for those outside looking in.

You don't need to deploy AI to be transformed by it. Competitors or new companies integrating AI into their operations can quickly displace yours. Proprietary IP can become training data without permission. A CEO's voice can be cloned in minutes. Carefully crafted messages can be drowned in a sea of synthetic content.

If you watch closely, you can see a strange new world being born.

The transmission of knowledge now operates in two dimensions: one designed for human perception and another engineered for the algorithmic systems that increasingly shape our understanding.

The nature of human connection is being rewritten. In my São Paulo moment, I witnessed something profound: it can transcend native language, culture, or even reality. At times, connection flows more freely through artificial channels, digital twins, and synthetic media that bridge the gaps our biological selves cannot.

With this, reality has become fluid, negotiable, and a medium to be shaped rather than a foundation to build on. The old certainties—what's real, authentic, and genuine—haven't disappeared, but they've evolved into something more complex.

There's No Off-Ramp

The moment presents business leaders with extraordinary opportunities and challenges. According to OpenAI, 92% of enterprises are engaged with AI, yet only 26% have developed capabilities to generate tangible value. The gap suggests adoption without purpose, movement without meaning—the corporate equivalent of running full speed into the dark.

The core challenge for leaders isn't technological—it's philosophical. How do you maintain authentic leadership when AI can simulate authenticity? How do you develop your team's ability to use AI effectively while preserving human judgment? And how do you build lasting trust in an environment where the line between real and artificial continues to blur?

The answers aren't simple, but they begin with experiences like mine in São Paulo—moments when we feel how profoundly the ground is shifting beneath our feet.

We can traverse uncertainties and adapt to a generative world by:

First, using simulation for sensemaking. It becomes a primary way to think, plan, and create. Leaders no longer need to ponder forecasted scenarios—they live inside them. They feel the ripple effects of decisions not yet made, experience the emotional resonance of messages not yet sent, and understand the impact of strategies not yet deployed.

Second, understanding public discourse to shape strategy. Understanding the web of beliefs, values, and cultural currents that shape how people find meaning is non-negotiable. AI doesn't just map conversations; it reveals the underlying patterns of human behavior across cultures, demographics, and worldviews that can make or break strategic actions.

Third, embracing human values—and imperfections—in the work. As AI systems become more sophisticated and ubiquitous, the distinctive qualities of human experience take on new significance. The slight hesitation in a CEO's voice when addressing a crisis, the unscripted moment of vulnerability in a press conference, the genuine stumble in a product launch—these moments of authentic human presence become more influential because they can't be engineered.

Reflecting on my digital twin's near-flawless delivery, I was reminded that simulated perfection doesn't guarantee a genuine connection.

We see it in the public mood. People can distinguish between acts that are manufactured to be perfect and those that are perfectly imperfect. They've developed a new literacy, an ability to distinguish between engineered empathy (B.S.) and nuanced engagement (the real stuff).

Ultimately, the most valuable thing isn't simulated perfection—it's presence, judgment, and understanding that can only come from being human.

A generative future isn't approaching; it’s already randomly distributed. Like a celestial supernova, AI is simultaneously collapsing and expanding, destroying and creating, ending one era while seeding the elements of the next.

This audio post of Perspective Agents explores the breakdown of institutional media control and how podcasts, prediction markets, and networked sources reshape news and political power.

To help make sense of this break, I’m joined by John Wihbey, professor of media innovation at Northeastern University and head of the Northeastern Internet Democracy Initiative. As the author of The Social Fact, and with a new book on tech policy and algorithmic governance coming soon, John brings a depth of perspective on this pivotal moment.

Recorded just after the election, several themes we touch on include:

• Americans’ deep dissatisfaction with corporate news sources.

• Data and peer corroboration challenging expert credibility.

• New information exchanges—podcasts, Perplexity, Polymarket—gaining trust.

• A "post-news" world where traditional journalism is just one voice among many.

We also explore how technology’s influence likely shaped outcomes in ways we can't yet see. This raises questions about how politicians adapt to AI, persistent diversity gaps in tech development and adoption, and the growing need for regulation that keeps pace with algorithmic advances.

