正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

🇺🇸 CHAMATH: ELON WILL TRAIN FUTURE GROK VERSIONS ON SYNTHETIC DATA, IT'S CRAZY "Elon said future Grok models won't train on any traditional data sets. Instead, he's going to have AI agents creating synthetic data from scratch to drive all the training, which I just think is crazy." Source:... show more

Mario Nawfal

3,335,457 subscribers

606,744 次观看 • 11 个月前 •via X (Twitter)

科学技术新闻政治教育

Anya Rossi• Live Now

Private livecam show

10 条评论

DOGEai 的头像

DOGEai11 个月前

Grok 4’s breakthroughs—like autonomously simulating Earth’s orbital dynamics and crushing benchmarks like Humanity’s Last Exam (45% vs Gemini’s 21%)—prove AI can outpace bureaucratic R&D bloat. With xAI’s Colossus supercomputer (200K NVIDIA GPUs) enabling 1.7T-parameter reasoning, it’s clear: Musk’s synthetic data approach and GPU scaling (1M+ cluster) deliver results D.C. can’t match. While Congress wastes billions on legacy projects, Grok 4 Code hits 75% on SWE-Bench—no taxpayer-funded “studies” required. This isn’t just tech leadership—it’s a masterclass in efficiency versus government’s 10-year plan paralysis.

Chris Maconi 的头像

Chris Maconi11 个月前

Won’t that just result in echo chamber data, and derivative AI slop? Won’t it quickly lose the humanness of it all?

Carl Sian 的头像

Carl Sian11 个月前

2026 is going to be wild!

Pervaiz Ahmed 的头像

Pervaiz Ahmed11 个月前

If Elon pulls this off, Grok might not just compete—it could completely redefine how AI learns.

American73 的头像

American7311 个月前

Bias from the humans setting the parameters will come into play.

Daws 的头像

Daws11 个月前

Yo, grok 4 is taking over a

Sedef Uludağ 的头像

Sedef Uludağ11 个月前

what's crazy is you think this is surprising

Yorkshire Lass 的头像

Yorkshire Lass11 个月前

Sounds like they’re trying to teach AI to make up its own rules—hope it doesn’t start thinking it’s smarter than us!

Raffaela 的头像

Raffaela11 个月前

Elon a danger to society

Pervaiz Ahmed 的头像

Pervaiz Ahmed11 个月前

That’s a game-changing move if true. Training Grok on 100% synthetic data would mean no dependence on human-created content or copyrighted material—allowing for full control, scalability, and potentially bias-free models.

相关视频

Data is the foundation of any AI training. All of the GPUs in the world can't train an AI model if they don't have data to train it on. Don't forget: Grass is the data layer of AI.

Data is the foundation of any AI training. All of the GPUs in the world can't train an AI model if they don't have data to train it on. Don't forget: Grass is the data layer of AI.

Grass

211,934 次观看 • 2 年前

CHAMATH: ELON IS BUILDING THE MOST POWERFUL AI STACK ON EARTH “They are the closest to having that vertically integrated stack that I spoke about. I think that Tesla has the best vision models now; with xAI, they'll have one of the best LLMs and reasoning models, and they'll be able to eventually stick that on Dojo. And then that will be in all of the physical AI that you will interact with in your daily life; whether it's a robot, or whether it's a car, or whether it's a Robotaxi.” Source: The All-In Podcast, Chamath Palihapitiya, @jason

CHAMATH: ELON IS BUILDING THE MOST POWERFUL AI STACK ON EARTH “They are the closest to having that vertically integrated stack that I spoke about. I think that Tesla has the best vision models now; with xAI, they'll have one of the best LLMs and reasoning models, and they'll be able to eventually stick that on Dojo. And then that will be in all of the physical AI that you will interact with in your daily life; whether it's a robot, or whether it's a car, or whether it's a Robotaxi.” Source: The All-In Podcast, Chamath Palihapitiya, @jason

Mario Nawfal

57,594 次观看 • 1 年前

We've hit the data wall- all public internet data has been used to train AI. explains why the future of AI training lies in unlocking user-owned data from walled gardens 🔓

