dora_da_destroyer
Master Baker
Everyday something else got the sky falling in this forum.
China just wrecked all of American AI. Silicon Valley is in shambles.
- Thread starter Secure Da Bag
- Start date
Cat piss martini
Veteran
fukk this country.
to be fair, ai has held this market up the past year
IIVI
Superstar
This post here really does a good job of explaining the ramifications.
Truly clever and innovative engineering. This is going to disrupt a whole lot of things.
F th USA and your tariffs
Destroy em.china
Destroy em.china
ALonelyDad
Veteran
just sent this to Trump, he said he will add 5% more to the Canadian tariffs
Yaboysix
That nikka
Oh yeah, along with the Biggest amount of Gold ever found recently, China is WINNING!
Where the Gold at, brehs...i ain't seen SHYT
Where the Gold at, brehs...i ain't seen SHYT
Yea the silicone valley and US markets are freaking out. Only wished they announced this after trump started investing some of that 500billion in a.i.
Have you seen open a.i and sam Altman's announcements?
Have you seen open a.i and sam Altman's announcements?
1/3
@kimmonismus
China is moving very very fast... first DeepSeek - now Kimi - and it's free with unlimited usage - and they said it beats 4o and 3.5 Sonnet on multiple benchmarks.
2/3
@DominicStLaure5
If they get an agent like operator, it is game over. I would be so glad to have an open source agent!
3/3
@D7QgO22P2fTo5kl
It's already released, and it's before the operator
GitHub - bytedance/UI-TARS-desktop: A GUI Agent application based on UI-TARS(Vision-Lanuage Model) that allows you to control your computer using natural language.
To post tweets in this format, more info here: https://www.thecoli.com/threads/tips-and-tricks-for-posting-the-coli-megathread.984734/post-52211196
@kimmonismus
China is moving very very fast... first DeepSeek - now Kimi - and it's free with unlimited usage - and they said it beats 4o and 3.5 Sonnet on multiple benchmarks.
2/3
@DominicStLaure5
If they get an agent like operator, it is game over. I would be so glad to have an open source agent!
3/3
@D7QgO22P2fTo5kl
It's already released, and it's before the operator
GitHub - bytedance/UI-TARS-desktop: A GUI Agent application based on UI-TARS(Vision-Lanuage Model) that allows you to control your computer using natural language.
To post tweets in this format, more info here: https://www.thecoli.com/threads/tips-and-tricks-for-posting-the-coli-megathread.984734/post-52211196
Fani Willis fan
Veteran
I would expect Chinas AI to be a lot more advanced, since they've been collecting vast amounts of data longer without much resistance from their citizens.
Full thread
1/37
@morganb
Finally had a chance to dig into DeepSeek’s r1…
Let me break down why DeepSeek's AI innovations are blowing people's minds (and possibly threatening Nvidia's $2T market cap) in simple terms...
2/37
@morganb
0/ first off, shout out to @doodlestein who wrote the must-read on this here: The Short Case for Nvidia Stock
3/37
@morganb
1/ First, some context: Right now, training top AI models is INSANELY expensive. OpenAI, Anthropic, etc. spend $100M+ just on compute. They need massive data centers with thousands of $40K GPUs. It's like needing a whole power plant to run a factory.
4/37
@morganb
2/ DeepSeek just showed up and said "LOL what if we did this for $5M instead?" And they didn't just talk - they actually DID it. Their models match or beat GPT-4 and Claude on many tasks. The AI world is (as my teenagers say) shook.
5/37
@morganb
3/ How? They rethought everything from the ground up. Traditional AI is like writing every number with 32 decimal places. DeepSeek was like "what if we just used 8? It's still accurate enough!" Boom - 75% less memory needed.
6/37
@morganb
4/ Then there's their "multi-token" system. Normal AI reads like a first-grader: "The... cat... sat..." DeepSeek reads in whole phrases at once. 2x faster, 90% as accurate. When you're processing billions of words, this MATTERS.
