Yeah, in practice feeding AI its own outputs is totally fine as long as it’s only the outputs that are approved by users.
Yeah, in practice feeding AI its own outputs is totally fine as long as it’s only the outputs that are approved by users.
This is an extremely complex topic, you’re unlikely to find any experts on it unfortunately.
My best (still bad) suggestion would be to temporarily switch color spaces and remap the colors on the fly.
Agreed, numpy really could/should be built in.
Why would it be a bad sign that the language has built in tools for common things you need to do?
Sleeping with your Vaporeon: https://m.youtube.com/watch?v=-EWMgB26bmU
Right, in my experience the majority of URLs generated by LLMs are just jumbles of letters that vaguely look like a URL. A fundamental architecture difference needs to happen in one way or another to properly cite sources, and it’s really bad for performance.
Good question! The answer can be found by looking at how most of the commercial open source products are monetized. Software hosting and technical support are quite lucrative if the software is valuable.
But let’s look bigger than just software. How do content creators get paid? That’s far less tested. I expect crowdfunding to be the primary vehicle for that. It’s popular for indies, but the big boys haven’t caught up with the times yet.
Indeed! My personal political alignment does in fact incorporate much of communism.
I’m a digital communist, at any rate. If something can be copied for free, it darn well ought to be free. Anything else is artificial and enforced by threat of violence.
Mostly due to part fatigue. They regularly reboost it into a higher orbit.
Pathetic and expected. It takes a lot of time and money to bring someone to trial. And since lawyers are paid hourly, both sides simply love to stretch it out as long as possible.
I see you just printed a bonchee.
Oh yeah, at the time Obama was right to try to repair relations with Russia. They’ve simply grown worse as they figured out how to poke at the world via information warfare.
Yup! Remember Romney saying that Russia was our geopolitical enemy and Obama being the one who called that ridiculous? Ever since then the Republicans have had an absurd love affair with Putin.
Apocalypticism will always be in fashion.
Asking that question is the first step people need in order to finally come to that conclusion. We all just completed the process a loooooong time ago.
Sure, go for it. But good luck paying an army of copywriters to summarize every article you read.
That’s not what I’m implying. What I’m saying is that wasting time and effort on quality is pointless when the threshold for success is low.
For example, I could use aerospace quality parts (perfectly machined to micron-level tolerances) to build a toaster. However, while this would not increase the performance meaningfully, the cost would be orders of magnitude greater. Instead I can use shitty off-the-shelf parts because it doesn’t really make a difference.
Maybe in other words, engineering tolerances apply to LLMs too. They’re crude devices, but it’s totally fine if you have a crude problem.
It might be all I care about. Humans might always be better, but AI only has to be good enough at something to be valuable.
For example, summarizing an article might be incredibly low stakes (I’m feeling a bit curious today), or incredibly high stakes (I’m preparing a legal defense), depending on the context. An AI is sufficient for one use but not the other.
Only up to the point where humans notice it. It’ll make AI images easier to detect, but still pretty for humans. Probably a win-win.