While it doesn't fit as nicely with your theme of "subtraction" as core to disruption, another psychological phenomenon that this reminded me of (and I think will slow large companies down) is the Endowment Effect - the over-attribution of value to something you own simply because you own it (a classic example is kids in a classroom handed a toy car and the price that they are willing to sell it to other classmates vs what their classmates are willing to pay for it highlights the disconnect).
The idea of starting from scratch (vs leaning into what has worked for them for so long) is scarier in larger orgs, where there are generally more people who want to maintain the status quo because it is associated with their quality of life. However, with AI, I think we will see products gain market share and become deprecated faster than we have traditionally seen. The idea of file > new project should be exciting, even for established companies that have successful products.
Alex, I love the way you're applying the Endowment Effect here. I was reading about the Endowment Effect while writing the article, but couldn't quite see how it fit. This makes total sense. Although I wasn't there in the mobile years, there was a sense of endowment at Microsoft about Windows, and it felt impossible that the Mac would be able to take meaningful share.
I think the Endowment Effect is playing another role as well -- it's probably contributing to the lower retention that AI apps are seeing. People feel a sense of ownership over what they create, and that feeling of ownership is stronger when they meaningfully participate in the creation. I think retention will increase as generative AI apps move from a one-time "prompt > generate" user experience to a more iterative and participatory experience. Sequoia talked about this as moving from "Copilot to Director’s Mode".
So well said. I need to educate myself on that Sequoia talk but completely agree- effort is probably a better proxy than dollars spent in predicting someone's perceived value in something they own/helped create- whether personal or with an organization. Enhancing the co-creation experience with LLMs (beyond tweaks to chat messages) will be an interesting trend to watch.
Small correction: Photoshop was never a contender in the UX designer software race. From Adobe, there was XD, or maybe InDesign and Illustrator. But the winner of that generation was Sketch, until Figma arrived.
Your main point stands though: multiplayer and browser native blew Adobe (and everyone else) out of the water.
Ever since Ben Thompson wrote about AI as a sustaining innovation (https://stratechery.com/2023/ai-and-the-big-five/), I've tried to find evidence both in support of that idea as well as to the contrary.
Where I saw some evidence to the contrary, e.g. specialized models such as your Harvey example, I have wondered about their longevity given the GPT-Store might trivialize their product -- but I'm too optimistic in the grit and craft of start-ups to write them off too soon.
Maybe 1-2 new AI-enabling companies (e.g. OpenAI, Anthropic) grow to be huge, and otherwise AI is a sustaining innovation? Time will tell.
I love your reference to Casey's technological shift vs distribution shifts -- I wasn't familiar with that framework. It's very helpful!
Hi Jason, thanks for sharing the Ben Thompson post -- always fascinating to read his perspective!
I'm also optimistic about startups , and I think there will be many that flourish in spaces that are under-appreciated or overlooked by the big platforms. My sense is that OpenAI won't go deep enough in solving domain specific problems. For example, hallucination is a nuisance for most use cases, but it's mission-critical in the legal setting where accurate, deterministic use of legal precedent and case citations is essential. This is why I think vertically focused startups like Harvey can win.
Love this, thanks Ravi!
While it doesn't fit as nicely with your theme of "subtraction" as core to disruption, another psychological phenomenon that this reminded me of (and I think will slow large companies down) is the Endowment Effect - the over-attribution of value to something you own simply because you own it (a classic example is kids in a classroom handed a toy car and the price that they are willing to sell it to other classmates vs what their classmates are willing to pay for it highlights the disconnect).
The idea of starting from scratch (vs leaning into what has worked for them for so long) is scarier in larger orgs, where there are generally more people who want to maintain the status quo because it is associated with their quality of life. However, with AI, I think we will see products gain market share and become deprecated faster than we have traditionally seen. The idea of file > new project should be exciting, even for established companies that have successful products.
Alex, I love the way you're applying the Endowment Effect here. I was reading about the Endowment Effect while writing the article, but couldn't quite see how it fit. This makes total sense. Although I wasn't there in the mobile years, there was a sense of endowment at Microsoft about Windows, and it felt impossible that the Mac would be able to take meaningful share.
I think the Endowment Effect is playing another role as well -- it's probably contributing to the lower retention that AI apps are seeing. People feel a sense of ownership over what they create, and that feeling of ownership is stronger when they meaningfully participate in the creation. I think retention will increase as generative AI apps move from a one-time "prompt > generate" user experience to a more iterative and participatory experience. Sequoia talked about this as moving from "Copilot to Director’s Mode".
So well said. I need to educate myself on that Sequoia talk but completely agree- effort is probably a better proxy than dollars spent in predicting someone's perceived value in something they own/helped create- whether personal or with an organization. Enhancing the co-creation experience with LLMs (beyond tweaks to chat messages) will be an interesting trend to watch.
Small correction: Photoshop was never a contender in the UX designer software race. From Adobe, there was XD, or maybe InDesign and Illustrator. But the winner of that generation was Sketch, until Figma arrived.
Your main point stands though: multiplayer and browser native blew Adobe (and everyone else) out of the water.
Ever since Ben Thompson wrote about AI as a sustaining innovation (https://stratechery.com/2023/ai-and-the-big-five/), I've tried to find evidence both in support of that idea as well as to the contrary.
Where I saw some evidence to the contrary, e.g. specialized models such as your Harvey example, I have wondered about their longevity given the GPT-Store might trivialize their product -- but I'm too optimistic in the grit and craft of start-ups to write them off too soon.
Maybe 1-2 new AI-enabling companies (e.g. OpenAI, Anthropic) grow to be huge, and otherwise AI is a sustaining innovation? Time will tell.
I love your reference to Casey's technological shift vs distribution shifts -- I wasn't familiar with that framework. It's very helpful!
Hi Jason, thanks for sharing the Ben Thompson post -- always fascinating to read his perspective!
I'm also optimistic about startups , and I think there will be many that flourish in spaces that are under-appreciated or overlooked by the big platforms. My sense is that OpenAI won't go deep enough in solving domain specific problems. For example, hallucination is a nuisance for most use cases, but it's mission-critical in the legal setting where accurate, deterministic use of legal precedent and case citations is essential. This is why I think vertically focused startups like Harvey can win.