DeepSeek launched a free, open-source large language model in late December, claiming it was developed in just two months at a cost of under $6 million.
I don’t understand why people are selling off. AI companies will just make bigger models, and then everything will be back to the same cost of compute again.
People who don’t understand tech (ie. investors) were led to believe the US had a decade headstart on AI and the GPU embargo would slow foreign competition.
Well, now here’s deepseek, giving competitive results, developed on a fraction of the compute power. Gap closed. Now these same people will believe tech companies won’t demand entire data centers of Nvidia B200s to compete.
Won’t the faster hardware be even better with more efficient models? I don’t see as much value loss as the market, especially since it’s already built.
It’s because it’s a bubble. There’s no tenable product that’s worth anywhere near the amount of money that people are/were pumping into the stocks.
Now there’s competition. So investors have to choose, who do they think is going to be more successful in the future? And that’s a decision based on the predicted value of the actual product.
And if we’re only talking about NVidia, they didn’t have any competition until now. So of course their stock was going to tank.
DeepSeek isn’t a competitor to Nvidia. They’re not making better hardware. They’re just making better software, and more accessible. Better access to software is only good for hardware companies.
And this doesn’t change the future outlook of AI. More efficient AI software just improves that outlook. Now there’s suddenly a great new potential for improved AI models.
This is bad news for OpenAI, who relies on their exclusive access to the GPT models. Good news for Nvidia, because now they’re going to sell hardware to anyone who want to take OpenAI’s throne.
Yeah, but if anything, this should be good news for Nvidia. Better access to software is good for hardware companies. This is mostly bad news for software companies like OpenAI who got lots of catching up to do.
I don’t get it either. If this makes it easier for smaller companies to run their OWN offline as well as computer hobbyists, etc…won’t it just mean more demand from all the individuals/companies vying to buy hardware to run it?
If I was a large company I sure as fuck would not be using some of the things like Copilot, etc. and sending all my data to a possible competitor. I’d be putting up a bit of cash and figuring out how to use models offline.
The reasoning is that people will stop buying their most expensive hardware offerings. This doesn’t make much sense either. Someone will come up with something clever to do with all that new compute, and then the demand for expensive hardware will be back again.
I understand why this is terrible news for OpenAI. Their entire competitive edge was that they’ve invested billions to build their GPT models. Anyone who wants to get to their levels must do the same investments.
If what’s being said about DeepSeek is true, then their entire competitive edge has just vanished overnight. Billion dollar investment: gone. It’s going to be hard for them to make the return they hoped to make from their investment.
Nvidia hasn’t lost their competitive edge. They’re still making the best hardware for AI computing. The demand for AI compute is likely going to increase.
It’s also embarrassing for OpenAI. How could they have missed this?
Agreed, I would think that hardware might still be turned to other uses like protein folding, and mining, possibly?
Personally, I think the “market” is usually filled with lots of hysterics pushing it both up and down and this is possibly just another example. I picture these stock guys as a precursor to all the dipshit Youtube “techfluencer” types (sidebar: sigh, what the hell has happened to my industry? Many of these people seem like they are one step away from video feeds moms unboxing useless cheap junk, but instead, it’s with the latest programming language or frontend Javascript framework. Apparently these types have lots of clueless followers thinking this is how to guide their careers in IT? WTAF. ) always with the hot takes and the clickbait thumbnails. Oh, and now these two groups work hand-in-glove - you can find stockbro YT hot takes all over the place reacting to DeepSeek, too. :)
i know absolutely nothing about investing. so this is a lay take:
before today i would invest in these ai big tech companies because ai is the next big thing, and there’s no other real competition around and any coming up might be ready in 3 or 4 years maybe.
after today’s news, why would i waste my money investing in companies that are building wasteful, energy consuming, and most importantly proprietary components when there’s a new generative ai that is just as good, if not better, than what’s available now, that runs on far cheaper components that is totally open sourced? i’m better off taking that money i was going to use to invest in big tech and putting it towards hiring some programmers to take the completely open sourced ai and modifying it to better suit my specific needs.
I don’t understand why people are selling off. AI companies will just make bigger models, and then everything will be back to the same cost of compute again.
