I would agree with this, I use Nvidia cards for professional work on Linux and I’ve never had a problem. Yeah there’s some upfront work configuring the drivers, but I’ve never had it take more than an hour to setup.

Factor 1: Not quite accurate. Yes there are categories of waste; the names change depending on the regulator. The lower level wastes are already disposed of in the US (there are already four such facilities). The politically charged problem is always the spent nuclear fuel itself.

Factor 2: Senator Reed (D-NV) was a former Senate majority leader. He extracted the defending of Yucca Mountain from the Obama administration as a concession to pass Obamacare. It’s still technically viable and not disposing of waste costs enormous amounts of money. The federal government is legally obligated to take spent fuel off the hands of operators. Obviously they have not, so the government is sued (and loses). This has cost the government roughly $20b for their inaction see here..

Factor 3: You can recycle spent fuel but there’s no concept as spent fuel with zero radioactivity.

Two largest problems in the US: Inability to manage waste and inability to execute on large scale construction required for nuclear.

I’m not really convinced that a GPU backend is needed. Was there ever a comparison of the different CLIP model variants? Or a graph optimized / quantized ONNX version?

I think the proposed solution makes a lot of sense for the task at hand if it were integrated on the pic-rs end, but it would be worth investigating further improvements if it were on the lemmy server end.