Juxtastat Test your knowledge of geography and statistics! New quiz every day

urbanstats.org/quiz.html

What is that one daily quiz called where you guess which us city has a higher xyz stat

I want a feature where I can see my own most liked posts

Notably, out of an abudance of caution, the Google team chose to “publish” its result via a cryptographic zero-knowledge proof that their circuit exists (so, without revealing the details to attackers). This is the first time I’ve ever seen a new mathematical result actually announced that way, although I understand that there’s precedent in the 1500’s, when mathematicians would (for example) prove their ability to solve quartic equations by challenging their rivals to duels. I’m not sure how much it will actually help, as once other groups know that a smaller circuit exists, it might be only a short time until they’re able to find it as well.

wild

Apollo 15 postal covers incident - Wikipedia

en.wikipedia.org/wiki/Apollo_...

sans mots www.brasstacks.blog/interview-wh...

You might be interested in gradual disempowerment:
gradual-disempowerment.ai

Interesting, i have two not yet unboxed. i was hoping to sell them for a profit some day, but thats clearly never happening lol.

In general anti-thing personalities tend to be more deranged than pro-thing personalities. Always better to orient around pro-thing

Can you play chess on RP^2

Reader + AI app | The Marginal Revolution Generative Book | Tyler Cowen Free reader and integrated AI companion for The Marginal Revolution: Rise and Decline, and the Pending AI Revolution

A book designed to be read with an AI. I enjoy the approach. One of the default options when highlighting a section is "find flaws in this argument". Love it.
tylercowen.com/marginal-rev...

We dont have a datacenter problem, we have an affordable datacenter problem

Drake no: competing in an efficient market Drake yes: consuming in an efficient market

Adverse selection rules everything around me

They don't need to be faster, just cheaper / hour.

And also: can you just repeat the repeated layers multiple times?

Makes me wonder if you could deliberately train a model to have this property

Showing the openAI tokenizer (https://platform.openai.com/tokenizer). The math writing software LaTeX takes 3 tokens to represent.

anti-math discrimination

Not comparing to other attention modifications makes this paper a bit sus

This feels like it does a similar thing to softmax1 , i.e. allows tokens to not attend to anything (by attending to itself)

100k gecs, when?

Probably not. Moving small amounts is not a solved problem

It is like with congestion pricing. Even if you burn the money it would still be more efficient.

Tbh, it wouldn't matter that much. The point is more to incentivize nicer behavior and less spam. But I suppose like money can go to the author, and tweet money goes to server maintenance ¯\_(ツ)_/¯

I want a twitter where each tweet costs 5¢ and every like costs 1¢.

I wonder if qwerty has subtlety nudged vocabulary more towards words that are easier to type on a qwerty keyboard.

Basic reading comprehension is (has become?) a 6 figure skill.

Getting dangerously close to actually talking about what I do here, but the boring answer here is kind of frustrating in a lot of policy related research on immigration because we endlessly look for policy related tweaks when there’s a giant neon sign with LET PEOPLE WORK floating overhead

Last one