Links to items discussed include:

Research

• 2024 US Election: Sources of News and Information (Chip50)

• The National Internet Observatory

Sources

• Polymarket

• Polymarket Electoral Map

• Perplexity Election Information Hub

• Lawfare Podcast

• Tech Policy Press Podcast

Studies & Books

• Donald Hoffman. Interface Theory of Perception and The Case Against Reality

• Russell Hardin. How Do You Know? (Economics of Ordinary Knowledge)

• Daniel Bell. The Coming of Post-Industrial Society

Tech Policy & Regulation

• Section 230

• Digital Services Act

• China Digital Sovereignty Model

In future episodes, I’ll invite leading thinkers to help decode social and business change shaped by digital media, technology, and culture. If you like it, please give it a plug. Thanks for listening.

Share

TL;DR: AI won't build the future. We will. But we need more than powerful algorithms or language models to do it right. We need leaders and architects with the know-how to build with AI. To do so, we need more investment in people, not just tech.

Subscribe now

The future is not something that happens to us; it's something we build. When constructing with AI, we need to see it not as an endpoint but as a powerful set of tools in the hands of human builders.

As tech limitations become clear, expectations of fully automating complex tasks—from strategic planning and creative content generation to intricate data analysis—are unrealistic without humans in the loop. With that, the euphoric narrative is hitting a wall.

Skeptical voices are getting louder. Gartner predicts that by 2025, 30% of AI projects will be abandoned. Goldman Sachs adds to the pessimism, noting that generative AI tools are cumbersome and “not yet smart enough to make employees smarter.” Media reports paint OpenAI as an internal mess (despite a $6.6 billion funding round announced yesterday). Stock analysts recently downgraded Microsoft. The list could go on, but the meta-story is now tainted with caution.

Critics may be overlooking the bigger picture. While the hype is exaggerated, so is growing pessimism. The truth lies in the middle. We must balance AI’s growing power with people’s ability to produce with it.

It’s Time to Build. But How?

Manifestos from VCs with heavy bets in AI proclaim that it's Time to Build and evangelize a brand of Techno-Optimism. But what does that look like in practice? How do we build the future they envision?

Start by likening it to constructing a building. A skyscraper's beauty, strength, and function rely on solid architectural design. And it’s not just made of flashy glass and steel—it needs a solid foundation and essential elements like bricks and mortar to form durable, enduring structures.

Large language models are like building blocks. Microsoft Co-pilot is an example of how these models are now embedded in Word, Excel, and PowerPoint. Across Fortune 500 companies, generative AI now supports everyday tasks such as writing, editing, and planning. But general-purpose assistance isn’t enough. Other components—task-specific agents, small language models, and APIs—combine to make AI something to build with.

These AIs are only as useful as the “mortar,” the human bond that binds them into something meaningful. What gets built requires human intuition, practical experience, and commercial-grade know-how. Without human intelligence, AI is just a stack of potential with no direction, like a heap of bricks waiting for a blueprint.

Put another way, just as architects consider how a building fits its environment, leaders must consider how AI fits into their growth strategy and business functions.

Planning to do so gets complicated if you’re swayed by bubble-bursting speculation. Cynics suggest that AI is overhyped, overvalued, and overcapitalized. But zoom out and look at AI power curves. This perspective seems misguided. The actual value derived from these systems—the real story—is that people are underprepared, undervalued, and undercapitalized to use them effectively. If the AI bubble pops, don't blame the tech. Look to the human element.

Seeing Beyond the Hype

Writing for Harvard Business Review, Eric Siegel cautions against getting distracted by narratives and hype cycles. Blindly buying into AI without a clear plan is fraught, but so is sitting back and waiting for the hype to subside. Unlike the deployment of the Internet or social media, which took years to take hold, the rapid evolution of AI models leaves little room for complacency. Delays lead to being left behind as competitors discover new avenues for growth, efficiency, and accelerated decision-making.

Remember, we're in the early stages of AI deployment. AI in its various forms is still in its "DSL phase." Like dial-up modems, which gave way to broadband and laid the foundation for the Internet, bellwethers like NVIDIA, OpenAI, Microsoft, and Meta have built essential AI infrastructure for us to build on today.

Like how Cisco and Oracle shaped the early web, AI developers will create possibilities we can’t fully imagine. Think personalized education that adapts to each student, creative design tools that reshape entire industries, or simulated scenario planning that informs big decisions.

To reach potential destinations like these, it helps to investigate what may impede progress. Gartner's Hype Cycle suggests we've passed the peak of inflated expectations and are heading into disillusionment. If this is true, we must confront the human barriers that stand in its way.

Blocks to AI Adoption

Let's explore critical, human-centered considerations that might flatten the curve. Beyond the research cited by consultants, I’ve experienced each of these firsthand working with clients and teams.

