Gotta say, for a state project this is unintentionally great PR. We are in the dark and discover what each part does, episode by episode. "How…
RT Bryan CheongSummarising this premise and telling every policymaker or policy-adjacent person I meet on both sides of the Pacific and telling them this is the…
Music to Wang's ears (elevator music? Slav phonk? who knows)Andrew Freedman, CFA 🦅: OpenAI is worth a lot less today than it was before $META released…
> Sinopec is also training its industrial large language model “Great Wall” (长城) here.I raally wonder what kind of LLM an oil and gas company needs…
There is something innate thoughI don't even mean IQ or "East Asian work culture". I think the biggest "innate" factor is reverence for competence. Everyone knows…
Maybe in another timeline I'd have taken the offer…Glinert 🇺🇸 🏭: @teortaxesTex
yeah, 1 MW of compute would be a beast for local inferenceman do I hate these LLM bots. big Screamers energyHenry Nguyen: @teortaxesTex That would be…
Nscale CEO the next morning after discussing pivot to TPU with Google"CUDA gang sends you their regards"Jukan: Jensen Huang is absolutely ruthless. He doesn’t leave even…
Xi will probably hype up the 950DT SuperPODsoon…I predict it's the 1024 NPU versionAndrew Curran: President Xi Jinping will attend the opening ceremony of the 2026…
> 91EB/year of supply versus 120EB/yr of demandAll these insane expansions, and no saturation in sightThis, mind you, is not even some unhinged singularitanian fast takeoff…