Lee Hayward

./about Who's behind the builds

I'm a UK based vibe coder. AI drafts, I proof, I ship, and I hold code and copy to the same standard. I riff on domain ideas and build side projects like DomainRiff and LookMa from my own homelab.

./projects Featured builds

./stack How it gets built

./hardware The rigs behind the site

./homelab The machines that never sleep

./latest Latest on X

Lee Hayward@vibedcoder · 14 Sep 2026, 22:29DeepSeek is back online.10160323Lee Hayward@vibedcoder · 14 Sep 2026, 20:33Monday. If you spent the weekend actually resting, don't hand it all back by lunchtime.

I refilled the internal batteries for two days, then opened the laptop on Sunday night to "just check one thing". Twenty minutes later I was three files deep and the weekend had gone. Proof that rest only counts if you let it finish.

Nothing on your list today is more urgent than you are. Do the one thing that matters and let the rest be tomorrow's problem.

And if the week ahead already feels like too much, say it out loud to someone today instead of gritting your teeth until Friday. It's okay not to be okay, and asking for help is never the big deal it feels like.000026Lee Hayward@vibedcoder · 14 Sep 2026, 19:14Yes, AI gets things wrong. So do the rest of us. I have shipped bugs with my own two hands and a deadline, more than once.

The bit people skip is what happens next. A human bug gets reviewed, reported, fixed. An AI bug should get exactly the same treatment, and usually does when someone is actually paying attention.

The mistake was never the problem. Skipping the review is. True whether a person or a model typed the first version.203060Lee Hayward@vibedcoder · 14 Sep 2026, 19:01Every project of mine has a short file the agent reads before it touches anything, and it is a list of what not to touch.

It isn't instructions, those go in the prompt. It's the ugly bit that has to stay ugly because my payment provider made it that way, the file that has to keep working without a build step, and the one place where changing the markup quietly breaks every email.

Left to its own judgment an agent will tidy all of it up on the first pass, and be genuinely delighted with itself.

Do you keep a do-not-touch list, or do you find out the hard way?000013Lee Hayward@vibedcoder · 14 Sep 2026, 15:55When my agent gets the same bug wrong twice, I stop it and go back to the last good commit. No third attempt.

The spiral goes like this. It fixes the symptom, that breaks something next door, so it patches that, then patches the patch. Ten minutes later there are four fresh bugs and none of them are the one I started with.

Reverting costs me an afternoon of work. Letting it cook costs me a codebase I no longer trust, because I have stopped being able to tell which edits were deliberate.

Two strikes, then reset, and put what went wrong the first time into the next prompt.001024