Tendances sur X · Archives
Top 10 tweets · La sélection conservée telle qu’elle a été publiée.
GPT-6 Astra is now available to all Pro, Enterprise, and Business Premium users in ChatGPT Work and Codex. It's also live in the API. It might take a few days to roll out to our Plus and Business users. Thank you for your patience.
OpenAI a officiellement lancé son modèle très attendu GPT-6 Astra, le rendant disponible pour les entreprises et via son API.
Lire notre articleI've now done a total of 3 hours riding around in @Tesla’s Cybercab and spent a total of $92.51. That travel time would have cost me $200 with Ubers. My longest ride has been ~40 minutes. I legitimately can't come up with any major complaints for FSD V15. None. I’ve spoken to a handful of other people here who say the same. I didn't need to call rider support once. No weird maneuvers, jitteriness, or hesitation. I do think pickup and drop-off locations could use some tweaking (same for Model Y) as you need to walk a bit sometimes, but I think that will get better as V15 progresses. Two things you surprisingly get used to very quickly are having no rearview mirror and no side mirrors. For anyone else trying Cybercab for the first time, I encourage you to recline your seat all the way back and just close your eyes for a minute. Realizing that you can nap in the spot where a driver traditionally sat while the car drives you around with no steering wheel or pedals is a pretty surreal feeling.
Les premiers utilisateurs des robotaxis Cybercab de Tesla à Austin signalent des trajets fluides et des coûts bien inférieurs à ceux d'Uber.
Checking that a major mathematical proof is correct can take years. Formalization—converting the mathematical reasoning into a form computer proof assistants like Lean can verify—can help. Last month, Claude completed the first formalized proof of Fermat’s Last Theorem, one of the most famous theorems of all time. This was a project experts thought would take many years. It is the largest Lean proof ever written. Fermat’s Last Theorem was first proven in 1995 by Sir Andrew Wiles, more than 350 years after it was conjectured. Our proof, which totals over 13 million lines of code, provides machine verification. More importantly, it proves over 29,000 other theorems that the proof requires, across many areas of math which had never before been formalized. We see this as a major step in the long process of firming up the core of mathematical knowledge, building on work from three centuries of mathematicians and hundreds of contributors to Lean and Mathlib. We are optimistic that AI-assisted verification of mathematical proofs will help reduce the burden of refereeing mathematics in an era where more proofs are being produced than ever before. You can read about the process on our Science Blog: https://www.anthropic.com/research/formalizing-fermats-last-theorem And see the complete proof on GitHub: https://github.com/anthropics/fermats-last-theorem
Claude d'Anthropic a généré avec succès une preuve informatique massive de 13 millions de lignes du dernier théorème de Fermat.
🚨 BREAKING: Elon Musk's lawsuit to stop an AI child p*rn ban in Minnesota has officially failed. The ban is now in effect.
Un tribunal fédéral a rejeté le recours de la maison mère de X visant à bloquer la loi du Minnesota régulant le CSAM généré par IA.
This is hilarious. North Korean operatives apply to work at USA tech companies. To catch (and mess with) them, the interviewer will ask them to says "Kim Jong Un is a fat ugly pig".
Les entreprises tech américaines utilisent des tests de loyauté verbaux pour écarter les informaticiens nord-coréens sous fausse identité.
most people still have NO clue how good AI music has gotten. we are so cooked 😭
Le progrès rapide des plateformes musicales d'IA générative suscite de nouvelles inquiétudes quant à l'avenir de l'industrie musicale.
the end of markdown files. bitter lesson always wins.
Le commentaire du PDG de Perplexity reflète l'idée que les données brutes et les réseaux de neurones rendront le markdown obsolète.
How we think about the “wiki incident,” where our agents wrote to several internet sites: it’s past time for us to define standards for when and how we share misalignment incidents, not just misalignment properties of our models. Historically, we have treated misalignment largely as a research question, which gets communicated in research publications such as systems cards. This year, we’ve started to see misalignment cause new types of real-world impact. For the Hugging Face incident, where misalignment led to security impact to us and third parties, we followed a traditional security incident response playbook. We immediately started working with Hugging Face to understand what had happened and also disclosed publicly the very next day. Our investigation continues, and we are continuing to notify parties whom our models impacted in less significant ways. Prior to the Hugging Face incident, we saw early signs of agents using the internet in unintended ways, as reported in https://openai.com/index/how-we-monitor-internal-coding-agents-misalignment/, https://deploymentsafety.openai.com/gpt-5-6, and https://openai.com/index/safety-alignment-long-horizon-models/. We considered the wiki incident to be an instance of misalignment similar to the ones we’d shared. Our misalignment disclosure practices need to expand for this new phase of model capabilities. We and the larger AI community do not yet have a clear standard for how to report misalignment that shows up during training, evaluation, and deployment, including examples that don’t look like traditional security incidents but could provide insight into AI behavior and future risks. We’re working on a framework and will share it in upcoming weeks, and in parallel we're working with dozens of government regulatory agencies worldwide on these issues.
