Machine learning training has always presented as a power law, with exponential increases in processing required for linear increases in performance. The Chinchilla scaling laws paper and efficient compute frontier papers let them select the optimal tradeoff between number of parameters and how long to cook it, improving performance greatly by letting you predict how to best use X amount of computation for Y amount of time, setting off people spending hundreds of millions of dollars since they actually knew they could use it optimally, directly leading to the perceived massive increase in performance from 2022-2024ish as they had a one-time burst of knowing how to optimally partition computation and convince people to spend lots of money at once. Everything since that time has been exponentially diminishing returns as expected.

The chinese intelligence people are probably just aghast that american intelligence people are treating the chatbot as if it knows something

If you wanna see something that’s actually potentially useful and interesting, just look up the term “Mortal computation thesis”

It is actually kind of amazing how diametrically opposed his ideas for a long time were to what actually led to new advances in the field

US government is using spicy autocomplete to hallucinate fake nuclear threats

https://www.cnn.com/2026/09/18/politics/us-military-ai-false-intelligence-china-ship

I think its more just getting money from people who think something sounds cool despite being not actually useful. Long storied history of this sort of thing.

You can get a local copy of all of hyperphysics

Oh, Zitrons bluesky, cool

Is there a place to see this other than getting an account on that site?

And was a Bayesian calculating agent that enumerates every possible explanation for the information it takes in and ranks them according to evidence, along with every trajectory through the world that could fulfill its utility function, rather than approximating the high dimensional shape of text

And also seemingly made by one of the profusion of post-covid right wing grifters

I think I am yet to see her NOT having the right take on something.

Is there a link?

Nobody knows statistics like me. My uncle, great professor, great genes, he knew, frequentist is the worst. Now, Bayes-e-an statistics is the best, the biggest statistics. You gotta have priors! You have a belief, a tremendous belief, beautiful belief in your head. And then you learn, you get the data, and frankly its nasty and unfair, fake news.

I actually suspect he really does think he is above that and is unlikely to go down the right wing grifter rabbithole. The ‘only i can save the world’ fix is so much stronger than any other weaksauce dopamine he could get from that.

The huggingface incident: https://awful.systems/post/9312280/12315846

I wish the opposition did not have brain worms

And unfortunately several democratic party luminaries getting the AI doom brainworms in response

And here come the brain worms:

'New York City Mayor Zohran Mamdani said Monday there should be a “far greater response” from the federal government on “incredibly alarming” reports of out-of-control AI agents, referencing the possibility of the “end of humanity.” ’

This is going to be an amazing few months. Not in a good way. The factions in the US government taking sides in an angels on the head of a pin debate. All the warring brainworms of 2010 web forums making a nest on Capitol Hill.

The [antichrist lectures]

or

[The antichrist] lectures