NVDA’s Statement on DeepSeek
- Thread starter NoleATL
- Start date
I was just gonna say something similar.
“Everybody should buy back in.” - me with an uncomfortable 20 percent of brokerage portfolio in $NVDA and $TSMC
“Everybody should buy back in.” - me with an uncomfortable 20 percent of brokerage portfolio in $NVDA and $TSMC
the current spin is that deepseek’s approach (pending reproducibility) reduced gpu needs by an order of magnitude but due to cost savings so many more apps will emerge that runtime need for compute will go through the roof. in other words nvda is actually the biggest winner! that’s a fantastic spin and plausible too.
props to the chinese for open sourcing. shrewd move for sure but mayb beneficial in the long run for everyone hopefully
that said i would be very nervous to run their app version
Deepseek just proved you need fewer Nvidia chips than previously believed to do the kind of computing they’re doing. Their position has materially changed. I’d question the valuation.
The tech bros asking for 100s of billions in funding only to be beat by China doing it ten times cheaper lol.
Trump making this huge announcement with $500 billion investment… lol buddy NVDA goes up and down $500 bill in a day bitch
THIS! DeepSeek was released 3 days before Trump’s announcement. They knew…
Has there been any verification of the claims made? Not from what I have seen. I believe nothing out of china. Nvidia will be just fine.
I've broken Deepseek multiple times asking it science questions. It apologizes when I said it was wrong and it still couldn't fix itself.
I guess my job is safe from my computer overlords.
I guess my job is safe from my computer overlords.
Why would NVDA claim you need so many of their chips to handle AI development? What’s their angle?🤫
It’s not gobs of money, I know that!
Agreed and IMO the money will be made by things that use AI to reduce costs and time lines. AI itself will get cheaper and cheaper, like TV sets.
I would bet on things like SAP and Salesforce
chatgpt otoh apparently cannot be broken. confident legit looking answers are always provided and many times turn out to be total bullshit which of course is exactly how humans also are.
I’d love to see how they coded the censoring of the things like the Tiananmen Square Massacre in their ‘open source’.
i understand that the code is open-source (as in we can compile and have it run on data of our choice) but am i correct in thinking we might never know what data they used for training or how their (chinese deepseek ai) weights affect answers, correct? So is it fair to think that even though the code is open-source their ai actually isn't?
The screen grabs I’ve seen where it starts to answer, and then stops itself as soon as it says the ‘forbidden thing’ have to be in the code, not it’s training. It’s ’catching itself’ after formulating the forbidden answer.
Can someone ask it why Xi looks like Winnie the Pooh?
Weird that creating laws to responsibly regulate something before a problem arises is regarded as strange, unless I'm misinterpreting the post. It certainly is different than the American approach which is to let people get harmed by something and then create laws to regulate that one specific instance. Gotta let the "innovators" grift first!
Not weird that one environment suffocates innovation, and the other doesn’t.
The regulation versus innovation issue has been highlighted in various forums. Nicolai Tangen, who manages the world’s largest sovereign wealth fund, has expressed concerns over Europe's growth prospects and innovation potential. As reported by the Financial Times, Tangen believes “Europe is less hard-working, less ambitious, more regulated, and more risk-averse than the US, with the gap between the two continents only getting wider.” His point was further emphasized in a LinkedIn post contrasting a European innovation—a bottle cap designed to prevent litter—with an image of Jensen Huang, CEO of Nvidia, showcasing the company's new super chip, underscoring the perceived disparity in innovation levels.
I break eggs with no intention of making an omelet. Just fried eggs. Does this make me a psychopath?The world is our omelet.
A few weeks back I went a few rounds with ChatGPT. I asked it how to do something and it kept giving me answers that were wrong. I'd tell it that the answer was wrong and it would apologize and come up with another solution that was wrong. When I told it that solution was also wrong it repeated the first wrong answer.
But you're not safe with idiots in charge, they don't always care about right and wrong.
Don't count on what any corporation says about this as truth. Out of China? Doubtful.
Maybe more like this: https://tubitv.com/movies/590869/the-china-hustle
No idea why it's rated R.
It has an Iowa connection btw. Byron Roth is Iowa born and raised former all state football player. Roth Capital.
Just don't believe anything about this. If it's true which is doubtful we are terrified of Tic Tok.
Maybe more like this: https://tubitv.com/movies/590869/the-china-hustle
No idea why it's rated R.
It has an Iowa connection btw. Byron Roth is Iowa born and raised former all state football player. Roth Capital.
Just don't believe anything about this. If it's true which is doubtful we are terrified of Tic Tok.
NVDA released a statement that pretty much said it was a revolutionary product and it opened a 3rd way of learning. As best as I could understand it.
Similar threads
Latest posts
-
-
-
HawkCast 122 BIG BOYS: 3 OL Commits, MBB to Miss B1G Tourney?, CC Jersey
- Latest: Eliot Clough
-