We’re headed for huge layoffs and a democratized market that sucks the profit out of most work. While individual forward-thinking companies may be able to move past other companies stuck in their ways, the overall market will not necessarily grow much in the sense of selling more clothing or more cars or more productivity software. There will be new markets created, for example on the consumer side, tutoring or AI friends etc, but I have a feeling they will be winner-take-all composed of a few oligopoly players or just one dominant player in each segment (if not rolled up across segments by google/openAI/Microsoft).
More generally, when something becomes easy, it doesn’t become more lucrative, it becomes less lucrative. Back when building websites was hard, or sourcing from China was hard, you could build a profitable business with a big moat. Now with Shopify or Alibaba or whatever, you have hundreds of thousands of entrants competing on who can undercut the most. It’s like an economics 101 class with perfect competition competing away all the producer surplus. AI will make white collar work of all kinds “easy”, and we will see the same result. The model providers (google/OpenAI/Microsoft/who-knows-mistral) will make all the money while millions of model users compete against each other
Great news about massive context windows and decreased latency.
Yet you end on a critical note. Not to wear out Wayne Gretsky's advice - but now is the time for us to skate to where the puck will be, not to where it is.
The suggestions that you provide for leaders are reasonable, but I am skeptical that a CxO of a generic trad corp will have the werewithal or desire to do what you suggest. I predict that most trad corps will use AI to do seemingly simple, but actually complex, stuff like chatbots, etc. And I expect those initiatives to fail spectacularly. It will take trad corps a while yet to figure out how to integrate AI into their workflows. This presents an interesting problem for a company like OpenAI, which is trying to get trad corps to sign enterprise contracts with it. Less so for Google/MSFT/etc, which have the cash to ride out the learning cycle.
Agree with your point about building for the future. Using AI tools today feels like doing web design in 1998. Pretty basic but exciting.
Learning prompt engineering is like learning html back then. But the web today doesn’t look like it did in the late 90s. Nor we need to know how to code to build a website. Yet, many people fail to consider this when it comes to AI. They dismiss AI because of its present state.
Writers everywhere are reluctant to leverage AI, thinking they will never be replaced, just because the quality of AI writing is not there yet. But if the history of technology teaches us something is that people have a hard time predicting how tech will affect them. Innovation and development occur incrementally, moving from one feasible possibility to the next.
Rather than taking giant leaps, progress often happens when you explore what's immediately attainable. And then the new possibilities that emerge from that step become the next adjacent possible.
Could you further articulate how AI itself - not merely its outputs - restructures not just individual tasks but the whole environment of an organization? thinking more along the lines of "The medium is the message."
also, why are we letting silicon valley decide that processing information quickly is needed in our society?
It’s not the massive context windows that are the game changer, but the improved recall in those large context windows that is. Claude has had 200K context windows for a while now but, without injecting an additional phrase in your prompt, recall/accuracy from the middle of the window was poor. What makes Gemini 1.5 Pro impressive is that its 1M context window was actually useful to you. I can only imagine what Gemini 1.5 Ultra will be like
question on the statement: "AI can provide great user journeys and personas, so your old product management approach is no longer a differentiator. "
Examples? As with any AI work, all is dependent on the data available and for most companies mapping user journeys is incomplete and source info on personas relies on interviews. thanks,
This is a very exciting statement for anyone getting into AI right now:
"One of the things I emphasized to them is that no one really knows what current LLMs are capable of, since most of the tests conducted by AI companies focus mostly on coding and testing benchmarks, not real-world applications. Since my students come from many industries and countries, they had a tremendous diversity of potential use cases. By the time they started building GPTs for their specific needs, whether private equity deal memos or suggesting a perfect wedding ring to their fiancé, they became the world experts in using AI for their specific field."
I have also been wondering if organizations should be working to identify areas where AI does not help at all and in fact, may actually hurt. They could use AI to get rid of mundane tasks while investing more time in these non-AI areas.
As the pace of acceleration continues the only position generating movement is internal clarity of purpose whereby one finds the fuel to do things without regard to their externalities. Trying to plan things based on externalized value misses the fact that our future conceptions of and relations with technology and one another are becoming inconceivable as we move towards a paradigm shift in how we create and conceive of value.
