Noël Michel ⏹️ (@Vert_Noel) header
Noël Michel ⏹️ (@Vert_Noel) profile picture

Nothing human makes it out of the near future. On discute ici pour se distraire, mais la seule vraie lutte qui compte est contre la superintelligence qui vient

Joined July 2010
Photo by Noël Michel ⏹️ (@Vert_Noel)Photo by Noël Michel ⏹️ (@Vert_Noel)Photo by Noël Michel ⏹️ (@Vert_Noel)Photo by Noël Michel ⏹️ (@Vert_Noel)Photo by Noël Michel ⏹️ (@Vert_Noel)Photo by Noël Michel ⏹️ (@Vert_Noel)Photo by Noël Michel ⏹️ (@Vert_Noel)Photo by Noël Michel ⏹️ (@Vert_Noel)
No need to worry about most things. If you're not working on making that safe, then it's time to chillmaxx and enjoy time with family.
Replying to @jam3scampbell
it seems mostly just a function of ppl not looking up at the world. they only interact with those physically in their bubble and don't listen to the batshit crazy things that top-cited scientists will say on podcasts
2
1
17
9,735
Noël Michel ⏹️ retweeted
AI devs are to become the most hated group in all of human history for what they are doing & have already done.
It might be hard to hold them accountable, but make no mistake: By knowingly risking all of our lives, AI companies are committing one of the greatest crimes in human history.
4
66
1,095
Noël Michel ⏹️ retweeted
PERSONNE N'EST CAPABLE DE DIRE A MELENCHON QU'ODILE TOUCHERAIT 4000€ / MOIS SI ELLE AVAIT INVESTI SES COTISATIONS SUR LE CAC 40 PUTAIN DE MERDE MEME APRES 50 CRISES DES SUBPRIMES CA RESTERAIT MIEUX QUE LA REPARTITION ILS SONT TOUS NULS NULS NULS NULS NULS
102
114
26
1,997
224,268
Noël Michel ⏹️ retweeted
C'est rigolo à quel point le monde associatif se considère (à raison) comme faisant partie de l'appareil d'état social, au point d'être choqué quand on lui fait remarquer ses conflits d'intérêts (chose normale pour le secteur libre, aussi appelé privé).
Depuis la commission de Bolloré contre France TV, ils doivent préciser qui couche avec qui. Et Anne-Cécile Mailfert, militante féministe intervenue à propos de Depardieu, n’apprécie pas qu’on la renvoie à son compagnon qui bosse sur France TV. On entend même un cri à la fin.
Readers added context they thought people might want to know
Aucun rapport avec Bolloré, c'était demandé par le SNJ de France TV : x.com/VLautard/statu…
7
33
234
4,514
Noël Michel ⏹️ retweeted
Didn't check X for 4 days while on vacation with my GF and it made me realize id rather experience the singularity as a normal person rather than an insider. I run an IRL Ai safety citizens group where I help people organize to book meetings with thier representatives. I run a YouTube channel where i often talk about x-risk to hundreds of thousands of people. I've spent a decent amount of time in SF and ive spoken to lab employees and safety researchers. I've been thinking about this since 2017 As an uneducated lower-middleclass yokel from Nova Scotia I've been uniquely lucky (or unlucky) to become quite close to AI as an industry, like top 1% close. Never saw this coming but here we are. In a way this is exciting, but it also feels like im a stage 2 cancer patient googling his symptoms every evening. I often feel guilty about daydreaming or planning for my future, I often feel guilty about trying to warn people, I don't even know what im warning about anymore. I will always advocate for a pause and will use what social capital ive accrued to advocate for a pause until the very end, short of that, im done caring about this. I dont care about the METR reports, I don't care about the forecasting, this knowledge does not help me. If we all die, reading the news the year before hand was useless. If we dont all die, it'll take a decade or more for the average person to adapt to our new world. Regardless, im sick of dooming. I'm sick of asking myself if this is my last Christmas, im tired of checking X every morning to see if "todays the day". I want to experience the singularity like a normal person. I want to check my news app and go "dog thats crazy" and then continue on with my day. Unless your job depends on it, I suggest you do the same. Maybe we're in our last 6 months, maybe true ASI is impossible, who gives a fuck? Grab a beer with your boys tonight, call your mom. What else is there to do?
11
6
4
155
4,433
Noël Michel ⏹️ retweeted
Altman: I don't trust all these other AI CEOs. Musk: I don't trust all these other AI CEOs. Amodei: I don't trust all these other AI CEOs. Zuck: I don't trust all these other AI CEOs. Nate: I agree with all of them. I just trust 1 less CEO than they do.
Why do AI labs keep racing to ASI if it's so dangerous?
4
39
792
Noël Michel ⏹️ retweeted
Le seul enseignement que je pourrais avoir retenu des punitions collectives c'est qu'il faut vraiment la jouer perso et n'en avoir rien à foutre. Au début du collège (je devais avoir 11 ou 12 ans) on avait pris une punition collective dans la classe parce que quelques uns bavardaient. Il fallait recopier dix fois un texte de biologie pénible. Je m'étais gâché une heure du week-end à faire ça, pour découvrir en sortie du week-end que la plupart avaient recopié seulement une ou deux fois et que les principaux responsables de la punition ne l'avaient simplement pas faite. Il n'y a eu aucune conséquence pour eux, et en définitive le plus puni a donc été le parfait innocent que j'étais.
Unpopop : les punitions collectives sont un très bon apprentissage de la vie en société. Dans la vie, on subit tous des "punitions collectives" pour des manquements dont on est pas responsable. Un des exemples est la fiscalité mais il y a aussi tous les dispositifs de contrôle.
