What Lee Robinson thinks about LLMs
Works on ML at SpaceX, as of 2026; formerly at Cursor, and five years at Vercel before that, ending as VP and working on Next.js. Writes at leerob.com.
Lee Robinson did not write this page.
We collected these quotes from things they published elsewhere, and every quote links to where it was said. They have no account here and have not endorsed this site. Quotes are word for word; the short line under each one is our own restatement, not their wording. Their own site. Is this you? Claim it or ask us to remove it. Or tell us what is wrong here.
Filter LLMs · Newest first
Only Lee Robinson on LLMs. See everyone on LLMs
10 dated positions, 2025 to 2026, in their own words. Our reading of what Lee Robinson has said — not written or endorsed by them.
-
Their wordsIt's funny, you don't really hear people talk about hallucinations anymore, because it's kind of a solved problem, mostly.
↗Stanford CS146S: How always-on agents workyoutube.com 4th of 22 in this recording
-
Their wordsBut, if you see the model doing, you know, 90% of its time doing shell commands for this one thing, it actually can be more efficient to make a dedicated tool for some of that stuff.
↗Stanford CS146S: How always-on agents workyoutube.com 9th of 22 in this recording
-
Their wordsOkay. Context engineering is al- also known as harness engineering. Like, I feel like it's kind of the same thing. They've evolved over time.
↗Stanford CS146S: How always-on agents workyoutube.com 12th of 22 in this recording
-
Their wordsAnd what we mean by this is you kind of have to internalize and assume that the models today are the worst they'll ever be, which means that in 6 months uh you might need to completely rebuild the UI.
↗Stanford CS146S: How always-on agents workyoutube.com 14th of 22 in this recording
-
Their wordsI also think that it's a pretty safe bet to assume that at this time next year there will be significantly, significantly more tokens flowing in the world
↗Stanford CS146S: How always-on agents workyoutube.com 19th of 22 in this recording
-
Their wordsBut I think it's helpful to understand just like when you're building a web app, it's helpful to know how databases work because increasingly the there are so many decisions you make at the product and the harness level that are influenced by the models themselves.
↗Stanford CS146S: How always-on agents workyoutube.com 22nd of 22 in this recording
- 3 months earlier
-
by
Their wordsKnowledge and experience are two different stats, and ideally you want to max both.
- 3 months earlier
-
Our readingAsking several models to fix the same bug and comparing results gives multiple shots on goal.
Their wordsSo, first, you can ask multiple models to fix the same bug, and then you can compare the results. Every model can take a somewhat different approach, and when you have multiple shots on goal, you can then pick the best option out of the few.
↗Cursor: coding agents tutorial (2026)youtube.com 9th of 18 in this recording
- 8 months earlier
-
Their wordsbut high reasoning is like a lot of reasoning and it might be too much. This is a good time where you need to really think about which model you're using.
↗Refactoring a Next.js & Tailwind app with Cursoryoutube.com 5th of 9 in this recording
-
Their wordsAnd this is the reality I think of how most people are working with AI models today.
↗Refactoring a Next.js & Tailwind app with Cursoryoutube.com 9th of 9 in this recording