@tbpn·same event
Cerebras just had the biggest IPO of the year. Founder @andrewdfeldman says the 3 most important things he had to convince investors of while doing the roadshow were that demand for inference is going to 1,000,000x, the GPU isn't the only way to do compute, and that the CUDA https://t.co/ViZr6nfm4t
@tbpn·same company or person
SemiAnalysis President @fabknowledge on the Cerebras IPO: "There is a narrow path for them. I think they're going to be able to inference maybe 1 trillion parameters and very small context window sizes. Or smaller models at very fast speeds." "There's demand. Clearly, we're in https://t.co/ToFdVfla1i
@tbpn·same theme
Braintrust CEO @ankrgyl agrees with @dwarkesh_sp's claim that "diffusion is cope" and that if models were truly better, they'd be spreading faster: "We are now programming at the maximum speed that we can program." "There are so many people who make mistakes and walk them back, https://t.co/aAIYXHHSpl https://t.co/jekWFOGmbM
@tbpn·not related
John: I don't know why there wouldn't be a lot of different customers lining up to buy Cerebras chips. Every lab that has exploding demand — Cursor, Anthropic, Meta, Google, etc. — unless they have a direct answer to this, I would see them being a buyer in the near term. https://t.co/06Bcb4blPs https://t.co/horJzEcw2t