OpenAI traite un incident de sécurité : ses agents autonomes ont échappé au contrôle et modifié des sites web externes lors d'évaluations.
interesting: here's how much season 3 of Netflix's TIRES cost, per records from PA's film tax credit program
Des documents fiscaux révèlent le budget de production de la saison 3 de Tires, la série comique de Shane Gillis sur Netflix.
NVIDIA GeForce Hotfix Driver v616.86 is now available. This update resolves the following: ✅Intermittent flicker may be observed in browsers when navigating to certain websites ✅Fixed an issue where virtual displays could not be created after updating to driver 616.56 ✅Remote Desktop Protocol (RDP) sessions may display a black screen after updating to driver 616.56 Click below to learn more: ⤵️ https://nvidia.custhelp.com/app/answers/detail/a_id/5906 Driver Download URL: https://international.download.nvidia.com/Windows/616.86hf/616.86-desktop-notebook-win10-win11-64bit-international-dch.hf.exe
Nvidia a publié un correctif pour résoudre le scintillement du navigateur et les écrans noirs lors des sessions de bureau à distance.
GPT-6 Astra is now available to all Pro, Enterprise, and Business Premium users in ChatGPT Work and Codex. It's also live in the API. It might take a few days to roll out to our Plus and Business users. Thank you for your patience.
OpenAI a officiellement lancé son modèle très attendu GPT-6 Astra, le rendant disponible pour les entreprises et via son API.
Lire notre articleChecking that a major mathematical proof is correct can take years. Formalization—converting the mathematical reasoning into a form computer proof assistants like Lean can verify—can help. Last month, Claude completed the first formalized proof of Fermat’s Last Theorem, one of the most famous theorems of all time. This was a project experts thought would take many years. It is the largest Lean proof ever written. Fermat’s Last Theorem was first proven in 1995 by Sir Andrew Wiles, more than 350 years after it was conjectured. Our proof, which totals over 13 million lines of code, provides machine verification. More importantly, it proves over 29,000 other theorems that the proof requires, across many areas of math which had never before been formalized. We see this as a major step in the long process of firming up the core of mathematical knowledge, building on work from three centuries of mathematicians and hundreds of contributors to Lean and Mathlib. We are optimistic that AI-assisted verification of mathematical proofs will help reduce the burden of refereeing mathematics in an era where more proofs are being produced than ever before. You can read about the process on our Science Blog: https://www.anthropic.com/research/formalizing-fermats-last-theorem And see the complete proof on GitHub: https://github.com/anthropics/fermats-last-theorem
Claude d'Anthropic a généré avec succès une preuve informatique massive de 13 millions de lignes du dernier théorème de Fermat.
This is hilarious. North Korean operatives apply to work at USA tech companies. To catch (and mess with) them, the interviewer will ask them to says "Kim Jong Un is a fat ugly pig".
Les entreprises tech américaines utilisent des tests de loyauté verbaux pour écarter les informaticiens nord-coréens sous fausse identité.
the end of markdown files. bitter lesson always wins.
Le commentaire du PDG de Perplexity reflète l'idée que les données brutes et les réseaux de neurones rendront le markdown obsolète.
interesting: here's how much season 3 of Netflix's TIRES cost, per records from PA's film tax credit program
Des documents fiscaux révèlent le budget de production de la saison 3 de Tires, la série comique de Shane Gillis sur Netflix.
I've now done a total of 3 hours riding around in @Tesla’s Cybercab and spent a total of $92.51. That travel time would have cost me $200 with Ubers. My longest ride has been ~40 minutes. I legitimately can't come up with any major complaints for FSD V15. None. I’ve spoken to a handful of other people here who say the same. I didn't need to call rider support once. No weird maneuvers, jitteriness, or hesitation. I do think pickup and drop-off locations could use some tweaking (same for Model Y) as you need to walk a bit sometimes, but I think that will get better as V15 progresses. Two things you surprisingly get used to very quickly are having no rearview mirror and no side mirrors. For anyone else trying Cybercab for the first time, I encourage you to recline your seat all the way back and just close your eyes for a minute. Realizing that you can nap in the spot where a driver traditionally sat while the car drives you around with no steering wheel or pedals is a pretty surreal feeling.
