He wanted to learn to code a Roblox game. It gave 3 modes for the tutorial - do it yourself, copy-paste some examples or use AI. He rejected AI because he's seen AI games and they're all bad.
You get the sense that Keir Starmer has found something that the silent majority really want to support him on. We'll see what it does to his approval ratings ahead of any leadership contest.
Whether it's achievable or not is a different question of course.
Next time we see something obviously AI-written we should ask for the prompt instead of the finished artifact. At least this way you'll know what the person was really thinking.
UK isn't perfect, obviously. Most recently there has been a lot of debate about whether Palestine Action is really a terrorist organization or whether allowing Kanye West into the UK is a good idea or not.
But it's clearly nonsense to suggest that the UK is more repressive than these other countries.
1. Guy I know who is an exec coach says that a chatbot can do a lot of what he does.
2. I am absolutely seeing lots of people asking ChatGPT a question and treating its answer as the truth.
So I can well believe that, in an age where you can get a quick answer to anything for free, that the market for books is collapsing.
Having said that, I also suspect that many of these self-help books are effectively a blog post stretched out over hundreds of pages, so maybe a decline in their sales is no bad thing.
DHH on the Pragmatic Engineer podcast gets right to the nub of what it feels like coding with agents in the Spring of 2026.
> I didn't want to be a project manager for agents
> Running a bunch of agents feels less like being a project manager for agents and more like stepping into this super mech suit where suddenly I don’t just have two arms, I have twelve.
I had been using ChatGPT for help with some coding stuff - sometimes it worked reasonably well but for anything tricky it just failed.
But then recently I asked Claude Code to upgrade a particular library for me that I had forked and to reapply the changes I had made in my fork. It just did it as if by magic. As if it already knew everything about the original library. Given that the library is open source and on Github then Claude probably does know everything about it.
But then I can still have times when Claude Code does something really helpful (e.g. at work it helped create the first version of an MCP server for our own agentic application), while other times it can do some downright dumb things (like editing directly a library's code so something worked rather than changing my code so it worked).
Any time I'm feeling like the machines are about to take over I just do this: 1. Ask Claude Code to write some tests. 2. Same model, but different window, ask it to delete any tests that were just added but which don't test anything substantive.
It always finds something to delete that it had just added.
I'm not even talking about using a different model for code review or anything like that. Just asking the exact same model 10 seconds after it wrote some tests to delete the bad ones.
He wanted to learn to code a Roblox game. It gave 3 modes for the tutorial - do it yourself, copy-paste some examples or use AI. He rejected AI because he's seen AI games and they're all bad.
Oh, also, ChatGPT lies.