Not that you should trust these things in the first place, but that's a surefire way to erode trust in these things.

Ah, gatekeeping, gotcha

Story as old as time, and you never know with non-ai. It's actually very very common.

It's funny when they say things like that: Clearly the computer turned itself on and spontaneously generated the music, the backstories, created the accounts, and uploaded the content with absolutely no human involvement, right?

Just to correct a small point of error. He wants all immigrants out, even documented, and they're willing to change legal status at any point to meet quotas.

You seem to be mistaking the jobs most replaceable with jobs most at risk. Logically, it may seem they are one in the same, but those in power are the ones who choose what to replace or not to replace.

I don't use cloud based LLMs. Instead, I run a local LLM that serves most of my needs just fine. I haven't measured my actual overall usage, but I do know this: My server idles at 140w and despite running dual 3090 gpus for LLM inference, only peaks at around 500w while generating tokens. Power limiting the cards has a negligible effect on token generation.

Overall, my electricity usage hasn't gone up compared to before I self hosted an LLM.

Yeah, I don't see why this is such a hard concept. Probably rooted in the capitalist west belief that labor gives humans value, thus a robot able to do any work a human can do is threatening the value of humans.

But imagine if every human just had a labor robot. 1 robot per person, that they could send to perform labor on their behalf while the human just did whatever they wanted.

Content of the courses and standard campus fare aside, they probably didn't like you because you exude "asshole" and carry yourself like a cunt. But go ahead and assert "reverse racism" and how they obviously didn't like you for being a white man or some delusional pity party shit. You felt judged because you decided to tell yourself you were judged. A cunt is a cunt, it ain't because you're white.

Now, explain why acknowledging things that you don't personally believe as fact is somehow bad. All you listed were statements you decided you disagree with. You've decided that your white christian self is superior so anything that veers from your beliefs is woke bullshit. Get some help.

Oh goodie! A weaker, less featured replacement!

No more "Set an alarm for me"
no more "What song is playing"
Only garbage AI responses that dodge doing anything useful.

Not that they aren't adding features to it, but until it's at least on par with regular assistant, I don't want to even hear about it.

That you know of. That is swiftly changing. Denial won't make it not happen.

Survivorship bias. The lesser did not help? Maybe the least awful candidate didn't make things go the way you wanted them to go, but if you honestly believe that things couldn't be worse? Well, we're seeing in real time right now that they could absolutely be worse than the status quo.

Grocery cost increase is related to monopolies, not just tariffs, tariffs will affect imported items, customer will foot that bill.

And neither do diffusion models. Humans, just like AI, has to learn how to produce art by observing existing art. No human creates art in a vacuum. Humans do have the advantage of learning from the world around them starting from an early age, and AI models can only learn from input images, that much is true, but conceptually it's not that different.

AI learns concepts. When you show AI a hundred pictures of a bird, and tell it what it sees is a bird, then it learns that the common features of the images are what a bird is. Meaning it won't associate leaves, the sky, or anything else in the image as being a bird. Though it may make tangential relations to those concepts since many images of birds also happen to includes such things.

Except AI does not actually crap out whole images at a time, not any more or less than humans do. You are right that when a human draws an image, they typically draw it shape at a time, since that is how they are usually taught to convey their mental image onto their medium. Image generators however start with a field of noise that has no particular shape and gradually, over many steps of re-evaluating, nudge that noise into the shapes that they've learned represent the concepts they know about.

You equate it being "not the same" as not having it at all. The entire point of diffusion models is to break things down into concepts and learn those concepts individually so that it can then create new images based on concepts. Of course, you're welcome to plug your ears and go back to believing your own bullshit, but I guess I shouldn't expect much from someone who calls themselves "drinkypoo".