Summary

Armin describes hacking a CarPlay dongle with LLM assistance, where the model itself surfaced the CatPlay project as a starting point — a discovery he didn't make independently. This experience crystallizes a broader pattern: when many people have similar conversations with the same models, they independently arrive at the same projects, tools, and ideas. The post asks whether the AI builder community's running joke about 'all working on the same things' reflects not just obvious ideas or social diffusion, but the shared latent capabilities being elicited from identical models. It ends open-endedly, questioning how much of what we build is truly our own versus a product of the paths models make most likely.

Key Insight

LLMs don't just accelerate our ideas — they supply the ideas themselves, and because we all use the same models, our supposedly independent discoveries may be convergent artifacts of shared latent capabilities rather than genuine originality.

Spicy Quotes (click to share)

• 4

  In the past, I would abandon many such projects for lack of tenacity. But my clanker is tenacious.

• 7

  But so are all of our clankers.

• 7

  In this case I did not find or decide on CatPlay, the model did.

• 3

  Some of the projects we're now attempting are happening because of conversations we have with them.

• 6

  It really got me thinking about how this could create situations in which completely independent people end up building things they believe are their own ideas.

• 6

  What if we took paths, because those were the paths that were more likely with current generation models?

• 7

  How much of what we build comes from eliciting the same latent capabilities from the same models?

• 7

  There is something powerful and strange about how LLMs diffuse knowledge and capabilities, while perhaps also nudging us all simultaneously and independently toward building the same things.

Tone

reflective, philosophical, curious