[AlanYx]

This piece ignores recent news that DeepSeek is doing inference on Huawei’s Ascend 910C GPUs.

The simplified export control analysis Amodei gives here gets a lot more complicated when increasing export controls potentially ends up spurring additional R&D in Nvidia competitors.

[rcarmo]

There are lots of people working on ASICs for inference. There are already people using FPGAs for some vertically integrated AI stacks, so it's only a matter of time until we start discussing literally hard-coding models (or flashing them) to hardware to make them go faster.