And the whole AI industry is holding up the stock market, while AI has historically always ran the hype cycle and crashed into an AI winter. Stock markets do crash after billions pumped into a sector suddenly turn out to be not worth as much. Almost none of these AI companies run a profit and don’t have any prospect of becoming profitable. It’s when everybody starts yelling that this time it’s different that things really become dangerous.
and don’t have any prospect of becoming profitable
There’s a real twist here in regards to OpenAI.
They have some kind of weird corporate structure where OpenAI is a non-profit and it owns a for-profit arm. But, the deal they have with Softbank is that they have to transition to a for-profit by the end of the year or they lose out on the $40 billion Softbank invested. If they don’t manage to do that, Softbank can withhold something like $20B of the $40B which would be catastrophic for OpenAI. Transitioning to a For-Profit is not something that can realistically be done by the end of the year, even if everybody agreed on that transition, and key people don’t agree on it.
They knew the housing/real estate bubble would pop, as it currently is…
… So, they made one final last gambit on AI as the final bubble that would magically become super intelligent and solve literally all problems.
This would never, and is not working, because the underlying tech of LLM has no real actual mechanism by which it would or could develop complex, critical, logical analysis / theoretization / metacognition that isn’t just a schizophrenic mania episode.
LLMs are fancy, inefficient autocomplete algos.
Thats it.
They achieve a simulation of knowledge via consensus, not analytic review.
They can never be more intelligent than an average human with access to all the data they’ve … mostly illegally stolen.
The entire bet was ‘maybe superintelligence will somehow be an emergent property, just give 8t more data and compute power’.
When all the hype dies down, we will see where it’s actually useful.
But I can bet you it will have uses, it’s been very helpful in making certain aspects of my life a lot easier.
And I know many who say the same.
That too is the classical hype cycle. After the trough of disillusionment, and that’s going to be a deep one from the look of things, people figure out where it can be used in a profitable way in its own niches.
… Unless its mass proliferation of shitty broken code and mis/disinformation and hyperparasocial relationships and waste of energy and water are actually such a net negative that it fundamentally undermines infrastructure and society, thus raising the necessary profit margin too high for such legit use cases to be workable in a now broken economic system.
The dot-com bubble (late 1990s–2000) was when investors massively overvalued internet-related companies just because they had “.com” in their name, even if they had no profits or solid business plans. It burst in 2000, wiping out trillions in value.
The “Internet hype” bubble popped.
But the Internet still has many valid uses.
The dot-com bubble isn’t the internet. The internet existed long before and continued to grow after. Companies that used digital posters on the internet had a crash, but the internet kept growing.
Digital posters do still have a use today, and there are still companies running on promises, but digital promises are simple and cheap. AI is not. At least not in this form.
Big tech hype will continue after this tech hype bubble pops, but that doesn’t mean the tech is good.
And the whole AI industry is holding up the stock market, while AI has historically always ran the hype cycle and crashed into an AI winter. Stock markets do crash after billions pumped into a sector suddenly turn out to be not worth as much. Almost none of these AI companies run a profit and don’t have any prospect of becoming profitable. It’s when everybody starts yelling that this time it’s different that things really become dangerous.
There’s a real twist here in regards to OpenAI.
They have some kind of weird corporate structure where OpenAI is a non-profit and it owns a for-profit arm. But, the deal they have with Softbank is that they have to transition to a for-profit by the end of the year or they lose out on the $40 billion Softbank invested. If they don’t manage to do that, Softbank can withhold something like $20B of the $40B which would be catastrophic for OpenAI. Transitioning to a For-Profit is not something that can realistically be done by the end of the year, even if everybody agreed on that transition, and key people don’t agree on it.
The whole bubble is going to pop soon, IMO.
Yep, exactly.
They knew the housing/real estate bubble would pop, as it currently is…
… So, they made one final last gambit on AI as the final bubble that would magically become super intelligent and solve literally all problems.
This would never, and is not working, because the underlying tech of LLM has no real actual mechanism by which it would or could develop complex, critical, logical analysis / theoretization / metacognition that isn’t just a schizophrenic mania episode.
LLMs are fancy, inefficient autocomplete algos.
Thats it.
They achieve a simulation of knowledge via consensus, not analytic review.
They can never be more intelligent than an average human with access to all the data they’ve … mostly illegally stolen.
The entire bet was ‘maybe superintelligence will somehow be an emergent property, just give 8t more data and compute power’.
And then they did that, and it didn’t work.
I agree with everything you said, but that doesn’t mean it can’t be very useful in many fields.
I mean, I also agree with that, lol.
There absolutely are valid use cases for this kind of ‘AI’.
But it is very, very far from the universal panacea that the capital class seems to think it is.
When all the hype dies down, we will see where it’s actually useful. But I can bet you it will have uses, it’s been very helpful in making certain aspects of my life a lot easier. And I know many who say the same.
That too is the classical hype cycle. After the trough of disillusionment, and that’s going to be a deep one from the look of things, people figure out where it can be used in a profitable way in its own niches.
… Unless its mass proliferation of shitty broken code and mis/disinformation and hyperparasocial relationships and waste of energy and water are actually such a net negative that it fundamentally undermines infrastructure and society, thus raising the necessary profit margin too high for such legit use cases to be workable in a now broken economic system.
The world revolves around the profit margin, so the current trend may even continue indefinitely… Sad.
Time will tell how much was just hype, and how much actually had merit. I think it will go the way of the
.combubble.LOTS of uses for the internet of things, but it’s still overhyped
The .com bubble had nothing to do with the Internet of Things.
Fair enough.
The dot-com bubble (late 1990s–2000) was when investors massively overvalued internet-related companies just because they had “.com” in their name, even if they had no profits or solid business plans. It burst in 2000, wiping out trillions in value.
The “Internet hype” bubble popped. But the Internet still has many valid uses.
The dot-com bubble isn’t the internet. The internet existed long before and continued to grow after. Companies that used digital posters on the internet had a crash, but the internet kept growing.
Digital posters do still have a use today, and there are still companies running on promises, but digital promises are simple and cheap. AI is not. At least not in this form.
Big tech hype will continue after this tech hype bubble pops, but that doesn’t mean the tech is good.