7/37
@morganb
5/ But here's the really clever bit: They built an "expert system." Instead of one massive AI trying to know everything (like having one person be a doctor, lawyer, AND engineer), they have specialized experts that only wake up when needed.
8/37
@morganb
6/ Traditional models? All 1.8 trillion parameters active ALL THE TIME. DeepSeek? 671B total but only 37B active at once. It's like having a huge team but only calling in the experts you actually need for each task.
9/37
@morganb
7/ The results are mind-blowing:
- Training cost: $100M → $5M
- GPUs needed: 100,000 → 2,000
- API costs: 95% cheaper
- Can run on gaming GPUs instead of data center hardware
10/37
@morganb
8/ "But wait," you might say, "there must be a catch!" That's the wild part - it's all open source. Anyone can check their work. The code is public. The technical papers explain everything. It's not magic, just incredibly clever engineering.
11/37
@morganb
9/ Why does this matter? Because it breaks the model of "only huge tech companies can play in AI." You don't need a billion-dollar data center anymore. A few good GPUs might do it.
12/37
@morganb
10/ For Nvidia, this is scary. Their entire business model is built on selling super expensive GPUs with 90% margins. If everyone can suddenly do AI with regular gaming GPUs... well, you see the problem.
13/37
@morganb
11/ And here's the kicker: DeepSeek did this with a team of <200 people. Meanwhile, Meta has teams where the compensation alone exceeds DeepSeek's entire training budget... and their models aren't as good.
14/37
@morganb
12/ This is a classic disruption story: Incumbents optimize existing processes, while disruptors rethink the fundamental approach. DeepSeek asked "what if we just did this smarter instead of throwing more hardware at it?"
15/37
@morganb
13/ The implications are huge:
- AI development becomes more accessible
- Competition increases dramatically
- The "moats" of big tech companies look more like puddles
- Hardware requirements (and costs) plummet
16/37
@morganb
14/ Of course, giants like OpenAI and Anthropic won't stand still. They're probably already implementing these innovations. But the efficiency genie is out of the bottle - there's no going back to the "just throw more GPUs at it" approach.
17/37
@morganb
15/ Final thought: This feels like one of those moments we'll look back on as an inflection point. Like when PCs made mainframes less relevant, or when cloud computing changed everything.
AI is about to become a lot more accessible, and a lot less expensive. The question isn't if this will disrupt the current players, but how fast.
/end
18/37
@morganb
P.S. And yes, all this is available open source. You can literally try their models right now. We're living in wild times!
19/37
@nikitabier
actually a good concise summary, thread boi.
20/37
@morganb
Thought I’d fukk around and go viral w/a thot piece on a Sunday night.
21/37
@kevinwtung
Fantastic summary Morgan.
22/37
@morganb
Thank you!!
23/37
@thee1of1
Thoughts on the safety of using it given that it’s from China? Risk of CCP accessing all data?
24/37
@morganb
Yeah a big difference between running the open source model on your own hardware and using the hosted app.
Definitely a bit squirrely using the ChatGPT version of TikTok
Will be interesting to see any infosec tear downs on the model but as open source you control weights and biases, fine tuning etc.
25/37
@dsog
nvda is $3.5T
26/37
@morganb
Hard to keep track
27/37
@AIML4Health
Morgan, the 3 points you’ve hit below are the ones that matter the most in that order. 8-bit; multi-token predictions; and MoE optimization.
Excellent thread. Reposting.
28/37
@morganb
Thanks ! Agree that the compounding of multiple innovations in one model is the key
29/37
@CharlesHL
ww @readwise save thread
30/37
@shouheiant
@readwise save thread
31/37
@PalveMantra
@threadreaderapp unroll
32/37
@threadreaderapp
@PalveMantra Halo! the unroll you asked for: Thread by @morganb on Thread Reader App Have a good day.
33/37
@naCrypto
@threadreaderapp unroll
34/37
@threadreaderapp
@naCrypto Guten Tag, you can read it here: Thread by @morganb on Thread Reader App Talk to you soon.
