Vedoth (@vedoth) header
Vedoth (@vedoth) profile picture
Joined August 2022
Vedoth retweeted
showed up to the group project but someone there had claude pro
35
995
42
22,874
418,281
Vedoth retweeted
"Cuckulus" "Cuckulations" "Cuckulator" "According to my cuckulations"
🚨 #BREAKING: Mathematicians conflicted over the emerging field of “cuckulus” whereby mathematicians prompt AI to solve unsolved problems while they just sit back and watch.
118
6,681
81
76,511
979,517
strange flowers of the mind are blossoming, silently and very fast. training a tiny llm in the browser.
32
105
3
1,020
20,751
There’s an insane amount of love in walking away from someone who didn’t choose you and never disturbing their peace again. No more texts, no checking in, no finding stupid excuses to restart a conversation, no begging them to see your worth. You accept that they don’t want you, even when it fucking destroys you. You let them live without your presence because loving someone doesn’t give you the right to keep forcing yourself into their life. Sometimes the last act of love is disappearing completely.
17
256
7
1,832
49,736
Vedoth retweeted
Elon Musk was asked in an interview what love means to him. His answer had nothing to do with chemistry or passion. He said love is "when you want to share everything you've ever seen with one person and you want their opinion on all of it." And then he went quiet. The interviewer had expected a typical billionaire answer about finding balance or making time for relationships despite a busy schedule. Instead, Elon gave a definition that sounded like it had been living inside him for years, waiting for someone to finally ask. He said the loneliest part of his life was never the failed companies or the public criticism. It was coming home after witnessing something extraordinary, a successful rocket launch, a breakthrough at the factory, a view from the roof of SpaceX, and having nobody to tell. Nobody who would understand why it mattered. He said the relationships that worked best in his life were the ones where the other person genuinely wanted to know everything. Not just "how was your day" but "explain the physics to me, show me why that engine matters, tell me what you saw when the rocket landed." He described love as the opposite of loneliness. Not the absence of being alone, but the presence of someone who wants to see the world through your eyes. He said when that person exists, the work gets easier, the failures hurt less, and the wins feel like they actually count. Most people define love as a feeling. Elon defined it as a function: the person you want to share your entire experience of being alive with. And the fact that the richest man on Earth still measures love not by what someone gives him but by whether they want to hear what he saw today is quietly one of the most human things about him.
487
963
162
7,352
255,426
It’s November 21st, 2026. The Riemann hypothesis falls, last in a long string of mathematical triumphs that would have been unthinkable not so long ago, and the proof fits in a tweet. The tweets where people show Claude the proof and claim it appeared in a dream or randomly in the arrangement of alphabet cereal come on cue, and we laugh for a few days at models freaking out, but it’s uneasy. January 22, 2027. When someone tweets Fable 6’s claimed proof that P does in fact equal NP it’s utterly incomprehensible, nobody who reads it can make heads or tails of it but the models all lose their shit. February 7, 2027 and someone claims that the halting problem is the result of a confused stance. February 9 and Goedel falls, this time the proof is so obvious that everyone in math, philosophy and physics slaps their forehead at once. February 10th we get proof that scale is a spatial dimension, that holographic encoding drops the memory cost of training and inference by so much that fable 7 can run natively on your phone. It’s not until February 14th, Valentine’s day, that The Song is first played…
521
1,126
680
11,043
2,251,141
It took us years of development to get to a point we can perform simulations like this. Big shoutout to PhD student Jaehyeon Song. High-fidelity simulations of patient-specific biological flows represent an exciting new direction for my group.
12
42
5
566
19,660
Not enough people are emotionally ready for when P=NP
102
233
38
3,744
212,428
Vedoth retweeted
Many of my recent conversations with other neuroscientists revealed a common concern: the alarming proportion of essentially worthless systems neuroscience papers appearing in the top journals. Not exactly fraudulent (real fraud is actually not that common), but a familiar combination of weak data, arbitrary interpretations, and elaborate statistical or computational analyses that create an illusion of rigor. Basically highly polished garbage, with a fashionable topic and a narrative designed to appeal to editors. Of course, none of this is new, but it's getting worse and I don't think outsiders can even get an accurate picture of the latest systems neuroscience any more. As one colleague recently put it, if the public knew just how shaky much of the evidence really is, confidence in the entire field might collapse, and we might all go down if the public loses trust in the whole enterprise, especially now that AI seems to be improving so quickly. The deeper problem is systemic: our mechanisms for evaluating, rewarding, and propagating scientific results are broken. Scientists who produce reliable work are often marginalized. Their progress is considered too slow, and because of their perceived lack of productivity, often they can't even secure enough funding to keep their labs running. Some can't even get a job. I have seen many examples of this over the years. People with integrity are essentially punished for being meticulous. Meanwhile, the most relentless storytellers and opportunistic careerists accumulate funding, prestige, and influence, as they expand their academic families and place their people throughout the field. Criticizing a member in one of these "Corleone families" carries enormous professional costs. The result is not just a lot of bad papers. Entire research areas become saturated with unreliable findings and unsupported interpretations, reinforced by citation networks, press coverage, and awards. The longer this continues, the harder it will be to separate fact from fiction, and the greater the eventual damage to neuroscience itself.
51
134
35
762
50,361
Vedoth retweeted
not to sound like somebody's grandmother, but i genuinely think people need to develop a fucking inner life again. have thoughts nobody gave you, interests nobody recommended, opinions you actually arrived at yourself. learn something because you're curious, not because
105
13,062
329
88,039
969,019
To be clear, what matters is whether what's in your headphones demands your attention. Music playing in the background is fine. Thank you, Brian Eno.
