Demand for chips isn't going anywhere but I can see how a former CFO turned CEO might be wary of extending credit to AI startups. For a few years now venture capital has been flinging money at AI anything and Nvidia has done a good job capturing some of those dollars. Easy funding won't go on forever though and when some % of AI startups falter it makes sense the CFO doesn't want to be left holding the bag.
lesuorac 9 hours ago [-]
Why would demand be going anywhere but down?
If you can’t sell to china there goes 25% of demand.
If you can’t build a data center in the US then there goes demand.
The big labs have all developed their own chips now and would rather use those.
4 hours ago [-]
simianwords 13 hours ago [-]
This article automatically thinks there's something off with the AI boom and there's a defacto reason to be skeptical of it.
> Jensen Huang was laying the groundwork to pull back the company’s approach to backstopping the AI economy.
> Questioning returns
I mean sure you can take a bull case but I find the confidence through which this article was written to be surprising. Does the general public really think all this is mostly a bubble?
lavendermon 10 hours ago [-]
Question is not whether AI is a bubble, it's whether the venture capital stuffed into AI startups will eventually dry up, consolidate or shift.
simianwords 9 hours ago [-]
i mean pretty much everyone knows that VCs invest in high variance and high payoff startups. so this is already factored in, but the expected value should be positive.
though the examples in the article are about investing in OpenAI and data centre deals along with investing in some startups
karahime 12 hours ago [-]
I think there's a subset online who take the same approach as the article of skipping the question and just asserting that one exists before moving on to collapsemongering.
I think it's actually a broader problem that shows up elsewhere now, too, of casting aspersions and trying to assign motives, skipping establishing that something is true in favor of asserting that it is and wiggling one's eyebrows before asking why the reader would be so nervous about it being true.
lithos 11 hours ago [-]
Stochastic Parrot and Statistical Average type deals are good enough for most companies. But I'd be wary of a SV type valuations at companies it's good enough for.
It's also going to be godsend for confrontational/delay necessary type customer support companies. Imagine how exciting it will be for Airlines and Insurance companies when they will have near perfect hits on hitting the legal maximums for how slow customer service is allowed to be.
If you can’t sell to china there goes 25% of demand.
If you can’t build a data center in the US then there goes demand.
The big labs have all developed their own chips now and would rather use those.
> Jensen Huang was laying the groundwork to pull back the company’s approach to backstopping the AI economy.
> Questioning returns
I mean sure you can take a bull case but I find the confidence through which this article was written to be surprising. Does the general public really think all this is mostly a bubble?
though the examples in the article are about investing in OpenAI and data centre deals along with investing in some startups
I think it's actually a broader problem that shows up elsewhere now, too, of casting aspersions and trying to assign motives, skipping establishing that something is true in favor of asserting that it is and wiggling one's eyebrows before asking why the reader would be so nervous about it being true.
It's also going to be godsend for confrontational/delay necessary type customer support companies. Imagine how exciting it will be for Airlines and Insurance companies when they will have near perfect hits on hitting the legal maximums for how slow customer service is allowed to be.