35/37
@yeeagency
I originally wanted to try to forward your content to my Chinese Twitter account. But I tried Deepseek, and I had deep doubts because its online search couldn't parse the content of X, resulting in a lot of thoughts but no results. In this case, how can friends abroad use it? I'll give some feedback to their staff.
36/37
@TheCryptoHubX
[Quoted tweet]
While Everyone Panics About China Taking Over AI, Here’s the Real Reason It Won’t Happen 
China’s Achilles Heel in the AI Market
Ask DeepSeek about Tiananmen Square, Taiwan, or Xi Jinping, and you’ll hit a wall of bias. This built-in censorship erodes trust—and trust is everything in AI.
Even better? Benchmark tests can easily include bias checks, exposing these flaws globally.
This is why China won’t dominate AI. Integrity > Propaganda.
#AI #ArtificialIntelligence #DeepSeek #China #Bias #AIMarket #TechEthics
37/37
@WilhelmMyhre
@jonasgahrstore a good idea to rethink those enormous energy-consuming data centres. @InnovasjonNorge hi there .... wakey wakey ;-)
To post tweets in this format, more info here: https://www.thecoli.com/threads/tips-and-tricks-for-posting-the-coli-megathread.984734/post-52211196
@morganb
Finally had a chance to dig into DeepSeek’s r1… Let me break down why DeepSeek's AI innovations are blowing people's minds (and possibly threatening Nvidia's $2T market cap) in simple terms...
2/37
@morganb
0/ first off, shout out to @doodlestein who wrote the must-read on this here: The Short Case for Nvidia Stock
3/37
@morganb
1/ First, some context: Right now, training top AI models is INSANELY expensive. OpenAI, Anthropic, etc. spend $100M+ just on compute. They need massive data centers with thousands of $40K GPUs. It's like needing a whole power plant to run a factory.
4/37
@morganb
2/ DeepSeek just showed up and said "LOL what if we did this for $5M instead?" And they didn't just talk - they actually DID it. Their models match or beat GPT-4 and Claude on many tasks. The AI world is (as my teenagers say) shook.
5/37
@morganb
3/ How? They rethought everything from the ground up. Traditional AI is like writing every number with 32 decimal places. DeepSeek was like "what if we just used 8? It's still accurate enough!" Boom - 75% less memory needed.
6/37
@morganb
4/ Then there's their "multi-token" system. Normal AI reads like a first-grader: "The... cat... sat..." DeepSeek reads in whole phrases at once. 2x faster, 90% as accurate. When you're processing billions of words, this MATTERS.
7/37
@morganb
5/ But here's the really clever bit: They built an "expert system." Instead of one massive AI trying to know everything (like having one person be a doctor, lawyer, AND engineer), they have specialized experts that only wake up when needed.
8/37
@morganb
6/ Traditional models? All 1.8 trillion parameters active ALL THE TIME. DeepSeek? 671B total but only 37B active at once. It's like having a huge team but only calling in the experts you actually need for each task.
9/37
@morganb
7/ The results are mind-blowing:
- Training cost: $100M → $5M
- GPUs needed: 100,000 → 2,000
- API costs: 95% cheaper
- Can run on gaming GPUs instead of data center hardware
10/37
@morganb
8/ "But wait," you might say, "there must be a catch!" That's the wild part - it's all open source. Anyone can check their work. The code is public. The technical papers explain everything. It's not magic, just incredibly clever engineering.
11/37
@morganb
9/ Why does this matter? Because it breaks the model of "only huge tech companies can play in AI." You don't need a billion-dollar data center anymore. A few good GPUs might do it.
12/37
@morganb
10/ For Nvidia, this is scary. Their entire business model is built on selling super expensive GPUs with 90% margins. If everyone can suddenly do AI with regular gaming GPUs... well, you see the problem.
