I’ve been a bit quiet on X recently. The past year has been a transformational experience. Grok-4 and Kimi K2 are awesome, but the world of robotics is a wondrous wild west. It feels like NLP in 2018 when GPT-1 was published, along with BERT and a thousand other flowers that https://x.com/DrJimFan/status/1944443447953498285
A humanoid robot policy trained solely on synthetic data generated by a world model. Research Scientist Joel Jang presents NVIDIA’s DreamGen pipeline: ⦿ Post-train the world model Cosmos-Predict2 with a small set of real teleoperation demos. ⦿ Prompt the world model to https://x.com/TheHumanoidHub/status/1945569893048619201
What if you could ask a chatbot a question the size of an entire encyclopedia—and get an answer in real time? Multi-million token queries with 32x more users are now possible with Helix Parallelism, an innovation by #NVIDIAResearch that drives inference at huge scale. 🔗 https://x.com/NVIDIAAIDev/status/1942389449498787920
Trump unveils $90 billion in energy and AI investments for Pennsylvania during summit in Pittsburgh – CBS Pittsburgh https://www.cbsnews.com/pittsburgh/news/trump-energy-ai-summit-pittsburgh-carnegie-mellon/
Google to invest $25 billion in data centers, AI infrastructure in PJM https://www.cnbc.com/2025/07/15/google-to-invest-25-billion-in-data-centers-ai-infrastructure-in-pjm.html
CoreWeave commits $6 billion to Pennsylvania data center amid Trump AI push | Reuters https://www.reuters.com/business/coreweave-commits-6-billion-ai-data-center-pennsylvania-2025-07-15/
Oracle to invest $3 billion in AI, cloud expansion in Germany, Netherlands | Reuters https://www.reuters.com/business/oracle-invest-3-billion-ai-cloud-infrastructure-germany-netherlands-2025-07-15/
🤝 Nvidia’s CEO Jensen Huang is walking a tightrope in Beijing, balancing US-China tech rivalry while keeping Nvidia at the heart of the AI revolution. The stakes? Trillions and global influence. https://x.com/fdaudens/status/1945537196884123923
Can Nvidia convince governments to pay for “sovereign AI”? Politicians are warming to the idea of national AI systems, but it might not reduce dependence on US tech. 🌍 https://x.com/fdaudens/status/1944759771212468733
Can Nvidia persuade governments to pay for “sovereign” AI? https://www.economist.com/business/2025/07/13/can-nvidia-persuade-governments-to-pay-for-sovereign-ai
Nvidia C.E.O. Treads Carefully in Beijing – The New York Times https://www.nytimes.com/2025/07/16/business/nvidia-jensen-huang-beijing.html
Nvidia CEO Jensen Huang’s rosy AI vision: “”There will be more jobs”” https://www.axios.com/2025/07/14/ai-jobs-nvidia-jensen-huang-dario-amodei
Nvidia just got the OK to sell AI chips to China after its CEO met Trump. Tech, trade, and geopolitics: all on the table. https://x.com/fdaudens/status/1945121369584234947
Nvidia says it will restart sales of a key AI chip to China, in a reversal of US restrictions | CNN Business https://www.cnn.com/2025/07/15/business/nvidia-resume-h20-chip-sales-to-china-intl-hnk
Google and Brookfield strike $3bn hydro power deal https://www.ft.com/content/d8bef8a3-5988-4080-ad7d-61bc9885e6ba
Google just inked a $3B deal for hydro power to run its AI data centers. Big Tech is scrambling for clean, reliable energy as AI’s appetite explodes. ⚡ https://x.com/fdaudens/status/1945121372465754471
Developing superintelligence is now in sight. We should act as if it’s going to be ready in the next 2-3 years.”” – Mark Zuckerberg About paying $100 million or $200 million pay packages, he argued that Meta will spend “hundreds of billions” on compute and data-center https://x.com/rohanpaul_ai/status/1945725129138597928
Meta to spend hundreds of billions on AI data centres, says Mark Zuckerberg https://www.bbc.com/news/articles/c1e02vx55wpo
Meta’s Zuckerberg pledges hundreds of billions for AI data centers in superintelligence push | Reuters https://www.reuters.com/business/zuckerberg-says-meta-will-invest-hundreds-billions-superintelligence-2025-07-14/
Our mission with the lab is to deliver personal superintelligence to everyone in the world. So that way, we can put that power in every individual’s hand.”” – Mark Watch Mark’s full interview with The Information as he goes deeper on Meta’s vision for superintelligence and”” / X https://x.com/AIatMeta/status/1945182467088113920