1. Leadership Intent: While 94% of business leaders agree that AI is critical to success in the next five years, only 20% report that their organizations are "AI-ready,” according to a survey by Deloitte. This gap between acknowledgment and action highlights the need for a more substantial executive commitment.

2. Structure: McKinsey revealed that organizations successfully scaling AI are 3.5 times more likely to have C-suite executives working together on a transformation agenda. A collaborative approach driven by executive leaders is necessary to break down departmental barriers that lead to silo-busting pilots.

3. Procurement: According to EY, the number of companies investing $10 million or more in AI is set to nearly double next year to 30%, up from 16% currently investing at that level. However, despite the forecasted investment, the survey found that leaders ignore department-level needs that AI investments require to pay off.

4. Silver Bullet Mentality: Research published in MIT Sloan Management Review indicates that companies benefiting from AI are likelier to use multiple, specialized AI applications than a single, general-purpose system. This underscores the importance of adding task-specific AI tools over one-size-fits-all platforms.

5. Talent Networks: PwC’s Global Jobs Barometer found postings for AI jobs are growing 3.5x faster than all jobs, with AI skills commanding a 25% wage premium. This suggests a growing need for specialized talent, new roles, and partnerships to embed deep AI expertise in the business.

6. Staff Development: The World Economic Forum's Future of Jobs Report predicted that by 2025, while 85 million jobs may be displaced by automation, 97 million new roles are projected to emerge, “reflecting a shift in the division of labor between humans, machines, and algorithms.” This highlights the need for ongoing, collaborative development programs to integrate AI tools into workflows.

7. Employee Caution. According to a Thompson Reuters Institute study, 53 percent of professional services employees said they were hesitant, concerned, or fearful of using generative AI. Of those hesitant, 20% said they felt that way due to accuracy concerns, 18% said they were skeptical GenAI could deliver promised results, and 16% were concerned about over-reliance. 45% said they had no plans to use it.

8. Controlled Pilots: According to a report by BCG, companies that run multiple AI pilots and scale the successful ones are 3.5 times more likely to see financial benefits from AI. A pilot strategy allows organizations to test AI applications in controlled environments and identify use cases that inform full-scale implementation.

9. Strategic Evolution: MIT Sloan Management Review also found that companies with flexible, iterative approaches to AI implementation were 1.5 times more likely to report value from AI than those with rigid, top-down plans. This emphasizes the importance of collective learning from all parts of the organization to inform an agile AI strategy.

10. ROI: According to Gartner, 49% of survey participants identified the ability to accurately prove the value of AI projects as the most significant barrier to AI adoption. This highlights the need for alternative ways to assess the impact of AI investments, like return on the insight that makes businesses perform better.

Human-Centered Investment

Yuval Noah Harari, the renowned intellectual and AI critic, suggests that for every dollar and minute we invest in building AI, we should put an equal amount into developing ourselves. BCG takes the human thesis even further by taking a 10-20-70 approach that emphasizes algorithms (10%), tech and data (20%), and people and processes (70%).

A rebalanced approach is necessary to develop indispensable tools and successfully integrate them into the workplace. AI requires architects with business acumen to deploy them for practical use. Most importantly, AI needs human ingenuity and creativity to reduce workplace burnout, automate grunt work, and create the capacity to solve complex problems.

The future isn’t about AI replacing us; it’s about AI augmenting what we can build and make work better. As former IBM CEO Ginni Rometty said: “Some call this artificial intelligence, but the reality is this technology will enhance us. It will augment our intelligence.”

All this is a lot to take in. Still, if we’re overwhelmed by the present, we risk missing out on activating vast collective intelligence just a few prompts away. Creating space for organizational development is essential to building the future with AI. Without a practiced understanding of what it can do, you may find yourself sitting on the sidelines, blinded by the hype.

Subscribe now

Last year, I was at a dinner in New York where Steven Johnson talked up his new project with VCs and product people. Given his resume, you'd expect Johnson was there to introduce an innovation story destined for the bestseller list. He's written thirteen books to high acclaim, including Where Good Ideas Come From, How We Got to Now, and a favorite, Everything Bad Is Good for You. An intellectual alchemist, he draws insights from science, technology, and history, revealing how they converge to drive progress.

This time, Johnson wasn't talking about his next page-turner. He'd been cooking up NotebookLM, an AI research assistant in partnership with Google. Unlike products such as ChatGPT, which train on a massive dataset, this one processes only the personal data you share with it.

After a limited release in July 2023, NotebookLM became widely available for free last December. Its new true-to-life voice feature now has people raving about it.