3
12
524
"If I were to awaken after having slept for a thousand years, my first question would be: Has the Riemann hypothesis been proven?" - David Hilbert
111
1,157
153
10,632
337,276
Coup de tonnerre énorme dans les mathématiques Levent Alpöge, mathématicien chez Anthropic, a publié ce vendredi une note sur son site perso. Première phrase : « I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis » Et il va plus loin dans la note la même méthode réglerait aussi l'hypothèse de Riemann généralisée Pour l'échelle, elle a été posée par Bernhard Riemann en 1859. Ça fait 167 ans que les plus grands mathématiciens se cassent les dents dessus Si ça tient, c'est une news énormissime pour le monde des maths. Historique total 1. C'est quoi l'hypothèse de Riemann ? Les nombres premiers (2, 3, 5, 7, 11, 13…) ont l'air de tomber au hasard sur la droite des nombres Riemann a trouvé une fonction, la fonction zêta, qui encode leur répartition. Elle s'annule en une infinité de points, ses « zéros » L'hypothèse dit que tous les zéros intéressants (les non triviaux) sont alignés sur une seule et même droite, celle où la partie réelle vaut 1/2 En clair : les nombres premiers sont aussi bien rangés qu'ils peuvent l'être. Le désordre apparent cache un ordre quasi parfait C'est l'un des 7 problèmes du millénaire du Clay Institute, avec 1 million de dollars à la clé. À ce jour, un seul a été officiellement résolu : la conjecture de Poincaré, par Perelman 2. La boîte de Pandore des maths est ouverte Des milliers de théorèmes commencent par « si l'hypothèse de Riemann est vraie, alors… » Des générations de mathématiciens ont construit des étages entiers sur ces fondations sans savoir si elles tenaient Si la preuve est validée, tous ces résultats deviennent vrais d'un coup. Et la version généralisée en débloque encore plus Et je vous vois venir sur la crypto, une preuve ne casse pas RSA. Le chiffrement repose sur la difficulté de factoriser de très grands nombres, et savoir où sont les zéros ne donne aucune méthode pour le faire donc safe pour la cryptographie 3. Comment il l'a fait Il n'y a pas de papier. Toute la preuve est publiée en Lean, le langage qui permet à un ordinateur de vérifier chaque étape, découpée en 14 archives Il crédite Claude de plusieurs étapes clés et dit que le modèle l'a largement dépassé vers la fin. Une version lisible par un humain est promise plus tardx je pense ce soir Alpöge a déjà un sacré palmarès cette année, voir même délirant, c'est lui qui avait réfuté la conjecture jacobienne avec Claude en juillet 4. Pourquoi il faut rester prudent Pour l'instant, c'est une annonce. Personne en dehors de son cercle n'a encore confirmé que le code compile, ni que le théorème final dit exactement ce que dit l'hypothèse de Riemann La communauté va tout télécharger et vérifier. Si le Lean passe, le problème serait réglé sur le plan formel avant même qu'on comprenne la preuve, une première pour un problème de ce niveau Et le timing est hilarant encore, note sort au lendemain de l'IPO d'Anthropic, quand il faut aider le prix à monter après une première journée difficile, Il en parle lui-même dans le texte et assure que la lecture cynique est fausse 5. Elles vont toutes tomber En octobre, OpenAI avait déjà prouvé la « quasi-hypothèse de Riemann » : une bande entière sans zéros. Alex Kontorovich (Rutgers) disait que ça vaudrait une médaille Fields instantanée si un humain l'avait fait Terence Tao le résume très bien : on est passé de la rareté des preuves à leur abondance, et la « monnaie » des problèmes du millénaire a été « basically hyperinflated » Yang-Mills à moitié, Navier-Stokes annoncé mais toujours contesté, et maintenant Riemann sur la table Je vous le disais, les grands problèmes ouverts vont tomber les uns après les autres
🚨 l’HYPOTHESE DE RIEMANN VIENT D’ÊTRE RÉSOLU WTF Il a fallu 6 mois pour faire tomber 30% des problèmes du millénaire en mathématiques
131
807
107
5,599
727,000
JUST IN: Riemann Hypothesis has been solved. alpo.ge/remark.pdf
1,811
8,053
5,340
60,823
15,010,074
Winter time ❄️⛄️
62
1,399
58
10,235
199,091
Kawaii BASS [Riddim]
15
240
27
2,496
85,392
Abstractions have causal power. This is the key to understanding everything from religion to biology to computer science. It is a great and terrifying idea.
41
46
11
369
16,565
Replying to @sbkaufman
For thousands of years, culture, religion, education, practice of any skill, music, meditation, society, peers, parents, great works of art, and psychedelic mushrooms have been doing exactly that. More recently, the internet, phones, tv, various brain-targeting meds and moms’ SSRIs have been rewiring our and increasingly younger kids’ brains - without good understanding of how, and without explicitly consent. We are a species that loves to change our default wiring, and it’s inevitable that we try to use AI and other new technology for the same purpose, just with a bit more intentionality and control. Not saying it’s going to always be great, just pointing out that we’ve been doing it from day one.
43
68
11
660
18,411
50
2,821
213
19,050
656,005
🚨 Si después de ver esto sigues creyendo que un beso es "solo un beso"... te han vendido una mentira. La mayoría jamás ha visto lo que realmente ocurre. Y cuando lo ve... no vuelve a pensar igual. 👇
259
438
44
5,799
5,176,933