The closest I could find to that claim was a Townhall article¹ mentioning that he posed for a picture with an imam whose mosque was attende by terrorists. Townhall as in the conservative radio/news site and not the NYC town hall.

1. https://townhall.com/news/jeff-charles/2025/10/20/fbi-agents...

Agreed, our real-life corporations are still trying to convince themselves (by 'them' I mean employees) that they're the good guys and wrap everything in a layer of HR-speak' "We're helping connect the world! Yay!", " We're collaborating with the ministry of War to zero people, but only if they have a signed judicial order", etc.

Their fictional counterparts don't bother.

Building Better Worlds ®

Publisher not author

When I was shopping for wearable body cameras last year, I searched for reviews on YouTube and about half the search results were creep shots of retail and restaurant workers. Gives you a good idea of who the early adopters for this tech are.

That's my personal theory too. The model is stuffing its own context with vaguely related tokens, which helps the attention heads retrieve the right tokens.

Yup. You basically just need something for probability to push off of

Reminded me of The Great Silence by Ted Chiang. https://electricliterature.com/the-great-silence-by-ted-chia...

I work on a small team maintaining multiple services and this is very similar to our approach. Although I generally find review easier _without_ splitting every PR by backend/frontend. The additional context helps and you don't have to switch back and forth between two sets of files/changes to make sure the interfaces align. Good ol' feature branches are also useful when you have multiple interdependent breaking changes.

P.S. Bruno if you see this please skip the LLM fluff-up next time. It makes it hard to read.

Don't blame LLMs if the zeitgeist for coffee shop websites is sepia tones and a Papyrus-like font.

My university's Principles of WebDev course 10 years ago had a similar assignment and the results all ended up looking like that too.

Unless he shows evidence of independent re-development like his LLM agent history or at the very least git commits, I'll consider that a lie

Another story about $BIGCORP's automated systems fucking over or locking out one of their users, but luckily the specific user was big enough or had a friend working at $BIGCORP that could escalate their complaint for proper human review.