atomic14

In my last post Claude and I failed to get reinforcement learning to play Manic Miner, and we ended up solving the whole game with plain old search.

Around 10 years ago Google DeepMind showed that you could use deep reinforcement learning to play video games.

TLDR: AI is really good at detecting deceptive adverts - why isn’t Google using it?

Last time I built a detector that counts four-letter runs (n-grams) and tells you whether Claude Opus 5 wrote something.

On 2 August 2026 Anthropic began watermarking the text that new Claude models produce - their support pages cover which products are marked.

F’you, Claude, you are driving me absolutely bananas: The house convention is a short phrase naming the defect, and the archive is full of it: “A citation that names nothing”, “The safe zone that only the spawner obeys”, “One word that means five things”.

I read a paper this week called The AI Layoff Trap.

Here’s a response Claude came up with recently.

In November 2006 I put everything I owned in my parents’ loft, had a great many injections, and flew to Guatemala.

I keep needing the same thing - a really basic dev board for whatever Espressif module I’m playing with.

Fresh from resurrecting Tanks!

Back in 2010 I wrote a tank game.

Every time I put text on a TFT display with TFT_eSPI I end up staring at the same built-in fonts: chunky, jagged, and very 1980s.