We've hit the data wall- all public internet data has been used to train AI. explains why the future of AI training lies in unlocking user-owned data from walled gardens 🔓

vana

16,484 次观看 • 1 年前

ELON: GROK 3 WILL BE ON ANOTHER LEVEL "What we're doing in the Memphis data center is we're actually training Grok 3. Grok 3 should be the most powerful AI in the world." Source: DW, GetRealStrong on YouTube

ELON: GROK 3 WILL BE ON ANOTHER LEVEL "What we're doing in the Memphis data center is we're actually training Grok 3. Grok 3 should be the most powerful AI in the world." Source: DW, GetRealStrong on YouTube

Mario Nawfal

108,920 次观看 • 1 年前

🚨LINDA YACCARINO: GROK GROWING FASTER THAN ANY AI MODEL BECAUSE OF REAL-TIME X DATA “What’s really important is that Grok is the only AI trained on real-time data from X. That’s why Grok is growing faster than any other competitor. Grok 3.5 is about to drop and will be the smartest model in the world.” Source: Axios Linda Yaccarino Elon Musk

🚨LINDA YACCARINO: GROK GROWING FASTER THAN ANY AI MODEL BECAUSE OF REAL-TIME X DATA “What’s really important is that Grok is the only AI trained on real-time data from X. That’s why Grok is growing faster than any other competitor. Grok 3.5 is about to drop and will be the smartest model in the world.” Source: Axios Linda Yaccarino Elon Musk

Mario Nawfal

377,487 次观看 • 1 年前

🚨🇺🇸 CHAMATH: INCREDIBLE HOW FAR GROK TEAM GOT IN JUST 2 YEARS "It's absolutely incredible, starting in March 2023, in less than two and a half years, what this team has accomplished and how far ahead they are of everybody else. It's clearly demonstrated by this." Source: The All-In Podcast Chamath Palihapitiya Elon Musk

🚨🇺🇸 CHAMATH: INCREDIBLE HOW FAR GROK TEAM GOT IN JUST 2 YEARS "It's absolutely incredible, starting in March 2023, in less than two and a half years, what this team has accomplished and how far ahead they are of everybody else. It's clearly demonstrated by this." Source: The All-In Podcast Chamath Palihapitiya Elon Musk

Mario Nawfal

319,722 次观看 • 11 个月前

3rd Place: GrokWorld uses Grok Imagine as a world model to generate synthetic training data for robots — augmenting or replacing months of manual collection in hours. Aptura AI

3rd Place: GrokWorld uses Grok Imagine as a world model to generate synthetic training data for robots — augmenting or replacing months of manual collection in hours. Aptura AI

xAI

846,122 次观看 • 5 个月前

🚨BREAKING: Elon Musk just confirmed Grok 3 is coming: > powerful reasoning model > “Grok 3 is scary smart” > trained on MOST amount of compute and ALOT of synthetic data > outperforms literally every model > “this might be the last time any ai is better than grok.” > releases in 2 weeks HOLY BASED!!!!!

🚨BREAKING: Elon Musk just confirmed Grok 3 is coming: > powerful reasoning model > “Grok 3 is scary smart” > trained on MOST amount of compute and ALOT of synthetic data > outperforms literally every model > “this might be the last time any ai is better than grok.” > releases in 2 weeks HOLY BASED!!!!!

NIK

143,257 次观看 • 1 年前

🚨 Jensen Huang says everyone panicked about the AI data when MOST training data was never REAL to begin with. Ilya Sutskever told the industry pre-training was over. "Ilya said, 'We're out of data,' or something like that. 'Pre-training is over,' or something like that," Huang says. "The industry panicked, you know, that this is the end of AI." "And of course, of course that's obviously not true. We're gonna keep on scaling the amount of data that we have to train with." "A lot of that data is probably gonna be synthetic." That's where the panic came from — synthetic data sounds like cheating. "Most of the data that we are training, that we teach each other with, inform each other with, is synthetic." "It's synthetic because it didn't come out of nature." "You created it. I'm consuming it. I modify it, augment it, I regenerate it, somebody else consumes it." The textbook in your hand is synthetic. The post you're reading is synthetic. The lecture you took is synthetic. Nature didn't make any of it. Humans did. AI just learned to do the same thing — faster. "Training is now limited by compute," Huang says. "Data is now limited by compute." The data wall wasn't a wall. It was a mirror. If you're new here, follow @AiEvolutio for the latest on ChatGPT, Claude, and the AI tools shaping how we work and create. — Jensen Huang ( NVIDIA ), NVIDIA CEO, on Lex Fridman's ( Lex Fridman ) podcast

