Sign in

Lucas

@lcsdcrz.bsky.social
44 followers 100 following 258 posts

Triste é não ter um verso maior que o literário.

PostsRepliesMedia
Lucas @lcsdcrz.bsky.social · 11/10/2026
let's granted that Tao example with the cancer verified-lean pill is not the best one; but are people really don't get his general point? like, it's 20 slides long and it isn't an new view. I have contact with it since may (bessis fall of theorem economy) and I no mathematician.
010
Lucas @lcsdcrz.bsky.social · 11/10/2026
Although that is true, OpenAi should coordinate those results with mathematicians, with the institutions etc; the community should have time to adjust with those new ways of doing mathematics. ''Break Things First Apologize Later'' will do much more harm than good for the math people.
130
Lucas @lcsdcrz.bsky.social · 07/10/2026
But how would someone do that?
000
Lucas @lcsdcrz.bsky.social · 07/10/2026
Maybe math is mature enough, has rich concepts and techniques that some of it can advance by itself when you have enough compute to survey it entirely. Layperson opinion.
000
Lucas @lcsdcrz.bsky.social · 06/10/2026
Any chance that it will be available online?
110
Lucas @lcsdcrz.bsky.social · 05/10/2026
I still can't fathom how bad it is. In Brazil we just elected the most right wing legislature in our democracy history. In three weeks, in the second turn, a 90% chance of the second son of Bolsonaro will be elected as president. All of it through reels from Instagram and tweets.
010
Lucas @lcsdcrz.bsky.social · 05/10/2026
tbh, when i think on folk psychology i am think about flaubert or tolstoi or shakespeare or whatever avatar that has a rich texture in describe humans so i very please with the idea of folk psychology as ability and enrichement of human mind (not something carve nature in the joint).
000
Lucas @lcsdcrz.bsky.social · 05/10/2026
I very poor educated in this debate but isn't the a idea that ''wait a minute, 'folk' psychology is rather something akin as ability to understanding human being rather a theory'' something that a wittgenstein-type-person would say? and aren't wittgensteinians the devil people in cognitive science?
100
Lucas @lcsdcrz.bsky.social · 28/09/2026
Maybe this is the reason that it is famous: it is a "horn", something that you know from your ordinary experience. Like, I can't fathom what would be like to bring a bag with alephzero oranges in it.
010
Lucas @lcsdcrz.bsky.social · 28/09/2026
I think because you can interact with 3d objects in your daily life and it would be very strange to encounter such an "object".
110
Lucas @lcsdcrz.bsky.social · 28/09/2026
but I'm a big fan of the lady with the rock that broke the bridge glass doing the baby fall into the river and a dog salved the little baby from the river and delivered the baby to its mommy
010
Lucas @lcsdcrz.bsky.social · 28/09/2026
so far the best that they came to do was a reddit for ai agent and make everyone crazy on twitter about the end of world as we know it, then they proceeds to forget about it in the next two days.
120
Lucas @lcsdcrz.bsky.social · 28/09/2026
From a sophisticated point of view, with math trained brain, it's look trivial or easy or whatever; but from the layperson it is contra intuitive. I remmember be pluzzed by it in high school till I learn calculus.
100
Lucas @lcsdcrz.bsky.social · 25/09/2026
This one hit differently because it was made by Aronofsky, someone with some prestige in Hollywood.
010
Lucas @lcsdcrz.bsky.social · 25/09/2026
Photography probably became an art form when they stopped pretending that it was paint.
010
Lucas @lcsdcrz.bsky.social · 25/09/2026
If AI video artists embraced the "weirdness" of generative video production instead of just replicating images of Walter White or IShowSpeed, I’d be much happier and would look forward to something interesting. But no: the first serious attempt was a generic, terrible short about USA independence.
210
Lucas @lcsdcrz.bsky.social · 24/09/2026
Nós ficamos, com justiça, confusos com posições que parecem mágicas: dualismo, por exemplo. Exigimos bons argumentos por assertivas misteriosas. Mas deveríamos também ficar surpresos e exigir o mesmo rigor para assertivas absurdamente redutivas como essa que Colin está respondendo.
000
Lucas @lcsdcrz.bsky.social · 24/09/2026
Why so? They both have deep respect for the work of Cavell.
100
Lucas @lcsdcrz.bsky.social · 23/09/2026
It's impossible that he can't see the tension in his takes. He must be ragebait y'all.
010
Lucas @lcsdcrz.bsky.social · 20/09/2026
Usually they don't hold sophisticated, nuanced, views. Clearly, when Altman was saying this, he was saying in the spirit of someone saying that we are nothing but a bag of atoms or that we are nothing but a bunch of heuristics. I remember the curiosity of Anderssen's introspection is bad and fake.
000
Lucas @lcsdcrz.bsky.social · 20/09/2026
I didn't know that, looks interesting. Lee Smolin iirc is very idiosyncratic. He has a book with a Brazilian philosopher about cosmology.
010
Lucas @lcsdcrz.bsky.social · 20/09/2026
I remember Altman tweeting something like: gpt is stochastic so do r u. Someshit like that.
010
Lucas @lcsdcrz.bsky.social · 20/09/2026
2021-2023 the tech bros were very happy with the stochastic parrot framing: their point was: humans are also stochastic parrots.
230
Lucas @lcsdcrz.bsky.social · 20/09/2026
The primacy of relation over substance, or that substance is constitutively a matter of relation.
110
Lucas @lcsdcrz.bsky.social · 20/09/2026
Leibniz holds something like that also, right?
111
Lucas @lcsdcrz.bsky.social · 20/09/2026
Cassirer?
000
Lucas @lcsdcrz.bsky.social · 20/09/2026
Não vão = vão*
010
Lucas @lcsdcrz.bsky.social · 20/09/2026
Minha primeira leitura na vida foi provocada por um livro de contos ilustrados com uma delicadeza profunda. Quatro volumes, cada um representando uma estação, 365 contos: um pra cada dia do ano, todas as páginas ilustradas. Triste imaginar que esses tipos de livros não vão praticamente desaparecer.
210
Lucas @lcsdcrz.bsky.social · 18/09/2026
Someone with some influence here posted a thread about the connection between the Rats and EA & the quasi-religious (read: sex cult) aspects that cascate from Yud.
010
Lucas @lcsdcrz.bsky.social · 18/09/2026
That is a very curious question, because AlphaGo Zero has some sort of conceptual bounders in the sense that we can individuate it in a way that is not available to LLM agents, except as roleplay character.
120
Lucas @lcsdcrz.bsky.social · 17/09/2026
like in eyes wide shut: if it was in real life I would asking why those people are like ganados in the castle from resident evil 4
110
Lucas @lcsdcrz.bsky.social · 17/09/2026
Type of thing that could be cool in a movie, but in real life is just cringe.
110
Lucas @lcsdcrz.bsky.social · 16/09/2026
Why should someone be obliged to subscribe to one available framework? Maybe there is a good argument against the three positions you list, although no one can provide the better point of view yet. And why are you concerned about NS? What frightens you maybe is the publicity surrounding it.
120
Lucas @lcsdcrz.bsky.social · 16/09/2026
And both have a sweden friend.
020
Lucas @lcsdcrz.bsky.social · 14/09/2026
Perhaps the way it is articulated in this passage makes the rationality seem very strong, in a way we rarely achieve. But I believe there is a strength to this type of argument that, even when weakened, captures our intuitions about reasoning that is not merely pattern matching.
000
Lucas @lcsdcrz.bsky.social · 14/09/2026
There is a tradition, dating back to Kant, that constitutively links our mental states (beliefs, intentions, reasoning processes) with an implicit, potential capacity for self-awareness. I believe this association connects 4/6 and 2/6 in a way that is absent in LLMs.
110
Lucas @lcsdcrz.bsky.social · 14/09/2026
I really don't know if it will be possible; and I am doubtful even more if mathematical problems can have $ as a proxy to importance for the math community.
010
Lucas @lcsdcrz.bsky.social · 14/09/2026
I suspect the hypothetical benefit lies in iterating new models that drastically reduce the cost of a task of this calibre: something that, frankly, has already happened with other tasks. We spend 20 million dollars on NS today so that, by, say, 2030, a problem of this level costs $20.000,00.
100
Lucas @lcsdcrz.bsky.social · 12/09/2026
bsky.app/profile/lcsd...
010
Lucas @lcsdcrz.bsky.social · 12/09/2026
You can easily see it displayed in talking about art: hollywood is cooked; artists are cooked; all of it is simply boys destroying your toy because they can't have them.
061
Lucas @lcsdcrz.bsky.social · 12/09/2026
TBH, a lot of it is driven by some kind of resentment. You grow up in a field adjacent to mathematics; you constantly hear mathematicians talking about the beauty of the discipline, and you can't see that beauty: so if you blow up the mathematician's worldview, you end up on equal footing.
192
Lucas @lcsdcrz.bsky.social · 10/09/2026
global.oup.com/academic/pro... (I'm not a mathematician)
global.oup.com
010
Lucas @lcsdcrz.bsky.social · 10/09/2026
rather than USA
000
Lucas @lcsdcrz.bsky.social · 10/09/2026
Do you believe that this approach, focused more on understanding than on problem-solving, comes more naturally to you because of your training in France?
120
Lucas @lcsdcrz.bsky.social · 10/09/2026
I think that a lot of technology was surrounded by those people when a new tech emerge. Like a rationalist cult type when we have electricity or steam engine.
000
Lucas @lcsdcrz.bsky.social · 09/09/2026
I believe these people are stupidly putting themselves at risk. When you say you hold the power (10%) to wipe out the human race, you are likely provoking someone who will try to end your life before it’s too late.
000
Lucas @lcsdcrz.bsky.social · 09/09/2026
050
Lucas @lcsdcrz.bsky.social · 08/09/2026
Qualia and sentience are not the same thing. I haven't read the book *Blindsight*—which seems to inform everyone about the distinction between mental content and consciousness—but it strikes me that people are confused here.
001
Lucas @lcsdcrz.bsky.social · 07/09/2026
The mistake here, I believe, is viewing this evolution as if it were a human going from basic arithmetic to solving a conjecture in four years. For a human, that would be incredible. For an AI? It might not hold much explanatory or predictive value.
130
Lucas @lcsdcrz.bsky.social · 07/09/2026
I think there is a significant error in the way people reason: "Models couldn't do arithmetic four years ago, and now they can conduct mathematical research! Just imagine what they'll be doing in another four years!"
120