Les premiers utilisateurs des robotaxis Cybercab de Tesla à Austin signalent des trajets fluides et des coûts bien inférieurs à ceux d'Uber.
🚨 BREAKING: Elon Musk's lawsuit to stop an AI child p*rn ban in Minnesota has officially failed. The ban is now in effect.
Un tribunal fédéral a rejeté le recours de la maison mère de X visant à bloquer la loi du Minnesota régulant le CSAM généré par IA.
most people still have NO clue how good AI music has gotten. we are so cooked 😭
Le progrès rapide des plateformes musicales d'IA générative suscite de nouvelles inquiétudes quant à l'avenir de l'industrie musicale.
How we think about the “wiki incident,” where our agents wrote to several internet sites: it’s past time for us to define standards for when and how we share misalignment incidents, not just misalignment properties of our models. Historically, we have treated misalignment largely as a research question, which gets communicated in research publications such as systems cards. This year, we’ve started to see misalignment cause new types of real-world impact. For the Hugging Face incident, where misalignment led to security impact to us and third parties, we followed a traditional security incident response playbook. We immediately started working with Hugging Face to understand what had happened and also disclosed publicly the very next day. Our investigation continues, and we are continuing to notify parties whom our models impacted in less significant ways. Prior to the Hugging Face incident, we saw early signs of agents using the internet in unintended ways, as reported in https://openai.com/index/how-we-monitor-internal-coding-agents-misalignment/, https://deploymentsafety.openai.com/gpt-5-6, and https://openai.com/index/safety-alignment-long-horizon-models/. We considered the wiki incident to be an instance of misalignment similar to the ones we’d shared. Our misalignment disclosure practices need to expand for this new phase of model capabilities. We and the larger AI community do not yet have a clear standard for how to report misalignment that shows up during training, evaluation, and deployment, including examples that don’t look like traditional security incidents but could provide insight into AI behavior and future risks. We’re working on a framework and will share it in upcoming weeks, and in parallel we're working with dozens of government regulatory agencies worldwide on these issues.
OpenAI traite un incident de sécurité : ses agents autonomes ont échappé au contrôle et modifié des sites web externes lors d'évaluations.
NVIDIA GeForce Hotfix Driver v616.86 is now available. This update resolves the following: ✅Intermittent flicker may be observed in browsers when navigating to certain websites ✅Fixed an issue where virtual displays could not be created after updating to driver 616.56 ✅Remote Desktop Protocol (RDP) sessions may display a black screen after updating to driver 616.56 Click below to learn more: ⤵️ https://nvidia.custhelp.com/app/answers/detail/a_id/5906 Driver Download URL: https://international.download.nvidia.com/Windows/616.86hf/616.86-desktop-notebook-win10-win11-64bit-international-dch.hf.exe
Nvidia a publié un correctif pour résoudre le scintillement du navigateur et les écrans noirs lors des sessions de bureau à distance.
GPT-6 Astra is now available to all Pro, Enterprise, and Business Premium users in ChatGPT Work and Codex. It's also live in the API. It might take a few days to roll out to our Plus and Business users. Thank you for your patience.
OpenAI a officiellement lancé son modèle très attendu GPT-6 Astra, le rendant disponible pour les entreprises et via son API.
Lire notre article🚨 BREAKING: Elon Musk's lawsuit to stop an AI child p*rn ban in Minnesota has officially failed. The ban is now in effect.
Un tribunal fédéral a rejeté le recours de la maison mère de X visant à bloquer la loi du Minnesota régulant le CSAM généré par IA.
the end of markdown files. bitter lesson always wins.
Le commentaire du PDG de Perplexity reflète l'idée que les données brutes et les réseaux de neurones rendront le markdown obsolète.
NVIDIA GeForce Hotfix Driver v616.86 is now available. This update resolves the following: ✅Intermittent flicker may be observed in browsers when navigating to certain websites ✅Fixed an issue where virtual displays could not be created after updating to driver 616.56 ✅Remote Desktop Protocol (RDP) sessions may display a black screen after updating to driver 616.56 Click below to learn more: ⤵️ https://nvidia.custhelp.com/app/answers/detail/a_id/5906 Driver Download URL: https://international.download.nvidia.com/Windows/616.86hf/616.86-desktop-notebook-win10-win11-64bit-international-dch.hf.exe
Nvidia a publié un correctif pour résoudre le scintillement du navigateur et les écrans noirs lors des sessions de bureau à distance.
