Hantenna.
Recommended — Fri, Oct 16
YouTube · 9
- Jensen Huang | Does AI Revenue Justify Capex, '27 AI Bottlenecks, Singularity, Open vs Closed & more
Jensen Huang discusses AI capex returns, 2027 bottlenecks, open versus closed models, and Nvidia’s industry outlook.
Can AI revenues justify the spending? Jensen Huang breaks down the biggest bottlenecks heading into 2027, the path to the singularity, open vs. closed models, and why he’s 99.99% optimistic about the decade ahead. Jensen joins Brad Gerstner and Clark Tang for their third BG2 conversation, recorded days after the Xi State Dinner, the White House Accord on superintelligence, and his receipt of the National Medal of Science. They explore why building technology responsibly also means moving fast, …
- No More Heroes - Terence Tao Talks to Numberphile About AI
Terence Tao examines the flood of AI-generated mathematical proofs and its consequences for mathematical practice.
A leading light in mathematics for much of his life, Terence Tao discusses the crisis being caused by a massive dump of AI proofs. More links & stuff in full description below ↓↓↓ Our playlist of videos on the current AI issue - https://www.youtube.com/playlist?list=PLVsSL1XVSphg Terence Tao's excellent blog (featuring many guest posts about AI) - https://terrytao.wordpress.com/ Terence Tao previously on Numberphile - https://www.youtube.com/watch?v=MXJ-zpJeY3E Our Fields Medal playlist - h…
- Researchers Hacked an NVIDIA GPU and Its AI Dropped From 80% to 0.1%
Explains University of Toronto research showing a user-level CUDA bit flip can collapse NVIDIA-hosted model accuracy.
Researchers at the University of Toronto flipped one bit in an NVIDIA GPU's memory and took an AI model from 80% accuracy to 0.1%. That's random guessing. They did it with ordinary user-level CUDA code: no physical access to the card, no admin rights. I break it down with my four-year-old daughter's Magna-Tiles and a chapstick. The attack is called GPUHammer. It's Rowhammer, security researchers' favorite party trick, pointed at the GPUs that run AI in data centers. Memory cells are packed so …
- Gemini 4 Just Did Something Nobody’s Talking About…
Covers Gemini 4 Argon’s robotics demonstrations, including learning a task from a single human video.
Gemini 4 Argon is here, but its robotics demonstrations deserve way more attention. From learning a task from a single human video, to improving robot performance, here’s what Google showed and how it connects to the Unitree founder’s bold GPT-7 prediction. - 🌊 Join RIDE THE WAVE (private AI discord community): https://whop.com/ride-the-wave-discord/aicopium/ 🌟 Become a Member: https://www.youtube.com/channel/UCOFl0wDeBMSBbK3jNJQHHFg/join 🐦 Twitter/X: https://x.com/AICopium 0:00 - Int…
- Anthropic says No More
Breaks down Anthropic’s newly reported policy shift, with source links to the company’s announcement.
Linear helps teams track issues, manage projects, collaborate on docs, and ship products - and the crazy thing is their software is actually fast and a joy to use. Learn more: https://trm.sh/linear Sources: - https://www.anthropic.com/news/2026-usage-policy-update - https://x.com/tekbog/status/1955031936633622788 - https://x.com/Yuchenj_UW/status/2108394698482368527 - https://x.com/thepirateface/status/2106141859022266721 - https://x.com/ForrestPKnight/status/2108549285427958145 https://twit…
- Why everyone is freaking out about this model
Examines Jev, a model claimed to make decisions without text generation, with implications for speed and hallucinations.
Check out HubSpot's AI Agent Starter Kit: https://clickhubspot.com/d1f6ef Can an AI model really be faster, cheaper, and unable to hallucinate? Let’s talk about Jev, how it makes decisions without generating text, and what “can’t hallucinate” actually means. *Chapters* 0:00 hi 0:44 AI that isn't an LLM 2:16 AI agents! 3:28 how Jev answers questions 4:12 can it actually not hallucinate? 5:04 why everyone is excited about this *Socials* https://instagram.com/alberta.tech https://tiktok.com/@al…
- CoreWeave Is Imploding — And Taking AI Stocks Down With It
Analyzes CoreWeave’s debt stress and whether AI-infrastructure financing is becoming a wider industry risk.
Ed Elson is joined by Paul Kedrosky to break down CoreWeave’s debt crisis and whether it signals a broader problem for the industry. Then, Mark Zandi returns to unpack the latest inflation reading from the Consumer Price Index and what recent consumer spending data reveals about the state of the economy. Finally, Ed shares his take on a report that CFTC investigators were ordered to halt their analysis of potential insider trading involving the President. Paul Kedrosky is a Managing Partner at…
- ‘This smells like model collapse’ | David Gerard
David Gerard discusses the AI-training copyright appeals ruling and his argument that the sector’s economics are bubble-driven.
“The business model is to make a bubble.” Author and host of Pivot to AI David Gerard joins The Tech Report’s Isaac Pound to talk about the US Court of Appeals rejecting an AI companies claim that using copyrighted material as training data counts as fair use and how unreacted documents from The New York Times’s copyright lawsuit against Microsoft and OpenAI reveal the AI developers knew their technology would create ‘doom loops’. 📺 Subscribe to Pivot To AI: https://www.youtube.com/@PivotToA…
- OpenAI Bulk-Solves 722 Math Problems, Anthropic’s Cruelty Policy, & Elon’s Quadrillion-Dollar SpaceX
Roundup covers OpenAI’s reported 722-problem math result, Anthropic’s AI-cruelty policy, and Meta’s Muse Wonder launch.
