I don't believe they're going to. But I do believe they're going to try and build something.
It never is
I agree.
I don't believe OpenAI is going to build all that they say they're building. Definitely not in 4 years. But it's not like they're not building anything. There's 1 data centre built already. It's important to have these conversations with a healthy dose of skepticism, but also as much reality as we can peppered in.
Let's just say, whatever good OpenAI (and similar companies in the AI space, which among the big companies is basically all of them) are promising with this so-called "revolution" of workflows and pipelines that the promises of AI bring, I think it's increasingly obvious it'll be outweighed by mountains of bad and new problems created by the very thing they are pushing.
Not that I want to endorse violence, but the only way to draw a sufficient 'negative' attention to AI is to have some mentally ill clown attack someone, writing in blood "CHATGPT MADE ME DO IT" in a public place, after actually conversing with ChatGPT in a way that convinces the model to suggest it. Imagine the scandal.
You know, that could work. But also possible it could backfire, i.e they'll just try and turn it into a political scapegoat situation and draw it down partisan lines, even if the culprit isn't political whatsoever.
Tough to tell what will happen since they're also still the poster-child of AI, given ChatGPT was one of the earliest adopted mainstream uses of AI LLMs.
I also worry they'll flame out, and there won't be any money left from them to prosecute out for the damage. Hopefully something good happens for 2027, or some other company offers alternative memory production for the consumer market.
True, they could just bleed themselves dry with all the spending. Though, a lot of that is just companies exchanging money with each other, as others ITT have already mentioned.
But it only takes enough of them to lose too much to where they can't keep playing that game, to screw over all the others.
We do? Seems like you're making quite an assumption there.
It's an educated guess based on a grouping of circumstances and factors that lend it high probability of being true.
Never. It benefits them buying chips from generic suppliers, as they benefit from the economies of scale. OpenAI buying fabs to make their own chips is a worthless exercise in burning money, because there's no economic benefit to that kind of vertical integration.
You could argue the upfront cost of the fab purchases could be profitable by magnitudes in the future, since they wouldn't need to pay companies who are themselves out for profit margins when they sell memory to clients. And it'd also enable a company like OpenAI to sell memory to clients themselves, opening up new profit streams.
It's entirely why TSMC pulled ahead of Intel in the CPU game, and why AMD sold off their fabs and focused exclusively on chip design.
Nah, TSMC pulled ahead of Intel because Intel are just generally incompetent AF. AMD had financial problems which is why they ultimately sold their fabs. I'd say it's just as fair to state, letting the fab industry consolidate to just a very small handful of companies was never a good thing in the long run, and we've been seeing the consequences of that for years by this point.
All the same, it would've been nice if other fabs like Global Foundries stayed competitive, and if companies like NEC and Sony were able to retain their fabs over the years (well, NEC at least spun theirs out into things like Elpida, so there is that, though they don't exist anymore).
FWIW, Sony have recently invested in a new fab in Japan, so they'd disagree with your earlier statement, at least in part.
Why would they have no customers for their AI products? The entire corporate world globally are their customers. There is literally no shortage of them, since there is almost literally no limit to AI's application across every conceivable sector of enterprise (both public and private).
I'm talking about non-corporate customers. You know, folks like myself and you, when you aren't at the 9-5 office. And some of the companies pumping big into AI have large non-corporate consumer markets they are actively damaging due to some of the decisions being made WRT AI.
Let's not also forget, a
MASSIVE component of this AI push is to fire a vast majority of the same corporate employees the AI technologies are being designed to replace. Unless by corporate customers you meant the corporations directly. In which case...why should any of us who are employees at those companies or consumers of their products give a damn?
Either way, we're getting screwed but the shareholders, board members and owners of those corporations are directly benefiting. Them being the customers in this case isn't a cause for celebration; it's a big reason things are as screwed up as they are today.
Bro totally rad of you to respond. But it would be totally rad if you protested in front of their campus and beat down the man like the totally stoked Radical you are.
I'm not going to jail over something like that o.0