btw every single compute infra provider’s chart, including render competitors, is looking like this. something broke in Dec 2025 and everything is becoming computer. forget GPU shortage, forget Memory shortage, the @fabknowledge pod on LS was right, there is going to be a CPU
AI HARDWARE
-
AI Copilots Revolutionize Medicine with MedOS: ChatGPT, XR Glasses, and Robotic Hands for Doctors
By
–
AI copilots are coming for medicine.
— AI Breakfast (@AiBreakfast) 14 mars 2026
MedOS is basically ChatGPT + smart glasses + robotic hands for doctors.
A physician can wear XR glasses, see a patient, and the system:
• watches the procedure in real time
• reasons with multi-agent AI models
• suggests diagnoses or… pic.twitter.com/PbxnYVO227AI copilots are coming for medicine. MedOS is basically ChatGPT + smart glasses + robotic hands for doctors. A physician can wear XR glasses, see a patient, and the system: • watches the procedure in real time
• reasons with multi-agent AI models
• suggests diagnoses or -
Energy as the Ultimate Constraint for AI Development
By
–
Energy is the constraint. Unlock it, and everything else follows. https://t.co/kwThpiO939
— Nina Schick (@NinaDSchick) 14 mars 2026Energy is the constraint. Unlock it, and everything else follows.
-
British AI Companies Depend on US Compute Subsidies
By
–
Who is DeepMind owned by? There is certainly incredible AI IP in London – IE @synthesiaIO / @wayve_ai . All of these companies need compute in order to incubate and scale their visions. That means that British AI is supported by the ‘compute subsidies’ provided by U.S.
-
Tenstorrent builds massive AI computers with 600-2000 processors
By
–
72 is actually a small number @tenstorrent is building big computers. 600 AI processors now, 1000-2000 soon. We program it as one computer. @satyanadella nitter.net/satyanadella/status/20… Satya Nadella (@satyanadella) We’re the first cloud to bring up an NVIDIA Vera Rubin NVL72 system for validation, another big step in building the next generation of AI infrastructure with NVIDIA. — https://nitter.net/satyanadella/status/2032515189086761005#m
→ View original post on X — @tenstorrent, 2026-03-14 17:25 UTC
-
Gigawatt AI Intelligence Factories Power Requirements US Infrastructure
By
–
The opposite. It’s powered up by multi-Gigawatt intelligence Factories in the US. There is a site quite near to where I live in Texas that is meant to expand to 5 GW. Compare that the U.K. where the whole national grid can support ~45GW at peak demand. Your ChatGPT query is
-
Chiplet Architecture Replaces Transistor Scaling for Heat Management
By
–
Transistor scaling is hitting thermal and cost ceilings, pushing chip design toward modular chiplets connected by high-speed links. With shared substrates and tailored process nodes, we manage heat and sustain performance without relying on miniaturization. Microblog @antgrasso
-
AI race hinges on scaling power and chip output
By
–
The AI race will come down to scaling power and chip output
-
Qwen 3.5 27B Matches Sonnet 4.5 on Single RTX 5090
By
–
That model, Qwen 3.5 27B Dense is equal to Sonnet 4.5 Runs great on a single RTX 5090 w/ full context But we are not anywhere near Opus 4.5 even with Qwen 3.5 397B MoE
-
Full Model GPU Offload Required to Avoid Performance Bottlenecks
By
–
Partial offloads wouldn’t work because this will create a severe bottleneck That’s why the full model has to be offloaded to the GPU to see actual performance gains