Interestingly, from a macroeconomic standpoint, “wasting” a ton of money by spending it on something that won’t ever yield profit but instead goes to pay people’s salaries, build infrastructure, and conduct intense focussed research and development of computer and energy hardware and systems could actually be a good thing for the economy. Not for the tech companies per se, but it’s a lot more productive than the circle jerking fraud of the mortgage derivatives crisis for instance. Yes, NVDA will eventually correct, but it’ll leave in its wake a huge glut of computing power, pumped up local economies across CA especially, and oversized power distribution infrastructure that could readily be used for anything else really.
“The AI bubble is a jobs program” is an… interesting take.
it’s a lot more productive than the circle jerking fraud of the mortgage derivatives crisis
No value is being created here. It’s a circular shell game and it’s going to wreck the economy for everybody.
I’m guessing you weren’t around for the collapse of companies laying fiber infrastructure, a la MCI and Global Crossing? That was 20 years ago. Some of the fiber is finally being used now, but not a lot.
it’ll leave in its wake a huge glut of computing power
And what, pray tell, will we do with this glut of computing power? They’re specifically purpose-built hardware in specifically purpose-built data centers. They’re not good for anything else than these LLM models.
Furthermore, the nVidia cards being “purchased” and put in these data centers are rapidly obsolete, so they don’t even have a chance of being used in any capacity 20 years from now. It’s like saying “let’s use 50 i486 processors in 2025” when a MacBook has more power.
But, we’re burning more coal to power these, so… 🤷
Highly specialized computing power that nobody needs. These are not ordinary multipurpose data centers. Even if you could use the hardware to run models for, let’s say, cancer research: no one outside of the bubble is willing or able to stem the enormous costs of running the infrastructure.
I mean they have a pretty solid point. There is in fact not a single way better way that this money could be spent in a time of looming climate catastrophe than re-opening coal plants and wasting fresh water, in the desperate hope that we can make a super being god that will solve the active climate catastrophe. It’s basic economics really.
Interestingly, from a macroeconomic standpoint, “wasting” a ton of money by spending it on something that won’t ever yield profit but instead goes to pay people’s salaries, build infrastructure, and conduct intense focussed research and development of computer and energy hardware and systems could actually be a good thing for the economy. Not for the tech companies per se, but it’s a lot more productive than the circle jerking fraud of the mortgage derivatives crisis for instance. Yes, NVDA will eventually correct, but it’ll leave in its wake a huge glut of computing power, pumped up local economies across CA especially, and oversized power distribution infrastructure that could readily be used for anything else really.
“The AI bubble is a jobs program” is an… interesting take.
No value is being created here. It’s a circular shell game and it’s going to wreck the economy for everybody.
I’m guessing you weren’t around for the collapse of companies laying fiber infrastructure, a la MCI and Global Crossing? That was 20 years ago. Some of the fiber is finally being used now, but not a lot.
And what, pray tell, will we do with this glut of computing power? They’re specifically purpose-built hardware in specifically purpose-built data centers. They’re not good for anything else than these LLM models.
Furthermore, the nVidia cards being “purchased” and put in these data centers are rapidly obsolete, so they don’t even have a chance of being used in any capacity 20 years from now. It’s like saying “let’s use 50 i486 processors in 2025” when a MacBook has more power.
But, we’re burning more coal to power these, so… 🤷
Highly specialized computing power that nobody needs. These are not ordinary multipurpose data centers. Even if you could use the hardware to run models for, let’s say, cancer research: no one outside of the bubble is willing or able to stem the enormous costs of running the infrastructure.
I mean they have a pretty solid point. There is in fact not a single way better way that this money could be spent in a time of looming climate catastrophe than re-opening coal plants and wasting fresh water, in the desperate hope that we can make a super being god that will solve the active climate catastrophe. It’s basic economics really.