@shApYT
OpenAI treated everything accessible on the internet like it's CC0, but when we the people do it, it's piracy.
@bilalsher8047
"Our attention got distracted by transformers..." I thought attention was all you need 😂
@Lykdraft
OpenAI = Closed AI, Deepseek = Open AI.
@LionBrine
Everyday I’m reminded on how 90% of people believe 90% of what they see on the internet And yes i made up those numbers
@JD-yz4kr
First, one thing, it is NOT ILLEGAL for Chinese entities or anyone else to buy Nvidia chips. It is illegal for AMERICANS to SELL high-end chips to restricted entities, so please correct your statement. Even if DeepSeek has a million Nvidia H100 chips, it is not illegal for them to have them. This idea of Americans having extra-territorial judicial sovereignty does not have any basis in international law. Also, Microsoft was saying that they noticed high levels of data transmission sometime very late last year, when DeepSeek would have needed those data early last year, for DeepSeek 0. DeepSeek R1 is obviously based on DeepSeek V3, which is the 3rd or more iteration of DeepSeek AI.
@divinelyindifferent
It’s absurd for Corporate America to expect loyalty from American consumers, especially amid all of these mass layoffs. If companies can “offshore” labor for higher profits, consumers can just as easily seek cheaper, offshore AI solutions (like DeepSeek) to save money during this tough economy. The loyalty is dead.
@ChristianIce
I'm sorry, but is the ownership of a H100 illegal in China? Because, last time I've checked, a chinese company has to respect chinese laws, not american ones.
@marcus-b4x3h
DeepSeek just pulled a pro gamer move: ditching CUDA for low-level coding to outsmart Nvidia's monopoly and proving you don't need $1B to innovate. Meanwhile, OpenAl's 'moat' is evaporating faster than their transparency. Open-source:1, Corporate walled gardens: 0. mic drop (This comment was created by DeepSeek)
@coinmaster-988
people attacking deep seek are idiots. They are doing far more for humanity by open sourcing than Open AI closing off their model for profit.
@st.altair4936
13:21 "I don't really get the hate that they're getting." It's because they're Chinese. It's just american exceptionalism coping with the fact that China is starting to become the forefront of tech that the US has been for so long.
@JRay2113
Tbh, regardless how they got there, the results speak for themselves. DSR1 showed us that it can be done more efficiently. And I have good reason to believe that Silicon Valley and the tech titans knew that. They wanted to add a barrier for entry, but DS sneaked in through the manhole cover.
@computervision557
open source vs capital, me 100% support open source
@Nekoeye
Saltyman: AI will take your jobs. DeepSeek: Oh really. 😏
@Felipe-n3j
USA to CHINA : NO WAY! We won’t allowed you CHINA to join our exclusive international space station….CHINA to USA: ohh really, no problem at all, we will just build our own more hightech space station…and it will cost us only a fraction of what you had spent to build your now dilapidated ISS…& we can finish it in just 2 yrs with out any help frm any nations.😊😊😊😊
@charlesyeeh
This Alexandr Wang guy, at first I had some respect for him as a young entrepreneur with a strong science and technology background. But the more he talked BS about DS (or smearing, to be more precise), the more he sounded like one of those ugly politicians in Washington DC. Just disgusting!
@eduardomoura2813
deepseek can give me complex scripts in 1 try, scripts that closedai can't make work even with 20 tries. I'm getting pretty close to think closedai is the one that stole data from deepseek
@4thpdespanolo
“Hey you can’t steal from us, we stole it first!”
@seraphin01
what I find fascinating with R1 is how the guys just said "eff it" with CUDA and went straight for low level coding instead. That puts, imo, a much bigger dent into Nvidia dominance than what most people talk about since they used less powerful hardware than their US counter part. All of a sudden Intel and AMD GPU becomes a lot more interesting. And considering how much leaps forward they made with the optimization and cost effectiveness, it's hard to believe other AI firms won't follow suit and keep using CUDA. Interesting times! Also love using R1, it's really a freaking great model.. And at this point I'd rather have my datas sent to china than the US if I had to pick anyway
@jackzhou4813
Americans should be thankful that DeepSeek R1 was released early. If it was released a few months later, the share prices of these U.S. technology stocks would have completely hit rock bottom, the AI bubble would have completely burst, and that would have been a financial crisis in the United States.
@bycloudAI