Today Mark announced Meta’s major AI compute investment. See his post: https://x.com/AIatMeta/status/1944783224288465165
Microsoft, US national lab tap AI to speed up nuclear power permitting process | Reuters https://www.reuters.com/business/energy/microsoft-us-national-lab-tap-ai-speed-up-nuclear-power-permitting-process-2025-07-16/
agree with lots of what jensen has been saying about ai and jobs; there is a ton of stuff to do in the world. people will 1) do a lot more than they could do before; ability and expectation will both go up 2) still care very much about other people and what they do 3) still be”” / X https://x.com/sama/status/1945541270438646270
Show HN: ArchGW – An intelligent edge and service proxy for agents | Hacker News https://news.ycombinator.com/item?id=44546265
Speaking of which: Nvidia CEO Jensen Huang disagrees with Anthropic CEO Dario Amodei on whether AI will create more jobs—or trigger a “white-collar apocalypse.” Huang believes AI will create vastly more, and better, jobs. ⚔️ https://x.com/fdaudens/status/1944759769803178334
New research from Exo done (in part) with MLX on Apple silicon: An algorithm for distributed training that leverages higher RAM capacity of Apple silicon relative to FLOPs and inter-machine bandwidth. https://x.com/awnihannun/status/1946270589729349645
Lets GOOO! @NVIDIAAIDev just dropped Canary Qwen 2.5 – SoTA on Open ASR Leaderboard, CC-BY licensed 🔥 > Works in both ASR and LLM mode (i.e. ask the model to summarise, QnA about the audio) > Achieves the lowest 5.62 WER > RTFx of 418 for a 2.5B model is impressive > https://x.com/reach_vb/status/1945869582520242502
RT @reach_vb: Lets GOOO! @NVIDIAAIDev just dropped Canary Qwen 2.5 – SoTA on Open ASR Leaderboard, CC-BY licensed 🔥 > Works in both ASR an…”” / X https://x.com/reach_vb/status/1946087224346313175
Google to pay $2.4 billion in deal to license tech of Windsurf, WSJ reports | Reuters https://www.reuters.com/business/google-pay-24-billion-deal-license-tech-windsurf-wsj-reports-2025-07-12/
Hardware independent LLM inference engine from ZML.”” / X https://x.com/ylecun/status/1946266201933349197
Manual Triton kernels demand tedious tuning and block everyday AI work. AUTOTRITON, a 8B model, learns to draft high performance Triton kernels from examples and live GPU tests. A custom pipeline collects PyTorch kernels, turns them into verified Triton pairs, and yields 14,102 https://x.com/rohanpaul_ai/status/1943950960587673971
Reposting for reach since this isn’t common knowledge. Don’t get tricked by GPU providers that offer you PCIe H100s! Pay the extra $0.30/hr for SXM H100s, they have 2x the power envelope and way faster interconnect!”” / X https://x.com/corbtt/status/1946268958589673662
RT @thdxr: holy shit groq + k2 is fast af – video not sped up stuff is changing real fast https://x.com/JonathanRoss321/status/1944988412357849128
Shipping WebGPU on Windows in Firefox 141 https://simonwillison.net/2025/Jul/16/webgpu-firefox/
Broadcom launches new Tomahawk Ultra networking chip in AI battle against Nvidia | Reuters https://www.reuters.com/business/broadcom-launches-new-tomahawk-ultra-networking-chip-ai-battle-against-nvidia-2025-07-15/
We rolled out a new inference engine built for NVIDIA Blackwell! DeepSeek R1 offers the best peak performance (386 TPS) across any service or silicon in production today for the full R1 model with impressive latency and throughput at higher batch sizes. You can try the”” / X https://x.com/vipulved/status/1945934641451675793
The paper argues that because the gains from extra compute keep shrinking, low‑budget models will soon rival today’s costly giants. Scaling laws show loss falls roughly as compute to the power -0.155, so doubling flops buys only a tiny loss cut. When hardware and algorithm https://x.com/rohanpaul_ai/status/1944356633351770559
Jensen Huang was asked about humanoid robots at Beijing Supply Chain Expo, what he said: ⦿ Timing for humanoid robot technology is right now ⦿ World is short tens of millions of workers, robots will grow world’s GDP a lot ⦿ China has unique combination of AI tech, https://x.com/TheHumanoidHub/status/1945600818100818202
