That awkward moment when you're getting the models to play Mafia, but one of them just proposes a conversational protocol along with a proof that it maximizes town's chance of winning, which the others verify, then all adhere to. Oops, another one saturated.

"Instrumentally important" questions can be formalized as those which provide high mutual information with "terminally important" questions about the future. And indeed, anytime I make predictions about the future, I first condition on several possible answers to this question.

There’s a kind of monetary deflation that comes with technological progress, as the economy becomes capable of addressing ever broader categories of needs. One day, currency and well-being will be expressed in the same units, and a dollar will be valuable indeed.

When the aesthetics of forecasting win the day, "life expectancy" will finally refer to the average age a person born today can expect to live. Hell, there'll be a distribution too. I wonder what people will think of the early hump and the long tail...

DAOs of Loving Grace

In retrospect, no surprise that Elon is using transformers to turn his car company into a robot company…

This is a good thread that would approach the heart of the matter, if only it changed one word. Mathematics is a fundamentally *instrumental* subject. But a mathematical idea's instrumentality, both inside and outside the subject, does not need to be defined anthropocentrically.

They say you have the average AI timelines of the five people you spend the most time with.

I found this moving.

"Pure mathematics itself is a special-case of the yet broader objective of verifying arbitrary software to given specifications." 👀

The Great Legibilization is well underway

I too am concerned about short-term AI water use!

Two new charity ideas: givedirectly.ai makeawish.ai

The core abstractions of the coming paradigm are taking shape.

453 days. (To be fair to Epoch AI I think the actual FM T4 turned out well.) I do genuinely wonder what there will be left to benchmark in AI math 453 days from now, besides human math's very toughest open problems.

"Experiments"? Oh, you mean cost-sensitive tool calls.

Nice to have these on the record ;)

When an LLM acts happy (“EUREKA!”) or sad (“I have failed…”), is that meaningless mimicry, or does it reflect something “real”? We don’t know if LLMs are conscious. But they increasingly seem to exhibit wellbeing, pain, and pleasure as they get smarter Paper 🧵: