Look I wanted to like bsky but the fact is that their userbase has been doing as much rage and culture wars and engagement bait as any other. I mean look at what just happened the other day with the paint.net guy
It seems the paint.net situation was more about anti-AI folks vs a pro AI dev. Not very red or blue coded, nor engagement bait, as far as I can tell. I suppose it is something of a culture war. Albeit without the CEO's thumb on the scales.
Good point, re: the CA law. The rule for text output may be limited to the EU AI Act, which Anthropic notes as a reason in the linked article about its implementation.
no, you don't have to make up some non-idiotic argument that's kinda vaguely like the dumb one and substitute it in there. it's not necessary! nobody's making you do it and there's no benefit to anybody from doing it!
back when the USSR was still a going concern, there were zero American leftists claiming that it wasn't really communist. that only started happening after they failed
Yeah, but how often is this? Certainly there are weird hobbyist edge cases that don't do well with this, that's true of anything, but a provider is loading weights once every few months
It all depends how many models you are serving from that flash and whether they all fit in there together. If you need to evict and load models, the flash will die a horrible death. This is one resource you should NEVER underprovision.
This is a game changer for local inference IMO, where you will basically never need to do this. On the other hand in a serverless/cloud setting it may be more problematic if you don't want to strand GPUs with locally attached HBF.
Every time I see somebody make an argument about "externalities" it's never actually backed up with any fact-based analysis about the magnitude of them. It's always exclusively vibes-based, "this thing you like is too cheap, it should be prohibitively expensive to make up for the thing I just thought of"
How do you expect people to discuss them across the scale of a whole economy without giant spreadsheets? Externalities are a well understood economic concept whose study goes back a century.
I'm not saying externalities don't exist, I'm saying that basically every time somebody makes an argument relying on the concept, there's no attempt to quantify the externalized cost, it's just a vibe. Just go upthread, the argument is just "the stuff you're buying is crap that I don't think you really need, therefore it should be more expensive, however much more it takes for you to stop buying it".
The economic concept implies that we're going to have an evidence-based discussion about how we're tracking specific dollar cost amounts of pollution or something, and then in practice that essentially never happens and it's just "I've decided you don't really need that thing you want"
That’s because it would not be possible to discuss reasonably these things without writing a thesis. It’s better to hand wave the cost of the externalities and those that are familiar with the domain will understand - rather than writing a thesis on “the externalities of manufacturing and distributing <product> in this <industry> <subsector> in <country> and relative cost to export versus produce locally.”
reply