The problem is that they consume resources and produce almost nothing useful. It’s the same as throwing those resources away.
If we were instead building out data centers, at the same cost and of the same capacity, for the purpose of moving Canadian government data out of the US-multinational-owned cloud and back within our borders, I would support it. It’s the waste, and the likelihood that the market will crash and take even the dubious benefit of $$ for building these with it, that causes me to be against building these particular data centers.
“produce almost nothing useful” is a naive take on this situation.
Clearly they’re producing something useful, otherwise the paid revenue for these companies wouldn’t still be increasing so rapidly despite the price increases they’re pushing out. Their revenue may not be near the investment at this point, but Amazon lost money for a decade while it got established too. Even if the big companies all stopped training new models tomorrow and laid off all their researchers, there would still be demand for running those existing models.
Besides, what’s the loss to Canada if we build a few massive datacenters and then the companies that are using them go belly up? It’s not the government building the datacenters, the companies would take a loss and sell them off to someone else to use at a discount. That’s happened before, and sparked entirely new innovation due to a glut of cheap specialized resources becoming available.
Clearly they’re producing something useful, otherwise the paid revenue for these companies wouldn’t still be increasing so rapidly despite the price increases they’re pushing out.
I disagree. It’s an economic bubble: The value of these companies is being vastly inflated by a collective delusion. Those tulip bulbs were worth one heck of a lot . . . until suddenly they weren’t.
It’s just not that massive, see my other comments. People keep failing to consider the scope of the world when seeing big numbers. The world is huge, industry in general is huge, and AI is a small fish in a very large pond when it comes to resource consumption.
Not a single person has posted evidence that AI uses up a massive amount of resources in the global or industrial context. Even the person who suggested that all currently proposed AI datacenters for all of Canada could use up to 8% of current electricity just proved that it is in fact a small number compared to our total energy use.
That first example specifically refers to how fast revenue is increasing.
Companies are going to try to lobby to save money any time they can, that’s how business and lobbying works. If they don’t get it, they aren’t going to stop, they will just use the next best option.
If you don’t want your government to subsidize industries that don’t actually need it, maybe you shouldn’t have voted in Trump.
The problem is that they consume resources and produce almost nothing useful. It’s the same as throwing those resources away.
If we were instead building out data centers, at the same cost and of the same capacity, for the purpose of moving Canadian government data out of the US-multinational-owned cloud and back within our borders, I would support it. It’s the waste, and the likelihood that the market will crash and take even the dubious benefit of $$ for building these with it, that causes me to be against building these particular data centers.
These centres are not for data storage. Legally, my research on human subjects data cannot leave Canada, but we can store the data locally easily.
“produce almost nothing useful” is a naive take on this situation.
Clearly they’re producing something useful, otherwise the paid revenue for these companies wouldn’t still be increasing so rapidly despite the price increases they’re pushing out. Their revenue may not be near the investment at this point, but Amazon lost money for a decade while it got established too. Even if the big companies all stopped training new models tomorrow and laid off all their researchers, there would still be demand for running those existing models.
Besides, what’s the loss to Canada if we build a few massive datacenters and then the companies that are using them go belly up? It’s not the government building the datacenters, the companies would take a loss and sell them off to someone else to use at a discount. That’s happened before, and sparked entirely new innovation due to a glut of cheap specialized resources becoming available.
I disagree. It’s an economic bubble: The value of these companies is being vastly inflated by a collective delusion. Those tulip bulbs were worth one heck of a lot . . . until suddenly they weren’t.
I’m not talking about their value. I’m talking about companies purchasing their services, which is why I used the term Revenue and not “valuation”.
The problem is the massive environmental impact of running these centres, not who runs them.
It’s just not that massive, see my other comments. People keep failing to consider the scope of the world when seeing big numbers. The world is huge, industry in general is huge, and AI is a small fish in a very large pond when it comes to resource consumption.
your other comments are wrong and have been corrected by others. Sorry bud, not buying your bullshit here.
Not a single person has posted evidence that AI uses up a massive amount of resources in the global or industrial context. Even the person who suggested that all currently proposed AI datacenters for all of Canada could use up to 8% of current electricity just proved that it is in fact a small number compared to our total energy use.
Is that why they’re all sucking up to Trump and asking for so much subsidies to survive?
Example 1
Example 2
Americans are effectively subsidizing these companies via their tax dollars by the billions.
These companies are also losing billions. So I don’t where where you got that idea.
That first example specifically refers to how fast revenue is increasing.
Companies are going to try to lobby to save money any time they can, that’s how business and lobbying works. If they don’t get it, they aren’t going to stop, they will just use the next best option.
If you don’t want your government to subsidize industries that don’t actually need it, maybe you shouldn’t have voted in Trump.
That isn’t the technology that’s at fault.