13/37
@morganb
11/ And here's the kicker: DeepSeek did this with a team of <200 people. Meanwhile, Meta has teams where the compensation alone exceeds DeepSeek's entire training budget... and their models aren't as good.
14/37
@morganb
12/ This is a classic disruption story: Incumbents optimize existing processes, while disruptors rethink the fundamental approach. DeepSeek asked "what if we just did this smarter instead of throwing more hardware at it?"
15/37
@morganb
13/ The implications are huge:
- AI development becomes more accessible
- Competition increases dramatically
- The "moats" of big tech companies look more like puddles
- Hardware requirements (and costs) plummet
16/37
@morganb
14/ Of course, giants like OpenAI and Anthropic won't stand still. They're probably already implementing these innovations. But the efficiency genie is out of the bottle - there's no going back to the "just throw more GPUs at it" approach.
17/37
@morganb
15/ Final thought: This feels like one of those moments we'll look back on as an inflection point. Like when PCs made mainframes less relevant, or when cloud computing changed everything.
AI is about to become a lot more accessible, and a lot less expensive. The question isn't if this will disrupt the current players, but how fast.
/end
18/37
@morganb
P.S. And yes, all this is available open source. You can literally try their models right now. We're living in wild times!
19/37
@nikitabier
actually a good concise summary, thread boi.
20/37
@morganb
Thought I’d fukk around and go viral w/a thot piece on a Sunday night.
21/37
@kevinwtung
Fantastic summary Morgan.
22/37
@morganb
Thank you!!
23/37
@thee1of1
Thoughts on the safety of using it given that it’s from China? Risk of CCP accessing all data?
24/37
@morganb
Yeah a big difference between running the open source model on your own hardware and using the hosted app.
Definitely a bit squirrely using the ChatGPT version of TikTok
Will be interesting to see any infosec tear downs on the model but as open source you control weights and biases, fine tuning etc.
25/37
@dsog
nvda is $3.5T
26/37
@morganb
Hard to keep track
27/37
@AIML4Health
Morgan, the 3 points you’ve hit below are the ones that matter the most in that order. 8-bit; multi-token predictions; and MoE optimization.
Excellent thread. Reposting.
28/37
@morganb
Thanks ! Agree that the compounding of multiple innovations in one model is the key
29/37
@CharlesHL
ww @readwise save thread
30/37
@shouheiant
@readwise save thread
31/37
@PalveMantra
@threadreaderapp unroll
32/37
@threadreaderapp
@PalveMantra Halo! the unroll you asked for: Thread by @morganb on Thread Reader App Have a good day.
33/37
@naCrypto
@threadreaderapp unroll
34/37
@threadreaderapp
@naCrypto Guten Tag, you can read it here: Thread by @morganb on Thread Reader App Talk to you soon.
35/37
@yeeagency
I originally wanted to try to forward your content to my Chinese Twitter account. But I tried Deepseek, and I had deep doubts because its online search couldn't parse the content of X, resulting in a lot of thoughts but no results. In this case, how can friends abroad use it? I'll give some feedback to their staff.
36/37
@TheCryptoHubX
[Quoted tweet]
While Everyone Panics About China Taking Over AI, Here’s the Real Reason It Won’t Happen China’s Achilles Heel in the AI Market
Ask DeepSeek about Tiananmen Square, Taiwan, or Xi Jinping, and you’ll hit a wall of bias. This built-in censorship erodes trust—and trust is everything in AI.
Even better? Benchmark tests can easily include bias checks, exposing these flaws globally.
This is why China won’t dominate AI. Integrity > Propaganda.
#AI #ArtificialIntelligence #DeepSeek #China #Bias #AIMarket #TechEthics
37/37
@WilhelmMyhre
@jonasgahrstore a good idea to rethink those enormous energy-consuming data centres. @InnovasjonNorge hi there .... wakey wakey ;-)
To post tweets in this format, more info here: https://www.thecoli.com/threads/tips-and-tricks-for-posting-the-coli-megathread.984734/post-52211196
Stock market is fighting for its life 
Last edited:
