Re:BarraCUDA Open-source CUDA compiler targeting AMD GPUs
by jorvi on Wednesday February 18, 2026 @09:05AM
(#47061028)
I mean.. I'm one of the staunchest skeptics of LLMs as agents, but they're amazing as supercharged autocomplete and I don't see anything wrong with them in that role. There's a position between handwritten and slopped that's pareto.
[ Parent
| on HN ]