Yeah, I am very open to skills and REST APIs, but REST APIs have an issue where the token gets injected, and so it's not really a good auth story for everyone. I think that skills have a feedback problem: how do you refine your skill that you've shipped to your customers? The suggestion you are giving isn't great for something like the ChatGPT or Claude iOS app. Where would you even store the token in that case, for example?
I wrote this piece to share my view about what cursor should do to manage some of the negative sentiment they’re getting online. What do you folks think?
For me it's not about difficulty, it's about friction. BJ Fogg's behavioral model B=MAP says Behavior = Motivation × Ability × Prompt. When you increase ability (lower friction), you get more behavior for the same motivation.
AI lowers my writing friction. Same motivation, more output. I write more, I get feedback faster, I iterate more. That's the value for me.
My reasoning: I use AI for development work (Claude Code), and better models = fewer wasted tokens = less compute = less environmental impact. This isn't a privacy issue for work context.
I regularly run concurrent AI tasks for planning, coding, testing - easily hundreds of requests per session. If training on that interaction data helps future models be more efficient and accurate, everyone wins.
The real problem isn't privacy invasion - it's AI velocity dumping cognitive tax on human reviewers. I'd rather have models that learned from real usage patterns and got better at being precise on the first try, instead of confidently verbose slop that wastes reviewer time.
I agree it's not about the 10x engineers or the greenfield. I think YC's selection process is still focused on finding distinguished individuals, but within two specific constraints.
I've been using Figma to build my decks for a while now, and I bet a lot of other folks have too. Figma Slides looks like it takes what we've been doing and makes it way smoother.
Agreed it's super hard. I think that's why you need to be careful about looking at data until you've got your hypothesis. Too much data and you'll overfit or introduce bias. I think of it more as a way to update your gut with an updated map of the state.
No definitely not. Kafka is for LinkedIn scale. Postgres is for everyone else.
Quick anecdote. I remember a friend who joined Shopify relatively early. He took ALL their data from many different services (analytics, Kafka etc) and loaded into Postgres on his laptop so he could query it efficiently.
It's amazing how over-engineered some startups are.
Previously: fforward, Delphia, Autodesk, Kera