They are also immune to embarrassment.

There are subtleties to this particular pissing match that aren't immediately obvious. The Vatican is starting to rein in the Opus Dei cult-within-a-cult. After John Paul II canonized its founder in 2002 on shaky grounds, later Popes have started to think that maybe they went a little too far. Meanwhile, Opus Dei has focused on gaining more secular influence in government (https://www.theguardian.com/us-news/article/2024/jul/26/kevi...).

But you knew that.

Still: does the combination of a deranged Truth Social posting, an obviously-pasted tweet from the Pakistani government, and a 10-point list of debatable provenance count as "clear public confirmation?"

I guess so, sort of, maybe? Fortunately I don't have a 7-figure wager at stake.

Kind of a disappointing piece of work from the guy who took down Theranos. His journalistic talents are sorely needed elsewhere right now.

I would be coughing up those email headers if I were him. Or forging some, if necessary.

That aside, I don't agree with the premise. Back might be Satoshi, but there's nowhere near enough evidence in Carreyrou's article to reach that conclusion. He should have run it by some other veteran figures in the crypto community, so they could point out how quotidian some of the language and tropes being cited really are.

It would be interesting to see some example problems along those lines. Design some games with complex rules, including one or two of the most subtle game-wrecking bugs you can think of, and ask the models if they can spot them.

In fact that sounds more interesting the more I think about it. Intensive RL on that sort of thing might generalize in... let's say useful ways.

Satoshi is either dead, or he lost his keys and probably wishes he were.

At that point, a lot depends on the quality of the preprocessing applied to the raw text dumps. It is reportedly not that trivial to go from DumpOfSketchyRussianPirateSite.zip to a data set suitable for ingestion during pretraining. A few bad chunks of data can apparently do more harm than one would expect.

AFAIK Google scans almost everything in print as part of the Google Books initiative, so they may have been able to skip the torrenting step.

I have other beefs with Amodei, including his pathetic, mewling appeals for regulatory capture and his forehead-slapping hypocrisy on copyright and ToS enforcement, but this seemed like a case where he was legitimately on the right side of the question and had the moral courage to stand by his position.

That's a valuable guarantee. So valuable, in fact, that you won't get it from Anthropic, OpenAI, or Google at any price.

Funny, I was about to say the same thing to you! Life is full of little coincidences.