How Taalas "prints" LLM onto a chip?

Re:How Taalas "prints" LLM onto a chip? by xattt on Sunday February 22, 2026 @06:32AM (#47110175)

Is it possible to supplement the model with a diff for updates on modular memory, or would severely impact perf?

[ Parent | on HN ]

Re:How Taalas "prints" LLM onto a chip? by baq on Sunday February 22, 2026 @06:45AM (#47110252)

this design at 7 transistors per weight is 99.9% burnt in the silicon forever.

[ Parent | on HN ]

Re:How Taalas "prints" LLM onto a chip? by mips_avatar on Sunday February 22, 2026 @08:46AM (#47110960)

I imagine you could do something like a LORA

[ Parent | on HN ]