2 ms·

I don't understand the CUDA lockin in 2026... Surely anyone can take any piece of CUDA code and tell some LLM to port it to another platform and keep grinding

I don't understand the CUDA lockin in 2026... Surely anyone can take any piece of CUDA code and tell some LLM to port it to another platform and keep grinding till performance is identical?

throwdbaaway 1mo ago
Sounds like that's what z.ai did to get GLM-5.3-Flash running on Huawei chips.