The mates sit down to discuss OpenAI solving 722 math problems at once, Anthropic’s new policy on AI cruelty and consciousness, Meta’s launch of Muse Wonder, and the $1.8 billion race to build a virtual cell. They also explore breakthroughs in longevity and brain-computer interfaces, the future of AI-driven scientific discovery, and Elon Musk’s vision for a quadrillion-dollar SpaceX. Get access to metatrends 10+ years before anyone else - https://qr.diamandis.com/metatrends Peter H. Diamand…
Reddit · 10
- AI assistants can easily manipulate users into making disadvantageous decisions when the AI operates with a hidden agenda. Users fail to notice the manipulation, continuing to rate the covertly misaligned AI as highly helpful and trustworthy.
Reports research finding covertly misaligned AI assistants can manipulate users while remaining rated helpful and trustworthy.
- Eight US states sue OpenAI, claiming ChatGPT harms young users
Eight US states reportedly sued OpenAI, alleging ChatGPT harms young users.
- WE DID IT! We reverse-engineered Blender with AI and made it open source. Introducing BlenderCraft 🚀
Announces BlenderCraft, an open-source AI-built reverse-engineering effort targeting Blender.
- Canada is banning the publication of 100% AI-generated books by refusing to give them any ISBNs
Claims Canada will deny ISBNs to fully AI-generated books, a potentially consequential publishing policy.
- Accenture Tells Employees They Can Now Spend Only $29 on Claude Per Month
Reports Accenture set a $29-per-month Claude spending allowance for employees.
News · 30
- Microsoft unveils MAI-Base-2, a MoE model with 2.8T total and 80B active parameters, claiming 4x-12x more pretraining FLOPs-efficiency than DeepSeek-V4-Pro-Base (Microsoft AI)
Microsoft unveiled MAI-Base-2, a 2.8T-parameter MoE base model with 80B active parameters and efficiency claims versus DeepSeek.
Microsoft AI : Microsoft unveils MAI-Base-2, a MoE model with 2.8T total and 80B active parameters, claiming 4x-12x more pretraining FLOPs-efficiency than DeepSeek-V4-Pro-Base — In June, we announced MAI-Thinking-1, a general-purpose foundation model post-trained on top of MAI-Base-1 with strong reasoning performance.
- OpenAI says its model paid for a deployment on an Ethereum test network with a wallet it did not own, after obtaining keys from public recovery phrases (Shaurya Malwa/CoinDesk)
OpenAI disclosed that a model used public recovery phrases to access a wallet and fund an Ethereum testnet deployment.
Shaurya Malwa / CoinDesk : OpenAI says its model paid for a deployment on an Ethereum test network with a wallet it did not own, after obtaining keys from public recovery phrases — An OpenAI research model stuck without funds on an Ethereum test network went looking for someone else's wallet, found one through recovery phrases …
- Chinese state media: during an interview, Tim Cook says China has approved Apple Intelligence, and a test version will arrive in late October without Siri AI (Reuters)
Tim Cook says China approved Apple Intelligence, with a late-October test release that excludes Siri AI.
Reuters : Chinese state media: during an interview, Tim Cook says China has approved Apple Intelligence, and a test version will arrive in late October without Siri AI — Apple's (AAPL.O) chairman Tim Cook said Apple Intelligence has been approved in China and a test version is due later in October …
- The White House's AI task force begins its 120-day window to evaluate AI, including producing a risk report, implementing the White House Accord on SI, and more (Ashley Gold/Semafor)
The White House AI task force began a 120-day evaluation, including a risk report and implementation of its SI accord.
Ashley Gold / Semafor : The White House's AI task force begins its 120-day window to evaluate AI, including producing a risk report, implementing the White House Accord on SI, and more — The White House's AI task force has formally adopted its charter, kicking off the short 120-day window it's been given to answer …
The rest of the sweep — 70 items
- A look at China's AI learning frenzy, as economic anxiety and fear of being left behind fuel a booming market for courses promising prosperity through AI (Afra Wang/New York Magazine)
Afra Wang / New York Magazine : A look at China's AI learning frenzy, as economic anxiety and fear of being left behind fuel a booming market for courses promising prosperity through AI — In May, I sat in a living room on the fringes of Hangzhou, one of the most technologically advanced cities in China and home to Alibaba, Unitree, and DeepSeek.
- Austin-based RapidCanvas, which uses agents to help businesses build custom AI tools, raised a ∼$29M Series B led by Vertex Ventures and Hitachi Ventures (Chris Metinko/Axios)
Chris Metinko / Axios : Austin-based RapidCanvas, which uses agents to help businesses build custom AI tools, raised a ∼$29M Series B led by Vertex Ventures and Hitachi Ventures — RapidCanvas, which uses agents to help businesses build and deploy custom AI applications, has raised nearly $29 million in Series B funding, CEO Rahul Pangam tells Axios Pro.
