deleted by creator

Well, it is true that computer programs have far surpassed humans in board games. They are very well suited for it. It just has nothing to do with the hypothesized abilities of future “AI” as rationalists conceive them.

Oh, I didn’t know that!

The paper is itself written by LLM.

flaviat explained why your counterexample is not correct. But also, the correct statement (Liouville’s theorem) is that a bounded entire function must be constant.

CS has a huge number of people who think you can derive the solutions to social problems from first principles. It’s impossible to reason with them.

In this case these are grade schooler’s (roughly ages 9-18) essays for a standardized test, so there isn’t a body of students who could grade them.

Given consistent trends of exponential performance improvements over many years and across many industries, it would be extremely surprising if these improvements suddenly stopped.

I have a Petri dish to sell you

codeword is banana bread

Will there be statues to swap as well?

Mistress

is this more evidence that robots are taking our jobs? the investigation continues…

The Democratic party’s presidential nominee for 2028 will be chosen by a primary election, which will also be in 2028. Until a candidate wins the primary election, there is no nominee. This is important because Newsom is a sack of shit and we have multiple years with which to oppose his run for nomination.

He is not the Democratic presidential nominee for 2028.

There’s a part where they quote someone saying “I am not particularly confident [in my p(doom)]” and I’m still remembering getting talked down to about how Actually all subjective uncertainty can be represented as probabilities because Bayesianism and you wouldn’t happen to be one of those stupid frequentists right?

That’s weird, The Register’s versions of the quotes are different (not just pared down).

“Replication of Quantum Factorisation Records with an 8-bit Home Computer, an Abacus, and a Dog”

You’re totally misunderstanding the context of that statement. The problem of classifying an image as a certain animal is related to the problem of generating a synthetic picture of a certain animal. But classifying an image of as a certain animal is totally unrelated to generating a natural-language description of “information about how to distinguish different species”. In any case, we know empirically that these LLM-generated descriptions are highly unreliable.

From the appendix:

TOTAL, COMPLETE, AND ABSOLUTE QUANTUM TOTAL ULTIMATE BEYOND INFINITY QUANTUM SUPREME LEGAL AND FINANCIAL NUCLEAR ACCOUNTABILITY

This week the WikiMedia Foundation tried to gather support for adding LLM summaries to the top of every Wikipedia article. The proposal was overwhelmingly rejected by the community, but the WMF hasn’t gotten the message, saying that the project has been “paused”. It sounds like they plan to push it through regardless.

The actual pathfinding algorithm (which is surely just A* search or similar) works just fine; the problem is the LLM which uses it.

I like how all of the currently running attempts have been equipped with automatic navigation assistance, i.e. a pathfinding algorithm from the 60s. And that’s the only part of the whole thing that actually works.