You’re trying to understand the stock market? Lol
That’s the problem. It’s irrational and has always been.
People who don’t understand tech (ie. investors) were led to believe the US had a decade headstart on AI and the GPU embargo would slow foreign competition.
Well, now here’s deepseek, giving competitive results, developed on a fraction of the compute power. Gap closed. Now these same people will believe tech companies won’t demand entire data centers of Nvidia B200s to compete.
Won’t the faster hardware be even better with more efficient models? I don’t see as much value loss as the market, especially since it’s already built.
It’s because it’s a bubble. There’s no tenable product that’s worth anywhere near the amount of money that people are/were pumping into the stocks.
Now there’s competition. So investors have to choose, who do they think is going to be more successful in the future? And that’s a decision based on the predicted value of the actual product.
And if we’re only talking about NVidia, they didn’t have any competition until now. So of course their stock was going to tank.
DeepSeek isn’t a competitor to Nvidia. They’re not making better hardware. They’re just making better software, and more accessible. Better access to software is only good for hardware companies.
And this doesn’t change the future outlook of AI. More efficient AI software just improves that outlook. Now there’s suddenly a great new potential for improved AI models.
This is bad news for OpenAI, who relies on their exclusive access to the GPT models. Good news for Nvidia, because now they’re going to sell hardware to anyone who want to take OpenAI’s throne.
Opportunity cost, why keep your money in company A that can give you 1.01 for each 1 invested, when company B is giving 1.011?
Yeah, but if anything, this should be good news for Nvidia. Better access to software is good for hardware companies. This is mostly bad news for software companies like OpenAI who got lots of catching up to do.
I don’t get it either. If this makes it easier for smaller companies to run their OWN offline as well as computer hobbyists, etc…won’t it just mean more demand from all the individuals/companies vying to buy hardware to run it?
If I was a large company I sure as fuck would not be using some of the things like Copilot, etc. and sending all my data to a possible competitor. I’d be putting up a bit of cash and figuring out how to use models offline.
The reasoning is that people will stop buying their most expensive hardware offerings. This doesn’t make much sense either. Someone will come up with something clever to do with all that new compute, and then the demand for expensive hardware will be back again.
I understand why this is terrible news for OpenAI. Their entire competitive edge was that they’ve invested billions to build their GPT models. Anyone who wants to get to their levels must do the same investments.
If what’s being said about DeepSeek is true, then their entire competitive edge has just vanished overnight. Billion dollar investment: gone. It’s going to be hard for them to make the return they hoped to make from their investment.
Nvidia hasn’t lost their competitive edge. They’re still making the best hardware for AI computing. The demand for AI compute is likely going to increase.
It’s also embarrassing for OpenAI. How could they have missed this?
Agreed, I would think that hardware might still be turned to other uses like protein folding, and mining, possibly?
Personally, I think the “market” is usually filled with lots of hysterics pushing it both up and down and this is possibly just another example. I picture these stock guys as a precursor to all the dipshit Youtube “techfluencer” types (sidebar: sigh, what the hell has happened to my industry? Many of these people seem like they are one step away from video feeds moms unboxing useless cheap junk, but instead, it’s with the latest programming language or frontend Javascript framework. Apparently these types have lots of clueless followers thinking this is how to guide their careers in IT? WTAF. ) always with the hot takes and the clickbait thumbnails. Oh, and now these two groups work hand-in-glove - you can find stockbro YT hot takes all over the place reacting to DeepSeek, too. :)
i know absolutely nothing about investing. so this is a lay take:
before today i would invest in these ai big tech companies because ai is the next big thing, and there’s no other real competition around and any coming up might be ready in 3 or 4 years maybe.
after today’s news, why would i waste my money investing in companies that are building wasteful, energy consuming, and most importantly proprietary components when there’s a new generative ai that is just as good, if not better, than what’s available now, that runs on far cheaper components that is totally open sourced? i’m better off taking that money i was going to use to invest in big tech and putting it towards hiring some programmers to take the completely open sourced ai and modifying it to better suit my specific needs.
From what I’ve read, bigger models don’t necessarily mean better results - at least not past a certain point.