
Lessons from Simon Willison
The items below are things Simon Willison said on podcasts about AI-assisted programming and building software. — Simon Willison on Talking Postgres with Claire Giordano, How AI is changing software development (August 2026).
On Talking Postgres, Willison said: "I want every line of code they write to have been exercised, and the easiest way to get them to exercise them is to tell them to write tests, 'cause they're very good at writing tests generally". — Simon Willison on Talking Postgres with Claire Giordano, How AI is changing software development (August 2026).
On Talking Postgres, Willison said: "I feel like our professional responsibility is we have to understand it". — Simon Willison on Talking Postgres with Claire Giordano, How AI is changing software development (August 2026).
On Talking Postgres, Willison said: "Like AI writing, if you're trying to be convincing or creative, I think it's garbage". — Simon Willison on Talking Postgres with Claire Giordano, How AI is changing software development (August 2026).
On Talking Postgres, Willison said: "You have to have put the effort in to make sure that the thing that you're communicating to them is good and it reflects the exact truth in your mind of what matters". — Simon Willison on Talking Postgres with Claire Giordano, How AI is changing software development (August 2026).
On Talking Postgres, Willison said: "We live in a world of where code has gone from the most valuable asset to almost the least valuable asset". — Simon Willison on Talking Postgres with Claire Giordano, How AI is changing software development (August 2026).
On Talking Postgres, Willison said: "If you don't have the discipline, you can end up with a Winchester Mystery House". — Simon Willison on Talking Postgres with Claire Giordano, How AI is changing software development (August 2026).
On Talking Postgres, Willison said: "You have to be able to express it very clearly, here is the thing that needs to be done, and figure out, okay, and here are the bits of information you will need to get those things done". — Simon Willison on Talking Postgres with Claire Giordano, How AI is changing software development (August 2026).
On Talking Postgres, Willison said: "At the same time, I think skill atrophy itself, at least as a software engineer, is a choice that you make". — Simon Willison on Talking Postgres with Claire Giordano, How AI is changing software development (August 2026).
On Talking Postgres, Willison said: "The whole point of the human in the loop is to protect other people from dumb mistakes that AI makes". — Simon Willison on Talking Postgres with Claire Giordano, How AI is changing software development (August 2026).
On Oxide and Friends, Willison said: "That's, I think that's something I keep on coming back to, is I do believe it is possible to write software that doesn't have exploits". — Simon Willison on Oxide and Friends, The Open Weight Revolution (July 2026).
On the Heavybit podcast, Willison said: "Like, producing software is a ferociously difficult thing to do, and you could give me all of the AI tools in the world, and it's still really difficult what we're trying to achieve here". — Simon Willison on Heavybit Podcasts, Ep. 9: The AI Coding Paradigm Shift (May 2026).
On the Heavybit podcast, Willison said: "Honestly, all of my influence in the AI industry comes from the fact that I have a blog, and I frequently update my blog". — Simon Willison on Heavybit Podcasts, Ep. 9: The AI Coding Paradigm Shift (May 2026).
On the Heavybit podcast, Willison said: "I'd say that November was the point when a coding agent can now be a reliable partner". — Simon Willison on Heavybit Podcasts, Ep. 9: The AI Coding Paradigm Shift (May 2026).
On the Heavybit podcast, Willison said: "I'm starting to treat the agents in the same way, and it still feels uncomfortable because human beings are accountable for what they do". — Simon Willison on Heavybit Podcasts, Ep. 9: The AI Coding Paradigm Shift (May 2026).
On the Heavybit podcast, Willison said: "It used to be if you found a GitHub repository with 100 commits and a good README and, like, automated tests and stuff, you could be pretty sure that the person writing that had put a lot of care and attention into that project". — Simon Willison on Heavybit Podcasts, Ep. 9: The AI Coding Paradigm Shift (May 2026).
On the Heavybit podcast, Willison said: "Like, the entire software development life cycleWas, it turns out, designed around the idea that it takes a day to produce a few hundred lines of code, and now it doesn't". — Simon Willison on Heavybit Podcasts, Ep. 9: The AI Coding Paradigm Shift (May 2026).
In the 2024 year-in-review post, Willison wrote: "I think people who complain that LLM improvement has slowed are often missing the enormous advances in these multi-modal models". — Simon Willison, Things we learned about LLMs in 2024 (December 2024).
In the 2024 year-in-review post, Willison wrote: "I think this means that, as individual users, we don’t need to feel any guilt at all for the energy consumed by the vast majority of our prompts". — Simon Willison, Things we learned about LLMs in 2024 (December 2024).
In the 2024 year-in-review post, Willison wrote: "The models may have got more capable, but most of the limitations remained the same". — Simon Willison, Things we learned about LLMs in 2024 (December 2024).
In the 2024 year-in-review post, Willison wrote: "We need to be talking through these problems, finding ways to mitigate them and helping people learn how to use these tools responsibly in ways where the positive applications outweigh the negative". — Simon Willison, Things we learned about LLMs in 2024 (December 2024).
In a 2025 post on using LLMs for code, Willison wrote: "They may well have stumbled on to patterns that work, but those patterns do not come naturally to everyone". — Simon Willison, Here’s how I use LLMs to help me write code (March 2025).
In a 2025 post on using LLMs for code, Willison wrote: "I find LLMs respond extremely well to function signatures like the one I use here". — Simon Willison, Here’s how I use LLMs to help me write code (March 2025).
In a 2025 post on using LLMs for code, Willison wrote: "The trick here is to dump the code into a long context model and start asking questions". — Simon Willison, Here’s how I use LLMs to help me write code (March 2025).
In a 2023 talk write-up, Willison wrote: "It turns out the moment you stick a chat interface on it the capabilities of the system suddenly become obvious to everyone!". — Simon Willison, Catching up on the weird world of LLMs (August 2023).
In a 2023 talk write-up, Willison wrote: "The more expert and obscure the question the more likely you are to run into convincing but blatantly wrong answers". — Simon Willison, Catching up on the weird world of LLMs (August 2023).
In a 2023 talk write-up, Willison wrote: "If you get into this space, the most important thing is that you share what you are learning with other people". — Simon Willison, Catching up on the weird world of LLMs (August 2023).
In a 2025 post on vibe engineering, Willison wrote: "This feels very different from classic vibe coding, where I outsource a simple, low-stakes task to an LLM and accept the result if it appears to work". — Simon Willison, Vibe engineering (October 2025).
In a 2025 post on vibe engineering, Willison wrote: "Just like human programmers, an LLM can only keep a subset of the codebase in its context at once". — Simon Willison, Vibe engineering (October 2025).
In a 2025 post on vibe engineering, Willison wrote: "Beyond automated tests, you need to be really good at manually testing software, including predicting and digging into edge-cases". — Simon Willison, Vibe engineering (October 2025).
In a 2025 post on vibe engineering, Willison wrote: "It signals that this is a different, harder and more sophisticated way of working with AI tools to build production software". — Simon Willison, Vibe engineering (October 2025).