4 ms·
It is incorrect to say that the moat is the software. The moat is primarily the compute hardware, which is still incredibly good for the price, plus really good
by why_only_15 3y ago
It is incorrect to say that the moat is the software. The moat is primarily the compute hardware, which is still incredibly good for the price, plus really good networking equipment. CUDA is not a significant moat for massive LLM training runs, as you can see from Anthropic moving from CUDA to Trainium (and so presumably rewriting all of their kernels to Trainium).