Karolis Dzeja

Exile

Nick Shirley uncovers an adult day care in Flushing, Queens with 7,000 phantom members. Nick: “This public document says you have 7,899 members.” Employee: “No, we don’t have 7,000 members.” Nick: “So you’re overbilling then? You’re getting paid $1,600 per patient — that’s how you got $12.9 million in 2024.” Employee: “Please leave.” American taxpayer dollars at work.

🚨 GPT-5.6 Sol's juice values (thinking budgets) have been severely degraded compared to release day If Sol now feels faster and more "efficient", this is probably why Terra and Luna juice values aren't affected, so their thinking budgets are now higher than Sol's

The fact that Agatha Christie had access to ChatGPT is undeniable proof that time travel is real

I still can’t figure out why Gemini struggles to compete with Claude and GPT. - Owns Chrome - Backed by Android - Stores most search results - Holds ~95% search history - Google has the biggest user data - Even incognito data isn’t fully private So what’s the problem?

Before bench press was invented (by Hackenschmidt, pictured) even bodybuilders had flat, underdeveloped chests. It's possible that not a single man in history had huge pecs until the 20th Century. Classical statues are all like this too. A little sad to think about.

The narrative of a rising philosopher class in the AI industry is just not plausible, I'm afraid—whatever anyone might say or sincerely think. Don't get me wrong, may all my philosopher friends get cool gigs at big AI labs! But it's structurally impossible that human philosophy could have any say in the ultimate destination of the top models, or top model (i.e. AGI). AI is a market phenomenon, subject to extreme competitive selection. It is simply not believable that any frontier AI lab could be optimizing for anything other than the present value of future discounted cash flows, simply because it would not have made it this far. And if the philosophers on staff are not driving some enlightened divergence from this objective function, then those are just called workers. Philosophy will probably matter greatly in the AI era, as a solution to the madness that AI will indirectly bring about, but it does not at all follow that any human philosophy or philosopher will have any say in that which AI brings about.

I think an unusual belief I hold is that books are honestly pretty poorly written on average, and AI-reading books is honestly a better way of extracting information from books

"visa, do you have a favorite bible verse?" i'm glad you asked

It’s worth appreciating a few years ago all LLMs blatantly hallucinated all the time—even if you tried to prompt around it (see “yo be real” for GPT-3). Now it’s rare enough in frontier models people find it interesting you can still make it happen with adversarial shenanigans:

Another white Christian preacher has been arrested by British police because his prayers caused anxiety to some people. This happens in Islamic caliphates.