It seems to me like this approach to custom-built ML hardware which can only be leased as a service, not bought outright (such as this or Google’s TPUs), is not very promising for research. Sure, this might be a cost-effective way for companies to deploy LLMs and maybe AWS can squeeze out a few interesting papers but since nobody else can do this without paying Amazon obscene amounts of money I don’t see this leading to the next great innovation in the field. It’s at least interesting that Google’s TPUs barely had an impact, I’m curious to see if this will also be true for Amazon.

Thank you for reading the source, the amount of FUD in this thread is crazy.

I happened upon two of these in much the same way, one of them unfortunately doesn’t work but the other one has been my home office keyboard (too loud for the office) for many years. I’ve always wanted to fix the broken one by swapping in a modern controller, maybe even make it wireless (because that would be truly absurd).

Ah, historical favorite

IIRC, this “accident” is simply because chocolate producers figured out it’s cheaper to make the milk slightly taste like vomit (something the Americans apparently didn’t mind too much) than cooling it properly. I think nowadays that wouldn’t really fly but back then cheaper chocolate was maybe so desirable that consumers didn’t mind the weird taste.

I disagree, something like Sponsorblock is working quite well already for ads embedded in the video and unless YouTube makes serious sacrifices in UX I don’t see how they could prevent this.

I don’t think so: Stack Overflow requires much more moderation for the comments and answers to actually stay on topic and be somewhat professional. Especially the “don’t just link somewhere, explain the thing” rule might require a lot of moderation.