andallthat

Yes. The logic is... I mean clearly it's because.... Anyway, yes

Not only that, from the article they are actively trying to become a Consultancy As A Service company, where somehow companies would pay a subscription fee to… talk to their AI, I guess?

It’s like watching a real-life version of Avengers, but one where Tony Stark says “hey, this Thanos guy is diarupting industries here!” and teams up with… Thiel and Musk to fund his quest for the Infinity Stones. You know, we can’t let China get them first!

It’s almost funny how all those AI doomsday scenarios are actually meant to prop up investment in AI. See how Amodei and Altman are usually the ones pushing these narratives on how worried they are by the incredible advancements of their respective companies’ creatures. They are so, so worried about the demise of the human race and how fast it’s coming. And I sort of understand them because whatever disruption they are peddling needs to happen very fast or they will all run out of money. But what does it tell about the rest of the human race. Ghat we are actually buying into it and pouring money into creating a dystopian rfuture?

there IS a very simple explanation, but it doesn’t help sell… “how can we have our customers share the massive costs of all the computing power AI needs, while at the same time keeping access to all their yummy private data?”

Not an expert at all, but I think to an extent this already happens with the current system in most countries, and it would probably need to be done much more now. Not that Automation pays more taxes, but that having employees generally qualifies companies for tax breaks. For instance, when Amazon said “we’re going to open a new HQ”, States tripped over themselves to try and give them the largest tax breaks. But that was under the assumption that the HQ would give jobs to tens of thousand of people, not to 5 data scientist and a massive, energy-hungry data center.

And yes, you CAN run Doom on it. But turns out it’s not a great idea

***k Hey, how do you know my password?

Yes it did. Making up variation of the same story in order to farm upvotes used to be done by humans. But the strategy of throwing shit at the wall and see what sticks has now been industrialized with AI, because the machine can produce tons of cheaper, faster, smellier shit. Socials are basically the perfect application for AI. Unreliable results are not a bug but a feature. You have humans helpfully training it for free by up or downvoting the result. And the AI companies get a machine trained to persuade large groups of people of any made-up story.

I like how there are all these terms with increasingly loose definitions, to which we attach different levels of evilness: algorithm - older, more reliable, mostly deterministic except when it’s “The Algorithm” in capital letters like “The Social Media Algorithm”, then it becomes evil machine learning - been out for decades, hasn’t destroyed the world, mostly does its job undetected. Mostly used only by technical people machine intelligence - The machine is starting to become conscious but it is still mostly helpful. “Machine intelligence” does brain surgery, folds and unfolds proteins whatever that means (but it sounds like a good thing, so we’ll give it a pass) artificial intelligence - machine intelligence’s evil twin. Takes credit for everything good that comes from the other ones and we tend to believe it, because it’s the only one we can actually speak to and can lie to us very convincingly. On its own it can draw pretty pictures and animate them, write code that occasionally works, pretend to love us and teach us the most effective way to slash our own wrists

If I know anything at all after I’m dead, I will already have the answer to the biggest question from all major religions