We’re headed for huge layoffs and a democratized market that sucks the profit out of most work. While individual forward-thinking companies may be able to move past other companies stuck in their ways, the overall market will not necessarily grow much in the sense of selling more clothing or more cars or more productivity software. There will be new markets created, for example on the consumer side, tutoring or AI friends etc, but I have a feeling they will be winner-take-all composed of a few oligopoly players or just one dominant player in each segment (if not rolled up across segments by google/openAI/Microsoft).
More generally, when something becomes easy, it doesn’t become more lucrative, it becomes less lucrative. Back when building websites was hard, or sourcing from China was hard, you could build a profitable business with a big moat. Now with Shopify or Alibaba or whatever, you have hundreds of thousands of entrants competing on who can undercut the most. It’s like an economics 101 class with perfect competition competing away all the producer surplus. AI will make white collar work of all kinds “easy”, and we will see the same result. The model providers (google/OpenAI/Microsoft/who-knows-mistral) will make all the money while millions of model users compete against each other
Great news about massive context windows and decreased latency.
Yet you end on a critical note. Not to wear out Wayne Gretsky's advice - but now is the time for us to skate to where the puck will be, not to where it is.
The suggestions that you provide for leaders are reasonable, but I am skeptical that a CxO of a generic trad corp will have the werewithal or desire to do what you suggest. I predict that most trad corps will use AI to do seemingly simple, but actually complex, stuff like chatbots, etc. And I expect those initiatives to fail spectacularly. It will take trad corps a while yet to figure out how to integrate AI into their workflows. This presents an interesting problem for a company like OpenAI, which is trying to get trad corps to sign enterprise contracts with it. Less so for Google/MSFT/etc, which have the cash to ride out the learning cycle.
Agree with your point about building for the future. Using AI tools today feels like doing web design in 1998. Pretty basic but exciting.
Learning prompt engineering is like learning html back then. But the web today doesn’t look like it did in the late 90s. Nor we need to know how to code to build a website. Yet, many people fail to consider this when it comes to AI. They dismiss AI because of its present state.
Writers everywhere are reluctant to leverage AI, thinking they will never be replaced, just because the quality of AI writing is not there yet. But if the history of technology teaches us something is that people have a hard time predicting how tech will affect them. Innovation and development occur incrementally, moving from one feasible possibility to the next.
Rather than taking giant leaps, progress often happens when you explore what's immediately attainable. And then the new possibilities that emerge from that step become the next adjacent possible.
Could you further articulate how AI itself - not merely its outputs - restructures not just individual tasks but the whole environment of an organization? thinking more along the lines of "The medium is the message."
also, why are we letting silicon valley decide that processing information quickly is needed in our society?
This was an informative and enjoyable post.
Thank you for sharing your thoughts and insights.
It’s not the massive context windows that are the game changer, but the improved recall in those large context windows that is. Claude has had 200K context windows for a while now but, without injecting an additional phrase in your prompt, recall/accuracy from the middle of the window was poor. What makes Gemini 1.5 Pro impressive is that its 1M context window was actually useful to you. I can only imagine what Gemini 1.5 Ultra will be like
Great questions and loads to think about.
question on the statement: "AI can provide great user journeys and personas, so your old product management approach is no longer a differentiator. "
Examples? As with any AI work, all is dependent on the data available and for most companies mapping user journeys is incomplete and source info on personas relies on interviews. thanks,
Is it bundled with the audio version in the US launch but not in the UK? Thnx
Thanks Ethan really helpful and particularly like the key questions for leaders.
This is a very exciting statement for anyone getting into AI right now:
"One of the things I emphasized to them is that no one really knows what current LLMs are capable of, since most of the tests conducted by AI companies focus mostly on coding and testing benchmarks, not real-world applications. Since my students come from many industries and countries, they had a tremendous diversity of potential use cases. By the time they started building GPTs for their specific needs, whether private equity deal memos or suggesting a perfect wedding ring to their fiancé, they became the world experts in using AI for their specific field."
I have also been wondering if organizations should be working to identify areas where AI does not help at all and in fact, may actually hurt. They could use AI to get rid of mundane tasks while investing more time in these non-AI areas.
I love these posts; I have converted this article to a high-quality AI narration for anyone interested:
https://askwhocastsai.substack.com/p/strategies-for-an-accelerating-future?sd=pf
Thanks for the post. Looking forward to your book!
As the pace of acceleration continues the only position generating movement is internal clarity of purpose whereby one finds the fuel to do things without regard to their externalities. Trying to plan things based on externalized value misses the fact that our future conceptions of and relations with technology and one another are becoming inconceivable as we move towards a paradigm shift in how we create and conceive of value.