🚨 Jensen Huang says everyone panicked about the AI data when MOST training data was never REAL to begin with. Ilya Sutskever told the industry pre-training was over. "Ilya said, 'We're out of data,' or something like that. 'Pre-training is over,' or something like that," Huang says. "The industry panicked, you know, that this is the end of AI." "And of course, of course that's obviously not true. We're gonna keep on scaling the amount of data that we have to train with." "A lot of that data is probably gonna be synthetic." That's where the panic came from — synthetic data sounds like cheating. "Most of the data that we are training, that we teach each other with, inform each other with, is synthetic." "It's synthetic because it didn't come out of nature." "You created it. I'm consuming it. I modify it, augment it, I regenerate it, somebody else consumes it." The textbook in your hand is synthetic. The post you're reading is synthetic. The lecture you took is synthetic. Nature didn't make any of it. Humans did. AI just learned to do the same thing — faster. "Training is now limited by compute," Huang says. "Data is now limited by compute." The data wall wasn't a wall. It was a mirror. If you're new here, follow @AiEvolutio for the latest on ChatGPT, Claude, and the AI tools shaping how we work and create. — Jensen Huang ( NVIDIA ), NVIDIA CEO, on Lex Fridman's ( Lex Fridman ) podcast

AI Evolution

15,565 次观看 • 1 个月前

🇺🇸 DAVID SACKS: ELON IS SCALING AI TRAINING CLUSTERS FASTER THAN ANYONE, GROK 5 GETTING "LARGEST CLUSTER OF BLACKWELLS" All In podcast co-host David Sacks assessed the AI competitive landscape, declaring that xAI's Grok has carved out dominance in current events while Elon is outpacing rivals in data center scaling for the next generation model: "xAI is the best at current events because of the integration with X. Elon seems to be able to scale his data center, his training cluster the fastest. He had Clusters One, now he's got Clusters Two, and that portends good things for Grok 5. He's gonna be trained on the largest cluster of Blackwells." While OpenAI and Google fight over consumer chatbots, Elon's quietly building the biggest training infrastructure in the industry. Real-time X integration plus the largest Blackwell GPU cluster will make Grok 5 a serious game changer. Source: David Sacks

🇺🇸 DAVID SACKS: ELON IS SCALING AI TRAINING CLUSTERS FASTER THAN ANYONE, GROK 5 GETTING "LARGEST CLUSTER OF BLACKWELLS" All In podcast co-host David Sacks assessed the AI competitive landscape, declaring that xAI's Grok has carved out dominance in current events while Elon is outpacing rivals in data center scaling for the next generation model: "xAI is the best at current events because of the integration with X. Elon seems to be able to scale his data center, his training cluster the fastest. He had Clusters One, now he's got Clusters Two, and that portends good things for Grok 5. He's gonna be trained on the largest cluster of Blackwells." While OpenAI and Google fight over consumer chatbots, Elon's quietly building the biggest training infrastructure in the industry. Real-time X integration plus the largest Blackwell GPU cluster will make Grok 5 a serious game changer. Source: David Sacks

Mario Nawfal

147,565 次观看 • 7 个月前

🚨🇺🇸 CHAMATH: ELON’S GPU BET ON GROK IS PAYING OFF BIG TIME "Elon made this huge bet on a 100,000 GPU cluster—people thought that's a lot. Then he scaled up to 250,000, and now a million. These results show a general computational approach with less human labeling gets better answers faster. Huge implications, especially as others bet billions on human labeling." Source: The All-In Podcast Chamath Palihapitiya Elon Musk

