“How is it possible that AI became productive and useful only six months ago, and they were somehow laying people off two years ago because of AI?” he added. “It doesn’t make any sense.”
This is very telling. Jensen is pulling this “6 months” figure completely out of his ass here, but the reason why he wants that number to be true is because it moves the goalposts. If AI hasn’t actually, really, been here for even a single fiscal year then it explains away everything. Suddenly the fact that it’s made zero impact on productivity, that no one is making any profit on it, all of that becomes justified. “It’s still early.” You’ll recall that this was the narrative around crypto too. Every time anyone criticized anything about it a herd of sheep would bleat “It’s still early” even over a decade into the technology existing.
Investors are starting to ask serious questions about when these tools are actually going to start delivering greater productivity to their companies. Managers are starting to get the screws put to them about why their budgets are ballooning to cover subscription and token costs with nothing to show for it. Jensen can’t have that, because AI is the whole reason why his company is on top of the world, so he’s trying to reset the clock.
For the record, there’s absolutely no evidence to suggest that AI has ever become productive and useful, but that wouldn’t fit Jensen’s narrative either. So instead he has to invent a world where AI is totally productive, 100% useful, just trust me! When did that happen? Oh, just now. That’s, um… Yeah, that’s why you didn’t notice. It just happened, right before you walked in.
If AI hasn’t actually, really, been here for even a single fiscal year then it explains away everything. Suddenly the fact that it’s made zero impact on productivity, that no one is making any profit on it, all of that becomes justified. “It’s still early.”
I think you’ve nailed it.
I clearly remember 2023 being the year where AI and ChatGPT hit the mainstream. Looking back, nvidia’s stock price had already doubled in the first half of 2023.
Transformer model AI has atrocious unit economics. The only way it really works is in some kind of post-scarcity environment where we simply don’t care how much it costs to run.
Crypto only solves problems it creates, or creates new problems out of the ones it solves. It’s a horrendously complicated way of wasting compute power to ultimately achieve nothing.
Monero is a good form of crypto, it’s ACIS resistant so it doesn’t require a big server to mine them. It’s also really good for anonymous transactions.
Transformers aren’t supposed to be used for big problems that require too much data. But there’s use cases for them.
This is very telling. Jensen is pulling this “6 months” figure completely out of his ass here, but the reason why he wants that number to be true is because it moves the goalposts. If AI hasn’t actually, really, been here for even a single fiscal year then it explains away everything. Suddenly the fact that it’s made zero impact on productivity, that no one is making any profit on it, all of that becomes justified. “It’s still early.” You’ll recall that this was the narrative around crypto too. Every time anyone criticized anything about it a herd of sheep would bleat “It’s still early” even over a decade into the technology existing.
Investors are starting to ask serious questions about when these tools are actually going to start delivering greater productivity to their companies. Managers are starting to get the screws put to them about why their budgets are ballooning to cover subscription and token costs with nothing to show for it. Jensen can’t have that, because AI is the whole reason why his company is on top of the world, so he’s trying to reset the clock.
For the record, there’s absolutely no evidence to suggest that AI has ever become productive and useful, but that wouldn’t fit Jensen’s narrative either. So instead he has to invent a world where AI is totally productive, 100% useful, just trust me! When did that happen? Oh, just now. That’s, um… Yeah, that’s why you didn’t notice. It just happened, right before you walked in.
I think you’ve nailed it.
I clearly remember 2023 being the year where AI and ChatGPT hit the mainstream. Looking back, nvidia’s stock price had already doubled in the first half of 2023.
Crypto and AI could have been a good thing. The politicians, criminals and free-loaders ruined it.
Transformer model AI has atrocious unit economics. The only way it really works is in some kind of post-scarcity environment where we simply don’t care how much it costs to run.
Crypto only solves problems it creates, or creates new problems out of the ones it solves. It’s a horrendously complicated way of wasting compute power to ultimately achieve nothing.
Monero is a good form of crypto, it’s ACIS resistant so it doesn’t require a big server to mine them. It’s also really good for anonymous transactions.
Transformers aren’t supposed to be used for big problems that require too much data. But there’s use cases for them.