14
11
181
11,287
Noël Michel ⏹️ retweeted
Yes, at the end of the day, the AI companies just really want to roll the dice. They don't want democratic oversight and they don't want to "pace the frontier". They want to gamble with our lives and nothing is too fast or too reckless for them to accept.
Replying to @KelseyTuoc
If that were true, then you're stuck convincing the AI companies to 'pace' on their own. But I think this is a terrible plan. The companies are simply not going to give embedded auditors full employee level access that isn't implicitly premised on playing nice.
5
43
563
Noël Michel ⏹️ retweeted
🔴 Sur TikTok, des jeunes interpellés à Paris se filment à l’intérieur du véhicule de police qui les transporte. Ils provoquent ouvertement les policiers et rigolent de la situation.
152
255
34
626
113,178
Noël Michel ⏹️ retweeted
Je ne sais pas si vous réalisez, mais l'État a dépouillé de ~450M€ le CNRS en 2 ans, en piochant dans sa trésorerie, pour payer des cotisations retraites et exige maintenant une restructuration budgétaire piloté par Bercy dû à son l'absence de trésorerie qu'il a causé 🤡
10
29
3
214
4,822
You have 11 months left.
in any case, its literally fucking over
24
39
8
552
39,136
Noël Michel ⏹️ retweeted
The window of Pangram having a ~0% false positive rate and the public simultaneously thinking that no useful AI detectors exist is probably closing soon and you can and should use it to your advantage
The college admissions essay is in a very weird place. It's impossible to say just how many of them are now written by AI, but at this point, experts agree that it's a lot of them. And while colleges generally frown upon that practice, they're using AI to evaluate the essays, too. cnn.com/2026/11/15/tech/ai-s…
24
51
4
2,797
130,677
"an anthropic employee organizes cnc parties" is that the best you got? multiple anthropic employees are barking at their owners as we speak
12
22
2
786
17,133
Noël Michel ⏹️ retweeted
You are entering 2027. Beyond this point, some things will not work the way you remember. Progress made after this point cannot be undone. It is recommended that you finish anything you wish to complete before proceeding. Continue? > Yes Yes
2027 has Q4 of humanity vibes
2
12
225
3,387
Noël Michel ⏹️ retweeted
Garrido: "La méthode Némésis. Aller provoquer tout le monde, se faire casser la gueule et ensuite pleurer." C'est comme ça qu'elle parle de la mort de Quentin, la Maritorne.
🔴🇫🇷 SÉQUENCE - Charles Consigny fait bégayer Raquel Garrido en direct sur les blacks blocs, elle est dans le déni total.
4
15
96
2,259
Noël Michel ⏹️ retweeted
Du parc locatif privé. 50% du parc locatif est détenu par l’État. Pour répondre à ta question : non, ça ne me semble pas juste, l’État concentre trop de logements, ce qui entraine corruption et mauvaise allocation du patrimoine.
3,5% de la population qui détient 50 (CINQUANTE, CINCUENTA, FIFTY) pourcent du parc locatif ça te semble juste peut-être ?
2
20
177
3,314
Noël Michel ⏹️ retweeted
My article about how a rogue Wikipedia editor misinformed the public by crediting Muslims with tons of inventions they had no part in was just syndicated by Pirate Wires! Go check it out: piratewires.com/p/the-invent…
77
1,542
68
7,845
114,582
Noël Michel ⏹️ retweeted
This feels like 2019. Terminally online ppl saw china freaking out and had maybe a week to prep while normies were oblivious Then they cancelled that NBA game. Days later everything shut down
55
134
13
3,771
292,939
Noël Michel ⏹️ retweeted
pov: a true intelligence explosion will eat your petty nationalisms for breakfast
.@paulfchristiano thinks that on the current trajectory (absent any slowing down), there is a 1 in 3 chance that we could have superintelligence powerful enough to overthrow the government within a year. (He articulates a concrete scenario). He argues that the labs should pause further algorithmic progress until they can secure their weights and research secrets. Otherwise China could simply steal the superintelligence in the middle of the intelligence explosion, and the labs would no longer be able to slow down even if they saw their models going off the rails. He also explains how AIs will be trained during the intelligence explosion, and why even a system trained entirely inside a data center could still have the capabilities profile such that humans would have very little chance of competing with it or controlling it. And he discusses the plan for alignment - basically, making a superintelligence that actually does what we ask the way we mean it. Fixing reward hacks and building better graders will help up till the next few model generations, but he thinks that no single current approach has better than even odds of letting us control superintelligence. The hope is that the AIs figure out the solution to this problem during the intelligence explosion. Paul is head of the Alignment Research Center. Previously, he was head of AI safety at the US AI Safety Institute. Before that, he led alignment at OpenAI, where he invented RLHF. He’s recently joined OpenAI’s board (though here he is speaking as a private citizen).
6
19
2
271
14,318
Noël Michel ⏹️ retweeted
it's funny how charlie hebdo published that muhammad comic, islamic terrorists attacked them (killing 12), a few newspapers published the comic in support, je suis charlie etc, and then basically nobody ever did that again. so i guess they won?
33
41
3
1,726
33,262
Noël Michel ⏹️ retweeted
I genuinely wish Hanson and others complaining about AI safety folks crying wolf prematurely had better forecasting records on AI capabilities. Disagreements on AI risk tend to be disagreements on AI capabilities, and to my chagrin, safety folks have the better record so far.
All these folks loudly crying wolf re AI are going to seem less credible in 5 years when not so much has happened by then. Making it harder to cry wolf at that point, if such cries are needed. There's an opportunity cost to crying wolf; best to wait til tis really needed.
2
19
313
6,448