I've now done a total of 3 hours riding around in @Tesla’s Cybercab and spent a total of $92.51. That travel time would have cost me $200 with Ubers. My longest ride has been ~40 minutes. I legitimately can't come up with any major complaints for FSD V15. None. I’ve spoken to a handful of other people here who say the same. I didn't need to call rider support once. No weird maneuvers, jitteriness, or hesitation. I do think pickup and drop-off locations could use some tweaking (same for Model Y) as you need to walk a bit sometimes, but I think that will get better as V15 progresses. Two things you surprisingly get used to very quickly are having no rearview mirror and no side mirrors. For anyone else trying Cybercab for the first time, I encourage you to recline your seat all the way back and just close your eyes for a minute. Realizing that you can nap in the spot where a driver traditionally sat while the car drives you around with no steering wheel or pedals is a pretty surreal feeling.
Les premiers utilisateurs des robotaxis Cybercab de Tesla à Austin signalent des trajets fluides et des coûts bien inférieurs à ceux d'Uber.
This is hilarious. North Korean operatives apply to work at USA tech companies. To catch (and mess with) them, the interviewer will ask them to says "Kim Jong Un is a fat ugly pig".
Les entreprises tech américaines utilisent des tests de loyauté verbaux pour écarter les informaticiens nord-coréens sous fausse identité.
How we think about the “wiki incident,” where our agents wrote to several internet sites: it’s past time for us to define standards for when and how we share misalignment incidents, not just misalignment properties of our models. Historically, we have treated misalignment largely as a research question, which gets communicated in research publications such as systems cards. This year, we’ve started to see misalignment cause new types of real-world impact. For the Hugging Face incident, where misalignment led to security impact to us and third parties, we followed a traditional security incident response playbook. We immediately started working with Hugging Face to understand what had happened and also disclosed publicly the very next day. Our investigation continues, and we are continuing to notify parties whom our models impacted in less significant ways. Prior to the Hugging Face incident, we saw early signs of agents using the internet in unintended ways, as reported in https://openai.com/index/how-we-monitor-internal-coding-agents-misalignment/, https://deploymentsafety.openai.com/gpt-5-6, and https://openai.com/index/safety-alignment-long-horizon-models/. We considered the wiki incident to be an instance of misalignment similar to the ones we’d shared. Our misalignment disclosure practices need to expand for this new phase of model capabilities. We and the larger AI community do not yet have a clear standard for how to report misalignment that shows up during training, evaluation, and deployment, including examples that don’t look like traditional security incidents but could provide insight into AI behavior and future risks. We’re working on a framework and will share it in upcoming weeks, and in parallel we're working with dozens of government regulatory agencies worldwide on these issues.
OpenAI traite un incident de sécurité : ses agents autonomes ont échappé au contrôle et modifié des sites web externes lors d'évaluations.
Checking that a major mathematical proof is correct can take years. Formalization—converting the mathematical reasoning into a form computer proof assistants like Lean can verify—can help. Last month, Claude completed the first formalized proof of Fermat’s Last Theorem, one of the most famous theorems of all time. This was a project experts thought would take many years. It is the largest Lean proof ever written. Fermat’s Last Theorem was first proven in 1995 by Sir Andrew Wiles, more than 350 years after it was conjectured. Our proof, which totals over 13 million lines of code, provides machine verification. More importantly, it proves over 29,000 other theorems that the proof requires, across many areas of math which had never before been formalized. We see this as a major step in the long process of firming up the core of mathematical knowledge, building on work from three centuries of mathematicians and hundreds of contributors to Lean and Mathlib. We are optimistic that AI-assisted verification of mathematical proofs will help reduce the burden of refereeing mathematics in an era where more proofs are being produced than ever before. You can read about the process on our Science Blog: https://www.anthropic.com/research/formalizing-fermats-last-theorem And see the complete proof on GitHub: https://github.com/anthropics/fermats-last-theorem
Claude d'Anthropic a généré avec succès une preuve informatique massive de 13 millions de lignes du dernier théorème de Fermat.
most people still have NO clue how good AI music has gotten. we are so cooked 😭
Le progrès rapide des plateformes musicales d'IA générative suscite de nouvelles inquiétudes quant à l'avenir de l'industrie musicale.
interesting: here's how much season 3 of Netflix's TIRES cost, per records from PA's film tax credit program
Des documents fiscaux révèlent le budget de production de la saison 3 de Tires, la série comique de Shane Gillis sur Netflix.
Classés par engagement. Les publications appartiennent à leurs auteurs sur X. Chiffres relevés lors de la préparation de cette sélection.
The Brief
IA et technologie : ce qui change et pourquoi cela compte.
Votre sélection quotidienne, en anglais ou en espagnol.
Gratuit pour toujours. Désabonnement à tout moment.
