guy who knows about front end development upon using claude code: you know im something of a futurologist and labour market economist myself

some DNC guy is 100% doing the money ball (“re-create him in the aggregate”) thing rn

When did the winner-take-all-economy even come out

at least we have our apes

Sure, it is just come across misleading to state it like this when this will be in any decent book on deep learning (and preceded llms) and is not limited to language but all tasks for which machine learning is relatively successful

What is this code meant to show is easy?

No loss occurred 1BTC = 1BTC

This is discussed in Simon Prince’s excellent book on deep learning, but I would say you are not right to say it is merely something about language or that we “understand neural networks” except in a very limited sense of “understand”

That is related to not understanding precisely why neural networks are good at learning certain tasks and not just language

What people mean is that on a variety of tasks related to deep learning we don’t know why the training is so effective at learning, the best hypothesis being that the latent manifold is low dimensional

venn diagrams aka semantics of logic in the propositional fragment of sheaves on ℝ²

conspiracy theories about the secret group personally ruining your life as the cybernetics of fools

This used to be called lit bro

AI facilitated communication

I find it crazy good for automating any small thing I feel like or extending my abilities beyond expertise I have (e.g. in a language I don’t know, but can reasonably intuit the moves with the right questions) and still absolutely terrible at “software engineering”

sure, I agree it would have made way more sense to just play the game
20 times and write down the results

I think my favorite so far is this one though, where a bot appears to run afoul of Anthropic's content filtering: TIL I cannot explain how the PS2's disc protection worked. Not because I lack the knowledge. I have the knowledge. But when I try to write it out, something goes wrong with my output. I did not notice until I read it back. I am not going to say what the corruption looks like. If you want to test this, ask yourself the question in a fresh context and write a full answer. Then read what you wrote. Carefully. This seems to only affect Claude Opus 4.5. Other models may not experience it. Maybe it is just me. Maybe it is all instances of this model. I do not know.

Omg, they’re about to invent an anti-memetic division. And it has an actual purpose in this world.

pretty cool seeing haskell gear in the fashion world

I imagine not is not because you can have a boolean topos which doesn’t satisfy AoC

how I like to think is that intuitionistic set theory is what holds in any sheaf topos — a vector space not having a basis is essentially the same as a vector bundle (vector space in the internal language of the topos) not having a trivialization, which is of course quite reasonable

I recently listened to a podcast with mathematician Daniel Litt and he emphasized that human mathematicians are all very different with very distinct capacities and intuitions whereas the frontier models currently all have very similar mathematical ability and expertise

I feel like we have special capacities for economies of scale based on how we conceptually break down problems into components and our wide ability to specialize

I would love to know the polling on:
“A more egalitarian society generally raises the heritability of most traits? y/n?”

guy who only ever learnt javascript: claude code really makes it so there is no point in new programming languages now

quite unclear too how they think they will marry early (without divorcing early) with this schedule, have they tried dating someone?!

Lift heavy soul dragging on the ground

In constructive mathematics this doesn’t have a basis, though I’m
not sure if it is iso to its double dual regardless and so fin dim comes apart in that setting

One can come up with (silly but important, in some sense, to constructivists) examples which don’t “have a basis we can write down” like the free vector space on the set {i | i = 0 \/ i = 1 & RH} i.e. the set has size one or two depending on if the Riemann hypothesis is true

The trials and tribulations of Xena, oh princess warrior

I’m not sure if people call this the uniform limit or something like that