I do think it is your moral responsibility to understand what AI is if you are going to participate in public debate about AI, because it's your moral responsibility to have some accurate understanding about whatever it is you are discoursing about in general. This appears to be a minority position.

pense of the timely, but rather "to present to the American public works of lasting value, produced with the greatest care and stress on quality.Our editorial concept is to help spread knowledge and understanding of the essential questions of human life and culture."

The mission of Kurt and Helen Wolff’s publishing house Pantheon Books in the 1940s. More of this please.

if u think feature creep and careless implementation caused by overenthusiastic engineers is a novel problem that wasn't already very serious before AI was a thing, clearly u have not met my coworkers

A round, fried hors d'oeuvre sits on a plate with a shocked expression. It is surrounded by others like it, except for an empty spot next to it.

“GUYS! WAKE UP! SOME DUDE JUST ATE CARL!”

Chalkboard with math symbols on it

Math wasn't always +, −, ×, ÷, or =—these symbols took centuries to develop. Next time you write an equation, give a little nod to history! magazine.amstat.org/...

#MathHistory #STEM #FunWithNumbers #StatsSky

we built God but Its business plan turned out not to be viable

This is a dangerous Technology that needs Regulation. Nobody is served by underestimating the capabilities it has.

Dr. Fenner Tanswell | Can Mathematics Be Hacked? Infrastructure, Artificial Intelligence, and the... YouTube video by INI Seminar Room 1

The video of my talk at the Isaac Newton Institute a few weeks ago is now online: "Can Mathematics Be Hacked? Infrastructure, "Artificial Intelligence" and the Cybersecurity of Mathematical Knowledge" www.youtube.com/watch?v=O11k...

We already had placed our towels there

🤣🤣🤣

We have already Seven of Themen.

LEEROOOOOY
JENKINS

We take a leave of absence from Long Words from time to time, us Germans.

If we are going back to a time when a divinely-appointed mad king tries to control the Pope, we might as well ressurect the Hanseatic League

VP: *glaring at me*

Me: You told me that if I work on Saturday I can wear weekend attire.

Manager: But you aren't wearing pants.

Me: Weekend attire.

Thanks for the link to the article, but also to its author. Very interesting guy.

I will read that for a while. Pen and paper, no LLM cheating…

No Anti AI blocks today. I am getting too soft on them?

Congrats.

We all should be in awe.

Purposeful Predictions What is simulation, and what is it good for?

I tried to pin down what people mean by "simulation" and place it in the broader context of the study and design of feedback systems.

The proof gets formalized and then formally verified. Rock solid.

Claude can do screenshots. Ask it to have a look at the results - it likes to forget it is a vision model.

I wonder if it's possible to tell whether LLM has a true "understanding" of causal relationships in a system or whether it just looks like it does. But then what does "understanding" mean?

I believe there may be value in evaluating LLM's counterfactual reasoning. There's some research in this area.

Pretty much none of the truly simple methods in ML scale well. SVM, kNN, random forests are some of the simplest methods out there, and they don't scale at all. Meanwhile "train a transformer via backprop and gradient descent" is a very high-entropy method,

So it's only four now? Not bad for a glorified word guesser...

AI Solved A Mathematical Problem That Had Stumped The World’s Best Minds For Decades GPT-5.4 Pro cracked a conjecture in number theory that had stumped generations of mathematicians, using a proof strategy that no human had ever considered.

"Three Erdős Problems Fell in seven Days" then, to be more precise. It does not matter how many have been solved already, the one still life are hairy. www.forbes.com/sites/anisha...

Neither AI Rationalism or AI-Con Thought is all that helpful in explaining the technologies we confront right now. The former tends to launch into fantasy, repeatedly demonstrating how starting from ridiculous premises allows you to reason your way to ridiculous results. The latter tends to curdle into denialism, claiming ever more loudly that disliked technologies are useless even as they find ever more uses. We ought to be much more worried about the claims of the triumphalists than the denialists, since they are far more influential. But to successfully deflate their claims, we need a more grounded perspective on what AI and related technologies are capable of than can be provided by the denialists. [Last sentence is highlighted by me.]

I've linked to this post from @himself.bsky.social before, and I think it's worth linking again. It's part book review, part essay about the AI debate.

www.programmablemutter.com/p/ai-has-lim...

Here is an excerpt:

Als selbstironischer Witz funktioniert es irgendwie nicht, als ernstgemeinte Aussage klingt es eher wie eine bösartige Karikatur auf Euch Humanisten. Ich will als alter STEM-Barbar gekonnt veräppelt werden, das könnt ihr Humanisten doch sonst so gut. Angriffsfläche gibt es ja bei uns genug 😁

"Jede Schreibmaschine kann Wörter..." - doch, genauso liest sich das.