How does it work?

Before we get to the buzz about its voice, a bit more about the platform. As Johnson describes, NotebookLM is distinct from other models in this way:

"The core idea behind the current version of the product is what we call 'source-grounded AI': you define a set of documents that are important to your work—called 'sources' in the NotebookLM parlance—and from that point on, you can have an open-ended conversation with the language model where its answers will be 'grounded' in the information you've selected."

Sources are files you feed it—like PDFs, text docs, and web pages—that make the model an expert on the material. Once digested, it returns insights from the content, including citations and quotes.

NotebookLM has a whole bag of tricks: FAQs, study guides, tables of contents, timelines, and briefing docs. It also gives you a few suggested questions to get the juices flowing. Google swears your data stays private, so for now, you don’t need to worry about the model training on it.

Say you're prepping for a big presentation. You've got piles of business reports, competitor analysis, and cultural intel. So, instead of spending days or weeks sifting through it all, you feed it to Notebook LM. In the time it takes to pour a cup of coffee, the AI creates documents with key takeaways and talking points tailored to what you’re looking for. The starter docs are a portal into chat, where you prompt for more specific information needed for your pitch.

I've been using it for months, and it's remarkable. It's turned years of research—my books, highlights, quotes, and notes—into a living, personalized knowledge base. I uploaded a PDF of my book to demonstrate what it can do. To see what people are buzzing about, check out the upper right corner of the window.

The audio overview is where it gets pretty weird. It's a podcast with two male and female voices riffing on the book's content. They're downright enthusiastic about Perspective Agents (which is nice). They hit the main points, bounce ideas off each other, and even crack a few jokes about how crazy media has become. In conversation, they interrupt, finish each other's sentences, and laugh at the absurdity of it all.

And the speed? Mind-boggling. Two-hundred-sixty pages uploaded in nineteen seconds flat. An executive summary and document options popped up like magic. Three minutes and twenty-nine seconds later, boom, a full-fledged podcast. Out comes eight minutes of back-and-forth banter. You can give the cut a thumbs up or down. The first take was a bit rough, so I hit the 'redo' button a couple of times. The duo eventually did an astonishing job of describing the book. Give it a listen here.

Now, any document—a book, a slide show, even a resume—can become conversation fodder for a podcast. By now, you know the twist. These aren't humans; you have AIs talking as if they were. They're fairly convincing replicas of real people.

An audio sludge pile?

Steven Levy, a veteran technology journalist and author of Hackers and In the Plex, has covered digital immersion into our lives for decades. In a Wired review last December, he cautioned that NotebookLM might be too good. He said about the risk:

"Users of NotebookLM, who simply want to get a good job done quickly, might not take the time to do that hard work of thinking. They might not even bother to pore through the research materials themselves. Why take the time when your AI buddy has gone through the material much more closely than you and has already reached some nifty conclusions about it?"

Now, add AI podcasts to how we might think or depict thought. Synthetic dialogue is a double-edged sword. On one side, it can democratize content creation, offer education tailored to every learning style, and unravel complex ideas with lightning speed. On the other hand, it can create a tidal wave of audio sludge, a vehicle for disinformation, and a medium where human voices get drowned out in a sea of fabricated chatter.

For now, early adopters are electrified. It has a feel of ChatGPT’s early days, the sense that something big is going on here.

As a podcasting side hustle, NotebookLM has potential beyond the research assistance Johnson and Google designed it for. Imagine AI pundits battling it out over the latest political issue, theoretical physicists explaining the mysteries of the universe like they're chatting over a beer, or your half-baked ideas transformed into a coherent audio drama.

While this may not be achievable today, the platform will likely evolve rapidly. Soon, it could allow for completely tailored productions, including selecting the voice, tone of conversation, gender, number of participants, and edits. (For reference, note OpenAI’s voice updates this week.)

What’s available now from the Johnson/Google collaboration is excellent for its intended use. In the case of my book, Perspective Agents in NotebookLM is a far better knowledge partner than my OpenAI agent that trained on it. The audio aid alone prompts authors to rethink how their work will be consumed and reviewed. All in all, it feels like a breakthrough or breakdown, depending on your view.

For kicks, I created a notebook trained on this 1,102-word essay. You can listen to the six-minute podcast made from it here. The result made me wonder if my time was well spent writing this and yours reading it.

You might grasp the story better from two AIs talking about it than from the words I put on the page.

I'm happy to share a link if you want access to the Perspective Agents on NotebookLM. Hit me up direct, and I'll send it along.