13 chunks · 12 episodes
If your AI-powered product treats the model as commodity, then it will improve as the best models improve. You can switch to whatever model is the best at any given time. That means the quality of your product is the max of any model. If your product is tied to any one model, then you can get stuck.
A trick to keep a toddler focused and engaged on the task: give them faux agency. Not: "Are you ready to get dressed for school?" "Do you want to pick the red shirt or the blue shirt?" You take for granted that the thing we're doing is getting ready for school, But because they have a choice, they f
An interesting economics paper about why OpenAI is moving so strongly into vertical integration. When there's a monopolist whose product is used as an input to a downstream product, they'll vertically integrate to maximize profit. The reason: downstream customers will underutilize the input from the
Anthropic is about to release a feature of LLM-powered software. Opal from Google is a similar shape, but without custom UI. Normal UI, but LLM guts underneath. Someone will figure out the right complement of normal software and LLMs, and it will change the world. It's a hard combination to nail, bo
Seems like a certainty that in 10 years, most US consumers will pay $100 a month for an AI-powered product. In order to not be a culdesac, it will have to be an open system that you can use for anything. It will need to subsume all of the other use cases. It will have to be bigger than just chat. Th
What is the order of magnitude that people will spend on LLM-powered services in a few years? If LLMs integrated deeply with your life, in a way that helps you live aligned with your intentions, are super valuable, perhaps it's way higher than we think. Maybe the right order of magnitude is not Netf
One of the reasons tech took over the world was zero marginal costs. But that doesn't apply to LLM-powered tools!
Imagine: a personal wiki that builds itself. There was never a way to use your personal knowledge graph before, so there wasn't a reason to distill and structure it. The graph was an end in and of itself. Only the most highly motivated organized would bother. But now with LLMs it can provide extreme
Tasks go from unstructured to structured as they exist for longer and get more baked. Chatbots are great for unstructured tasks but can't do structured well. To help with orchestrating our lives, LLM-powered tools will need more structure.
We're lucky to live in a universe with high-quality LLM APIs. Here's an alternate universe that is totally possible to imagine. OpenAI releases ChatGPT before they release any API. They don't ever release an API because it would be "dangerous" …and also undermine their app's differentiation and powe
There's a gap between Anthropic's Artifacts and OpenAI's GPTs. Anthropic Artifacts makes it super simple to create a little sandboxed live demo app with whatever UX you want that you can share with anyone you want. But the Artifact you distribute can't run additional LLM calls on the viewer's behalf
Pricing AI assistance for businesses is easier than for consumers. For a business, they can see the value is "the amount of salary I don't have to pay to get the same result." You can think of SaaS payments as being like mini-salaries to tools (and, increasingly, AI-powered agents). Value-based pric
With an early stage technology that has promising but uneven quality, you have to design the UI to be a good enough experience for the worst case, not an exceptional experience in the best case. You have to meet the technology where it is. If your UI sets an expectation of a quality level your backe