That is true philosophically but not in the case of AI. I use Claude heavily to write code. My mother in law uses it to query legal documents and standards. My wife uses it as a roided up Google.
Generally speaking, that's incorrect. That's like saying "I don't like cars, and don't see the value in cars, therefore the market for cars is fraudulent".
In AI, buyers are getting what they want. The demand is real. YOU might not value what they're getting, but that doesn't make it fraudulent.
This is why people misunderstand why AI isn't a bubble. A bubble is asset prices rising due to speculative demand far beyond what the actual demand is.
AI - specifically chip and memory markets feeding AI - is a demand shock on par with World War 2 in its impact. NVIDIA is legitimately forecasting demand of $1 trillion in their chips+memory by the end of 2027.
This is actual, real, shipping physical product: not vapor, not something that will disappear, not something that will "crash" suddenly.
Yes, there is some speculation among AI providers training new models in the race to AGI, but that is not the majority of demand, inference is 65-80% of demand. If the current pace of training slows, that excess capacity for training will get easily sorted out through resale markets.
> Generally speaking, that's incorrect. That's like saying "I don't like cars, and don't see the value in cars, therefore the market for cars is fraudulent".
It's arguable that the car market is indeed fraudulent and the result of years of lobbying, destroying public transportation and car-centric architectures.
There are circular deals, the product is heavily subsidized. Managers are promised the moon and are deceived about the actual capabilities of the product.
Mythos is marketed like a nuclear weapon to make people jealous.
AI model government approval is floated, perhaps in preparation for a government bailout justified by "national security".
Kushners are invested in OpenAI.
There is a lot of fraud going on.
> This is actual, real, shipping physical product: not vapor, not something that will disappear, not something that will "crash" suddenly.
"Not X, not Y, not Z, just A" works better than "Actually A, not X, not Y, not Z".
> This is actual, real, shipping physical product: not vapor, not something that will disappear, not something that will "crash" suddenly.
Tulips were real shipping physical products. Railways were real. Housing was real. Whether or not the demand is speculative is largely disconnected from the actual subject of the bubble.
> NVIDIA is legitimately forecasting demand of $1 trillion in their chips+memory by the end of 2027.
Forecasts do not make it so.
> but that is not the majority of demand, inference is 65-80% of demand.
Inference is massively subsidized. The demand is fictitious just because of that. Once prices go up, especially once free or cheap inference dries up, demand will collapse.
But it's not even just the subsidies. AI is forced onto the workplace top-down. Executives demand AI be used before careful evaluation. That's all demand that can collapse at any moment if public opinion sours.
> The world has changed.
It hasn't. For all the claims that AI has made any given job so much easier, developers who claim "It'd have cost me a billion years to do so" (next time bring a counterfactual), the actual economic benefit appears to be a big fat zero. We're right back to the Solow paradox.
Except AI companies are dumping trillions of dollars into this, expecting tens of trillions in return ... from where? Where will these tens of trillions come from if the aggregate economic benefit doesn't exist? Joe Slopman making a dozen CRUD apps a week for half a million in revenue, but there ain't a million of him.
So much of the demand for inference is driven by hype. Companies using AI in the expectation of an ROI that has not materialized, and in many places, is very unlikely to. In no small part because any "efficiency" or "productivity" gains realized by AI immediately drives down the cost of the good or service produced.
Just because something is heavily in demand doesn't mean it's not fraudulent.
That is true philosophically but not in the case of AI. I use Claude heavily to write code. My mother in law uses it to query legal documents and standards. My wife uses it as a roided up Google.
It's here, it's real and it works.
That still does not mean it is not fraudulent.
Generally speaking, that's incorrect. That's like saying "I don't like cars, and don't see the value in cars, therefore the market for cars is fraudulent".
In AI, buyers are getting what they want. The demand is real. YOU might not value what they're getting, but that doesn't make it fraudulent.
This is why people misunderstand why AI isn't a bubble. A bubble is asset prices rising due to speculative demand far beyond what the actual demand is.
AI - specifically chip and memory markets feeding AI - is a demand shock on par with World War 2 in its impact. NVIDIA is legitimately forecasting demand of $1 trillion in their chips+memory by the end of 2027.
This is actual, real, shipping physical product: not vapor, not something that will disappear, not something that will "crash" suddenly.
Yes, there is some speculation among AI providers training new models in the race to AGI, but that is not the majority of demand, inference is 65-80% of demand. If the current pace of training slows, that excess capacity for training will get easily sorted out through resale markets.
The world has changed.
> Generally speaking, that's incorrect. That's like saying "I don't like cars, and don't see the value in cars, therefore the market for cars is fraudulent".
It's arguable that the car market is indeed fraudulent and the result of years of lobbying, destroying public transportation and car-centric architectures.
There are circular deals, the product is heavily subsidized. Managers are promised the moon and are deceived about the actual capabilities of the product.
Mythos is marketed like a nuclear weapon to make people jealous.
AI model government approval is floated, perhaps in preparation for a government bailout justified by "national security".
Kushners are invested in OpenAI.
There is a lot of fraud going on.
> This is actual, real, shipping physical product: not vapor, not something that will disappear, not something that will "crash" suddenly.
"Not X, not Y, not Z, just A" works better than "Actually A, not X, not Y, not Z".
> This is actual, real, shipping physical product: not vapor, not something that will disappear, not something that will "crash" suddenly.
Tulips were real shipping physical products. Railways were real. Housing was real. Whether or not the demand is speculative is largely disconnected from the actual subject of the bubble.
> NVIDIA is legitimately forecasting demand of $1 trillion in their chips+memory by the end of 2027.
Forecasts do not make it so.
> but that is not the majority of demand, inference is 65-80% of demand.
Inference is massively subsidized. The demand is fictitious just because of that. Once prices go up, especially once free or cheap inference dries up, demand will collapse.
But it's not even just the subsidies. AI is forced onto the workplace top-down. Executives demand AI be used before careful evaluation. That's all demand that can collapse at any moment if public opinion sours.
> The world has changed.
It hasn't. For all the claims that AI has made any given job so much easier, developers who claim "It'd have cost me a billion years to do so" (next time bring a counterfactual), the actual economic benefit appears to be a big fat zero. We're right back to the Solow paradox.
Except AI companies are dumping trillions of dollars into this, expecting tens of trillions in return ... from where? Where will these tens of trillions come from if the aggregate economic benefit doesn't exist? Joe Slopman making a dozen CRUD apps a week for half a million in revenue, but there ain't a million of him.
So much of the demand for inference is driven by hype. Companies using AI in the expectation of an ROI that has not materialized, and in many places, is very unlikely to. In no small part because any "efficiency" or "productivity" gains realized by AI immediately drives down the cost of the good or service produced.