AMD has ROCm which tries to get close. I’ve been able to get some CUDA applications running on a 6700xt, although they are noticeably slower than running on a comparable NVidia card. Maybe we’ll see more projects adding native ROCm support now that AMD is trying to cater to the enterprise market.

Maybe don’t allow autonomous cars on public streets then? The tech is nowhere near ready for prime time.

Lol. Guess it’s time to add the rest of my subbed channels to YT-DL and ditch their shitty ad-filled site entirely.

Because 90% of their users don’t care about APIs or 3rd party apps, they just want the content however reddit makes them consume it.

It’s not. Lemmy.ml and lemmygrad.ml are the commie instances, .world is run by a good team.

This is like asking a cop if they’re a cop, they can answer however they want. I’m no tankie but I won’t be replying back some message with a copypasta just because someone asks me to.

I haven’t seen any obvious bots in the wild here (yet), but they’ll come soon enough. There are at least a couple Github repos out there (i.e. this one) with bot libraries so I’d expect some of the old reddit bots could make some sort of a comeback pretty soon.

You can see what instances your server has blocked at {instance_url}/instances, there’s also a link in the footer. If you’re unhappy with your home instance blocking too much content, you can always make an account elsewhere and sub to the same communities as before. Account export is a feature that’s currently requested on the Lemmy Github so maybe that process will become easier soon.

AI content moderation will be chump change if they get enough takers on this new API pricing. Plus they’re wanting to IPO at a time when every Fortune 500 is jumping on the AI bandwagon in quarterly reports, so I’d be surprised if they don’t already have teams researching this internally.

They’ll probably just plug Automod into OpenAI’s content moderation API or something if too much spam piles up

I just sub to both if I run into a sublemmy collision where both are sizable. It is a little weird and I’d like to see some clean way to merge them in the future (i.e. with content migration and redirects), but for now it is what it is.

Communities can unfederate themselves at the click of a button (by an admin, of course). Or they can blacklist “bad” instances. Or whitelist specific instances and connect to nobody else.