Got another one from a reader today. Check out https://www.perplexity.ai/ -- an "any question" knowledge engine with sources cited. Others worth adding to the list??
I think we're very quickly going to some some degree of commodification (integration into Adobe, Figma, etc.), but also a quick erection of barriers in commercial settings/uses as questions of Intellectual Property Rights, Ethics and Veracity overcome the initial wave of (much deserved) excitement.
If that holds true, I think we'll see growing interest in products and services that specifically pitch themselves as addressing these issues (i.e. Neeva for generative search and source citation, or Anthropic as a more ethical general purpose model).
This will also open the door for brands to go beyond experimenting with out of the box models and start fine tuning LLMs and Diffusion models based on their own brand voice and IP. Greater control over the inputs and outputs mitigate known risks that I expect will put a damper on just how widely these tools are used by brands to produce public-facing content. That in turn could lead to a broader impact as brands feel safer using the tools to assist in the creation of public-facing content.
I imagine innovation from AI will occur in two cycles: the first is the application of AI to the known (e.g. search engines, copy writing, etc.), and the second being wholly new applications we haven't necessarily thought of yet. Over the next 6-18 months I imagine we'll see AI applies to tons of low-hanging-fruit, but I'm particularly excited about what comes next.
What might some of those new applications be? Well, GPT and other LLMs are particularly good at "translating" our language inputs into requirements and outputs, but they're not necessarily good at logic. CNET discovered ChatGPT can write plenty of articles across a variety of verticals, but wasn't particularly effective in finance because it's bad at math.
But, what happens when LLMs are combined with knowledge engines like Google's Knowledge Graph or Wolfram Alpha? Instead of an engine like ChatGPT being there to answer relatively simple queries or draft rote content, could it become a math tutor to a child, an engineering assistant to an architect, or a thought partner to an entrepreneur?
I'm excited by the ability for these tools to accelerate content creation and even free one's own creative mind from the shackles of under-developed artistic craft (speaking for myself), but I'm eagerly awaiting the inevitable evolutions in AI tech which can augment my thinking and not just my execution.
Type in a city and the # of days and within seconds you have an itinerary built. Imagine how much better this will get once it incorporates review data (Yelp, Google, etc.) and incorporates personal taste based on previous trips.
Got another one from a reader today. Check out https://www.perplexity.ai/ -- an "any question" knowledge engine with sources cited. Others worth adding to the list??
I think we're very quickly going to some some degree of commodification (integration into Adobe, Figma, etc.), but also a quick erection of barriers in commercial settings/uses as questions of Intellectual Property Rights, Ethics and Veracity overcome the initial wave of (much deserved) excitement.
If that holds true, I think we'll see growing interest in products and services that specifically pitch themselves as addressing these issues (i.e. Neeva for generative search and source citation, or Anthropic as a more ethical general purpose model).
This will also open the door for brands to go beyond experimenting with out of the box models and start fine tuning LLMs and Diffusion models based on their own brand voice and IP. Greater control over the inputs and outputs mitigate known risks that I expect will put a damper on just how widely these tools are used by brands to produce public-facing content. That in turn could lead to a broader impact as brands feel safer using the tools to assist in the creation of public-facing content.
I imagine innovation from AI will occur in two cycles: the first is the application of AI to the known (e.g. search engines, copy writing, etc.), and the second being wholly new applications we haven't necessarily thought of yet. Over the next 6-18 months I imagine we'll see AI applies to tons of low-hanging-fruit, but I'm particularly excited about what comes next.
What might some of those new applications be? Well, GPT and other LLMs are particularly good at "translating" our language inputs into requirements and outputs, but they're not necessarily good at logic. CNET discovered ChatGPT can write plenty of articles across a variety of verticals, but wasn't particularly effective in finance because it's bad at math.
But, what happens when LLMs are combined with knowledge engines like Google's Knowledge Graph or Wolfram Alpha? Instead of an engine like ChatGPT being there to answer relatively simple queries or draft rote content, could it become a math tutor to a child, an engineering assistant to an architect, or a thought partner to an entrepreneur?
I'm excited by the ability for these tools to accelerate content creation and even free one's own creative mind from the shackles of under-developed artistic craft (speaking for myself), but I'm eagerly awaiting the inevitable evolutions in AI tech which can augment my thinking and not just my execution.
Another perspective agent in the travel domain: https://gpt-travel-advisor.vercel.app/
Type in a city and the # of days and within seconds you have an itinerary built. Imagine how much better this will get once it incorporates review data (Yelp, Google, etc.) and incorporates personal taste based on previous trips.