It also gamifies the moderation, letting admin ban people for being kinda annoying or posting a profoundly bad take instead of quibbling over the rules all the time. You want back in? Ten bux.

"A letter of the alphabet also contains an immediate secret, like the corals in the petrified reef. The molecules have remained as they were shaped by life and they can be reanimated." is impressive intuition on Junger's part to say the least.

That Wikipedia article isn't very good, here you go.
cyborgism.wiki/hypha/eumeswil

I don't think they did. However Ernst Junger also predicted a specific machine like this, of all people.

en.wikipedia.org/wiki/Eumeswil

*smiles slightly*
Indeed. :3

You will enjoy this post if you haven't read it.
minihf.com/posts/2025-0...

I continue to feel like there's room for foundation model assisted media studies where you explore what various tropes and rhetorical devices "mean" by perturbing and parameter sweeping them in the machine. Looking at what happens with different frames and phrasings, etc.

Along with Baguettotron we release the smallest viable language model to date. Monad, a 56M transformer, trained on the English part of SYNTH with non-random performance on MMLU. Desiging Monad an engineering challenge requiring a custom tiny tokenizer. huggingface.co/PleIAs/Monad

How did you get into AI?

Apparently not!
bsky.app/profile/girl...

Where's the list?

Like basically if you have to insure a product made by a cabal who only give quotes after you buy the product (???) and don't do consistent pricing and want to bilk you for 10x in various line items at some point it would presumably be cheaper to buy into the cabal to figure out what the costs are.

"Why are the insurance companies not paying the money to get this information if they're so deeply adversarial and in such a fucked epistemic space that you're arguing about which order of magnitude a bill is in?"

Not quite? I'm not talking about like. "Why not vertically integrate?" I mean "Why is there this huge information asymmetry about information that like, exists, that is convergently accessible and replicable?"

ChatGPT says it's because insurance companies prefer to be asset light and worry about being accused of having a conflict of interest but if you're at the point where you're literally arguing about which *order of magnitude* a bill should be in is it really cheaper to haggle over that at scale?

I think it's definitely not cool for Google to take a product with decades of built up expectations about what its capabilities and affordances are, what its failure modes are, and then suddenly bolt onto the top of that a poor quality version of a product with totally different failure modes.