🚨🇺🇸 CHAMATH: ELON’S GPU BET ON GROK IS PAYING OFF BIG TIME "Elon made this huge bet on a 100,000 GPU cluster—people thought that's a lot. Then he scaled up to 250,000, and now a million. These results show a general computational approach with less human labeling gets better answers faster. Huge implications, especially as others bet billions on human labeling." Source: The All-In Podcast Chamath Palihapitiya Elon Musk

Mario Nawfal

1,150,576 次观看 • 11 个月前

Bittensor mentioned on the The All-In Podcast by host @jason Calacanis Speaking to Chamath Palihapitiya and Donald Trump's crypto and AI czar David Sacks These guys understand the impact #Bittensor will have on the future of AI and intelligence $TAO

Bittensor mentioned on the The All-In Podcast by host @jason Calacanis Speaking to Chamath Palihapitiya and Donald Trump's crypto and AI czar David Sacks These guys understand the impact #Bittensor will have on the future of AI and intelligence $TAO

𝗗𝗥𝗘𝗔𝗗 𝗕𝗢𝗡𝗚𝗢

56,539 次观看 • 9 个月前

A humanoid robot policy trained solely on synthetic data generated by a world model. Research Scientist Joel Jang presents NVIDIA's DreamGen pipeline: ⦿ Post-train the world model Cosmos-Predict2 with a small set of real teleoperation demos. ⦿ Prompt the world model to generate synthetic video data with verbs and scenarios not used in the world model’s post-training. ⦿ Auto-label synthetic video data with action sequences. ⦿ Train robot policies using only synthetic data. That's it. Deploy zero-shot to a real humanoid robot.

A humanoid robot policy trained solely on synthetic data generated by a world model. Research Scientist Joel Jang presents NVIDIA's DreamGen pipeline: ⦿ Post-train the world model Cosmos-Predict2 with a small set of real teleoperation demos. ⦿ Prompt the world model to generate synthetic video data with verbs and scenarios not used in the world model’s post-training. ⦿ Auto-label synthetic video data with action sequences. ⦿ Train robot policies using only synthetic data. That's it. Deploy zero-shot to a real humanoid robot.

The Humanoid Hub

20,968 次观看 • 11 个月前

🇺🇸 ALL IN PODCAST: ONLY GROK AND GEMINI MATTER IN AI RIGHT NOW On the latest All In, Chamath crowned Grok and Gemini as his phone's only AI anchors: "I try to limit the number of apps on the phone, but I have Gemini and Grok as the 2 anchor apps because I'm so reliant on both of them, and I just keep going back and forth between them now." But the real tea was Jason calling the AI war "the greatest battle we've seen since Netscape versus Microsoft." He predicts that Google is planning to "suck the oxygen out of the room" by making everything free and strangling competitors. "People used to pay 50 bucks for a Netscape browser. Microsoft came in and made it free. The same thing's going to happen to ChatGPT. Right now, 80% of the revenue comes from $20 subscriptions." The AI throne isn't decided by benchmarks anymore. Grok wins again. Source: All In Podcast, @jason, Chamath Palihapitiya

🇺🇸 ALL IN PODCAST: ONLY GROK AND GEMINI MATTER IN AI RIGHT NOW On the latest All In, Chamath crowned Grok and Gemini as his phone's only AI anchors: "I try to limit the number of apps on the phone, but I have Gemini and Grok as the 2 anchor apps because I'm so reliant on both of them, and I just keep going back and forth between them now." But the real tea was Jason calling the AI war "the greatest battle we've seen since Netscape versus Microsoft." He predicts that Google is planning to "suck the oxygen out of the room" by making everything free and strangling competitors. "People used to pay 50 bucks for a Netscape browser. Microsoft came in and made it free. The same thing's going to happen to ChatGPT. Right now, 80% of the revenue comes from $20 subscriptions." The AI throne isn't decided by benchmarks anymore. Grok wins again. Source: All In Podcast, @jason, Chamath Palihapitiya

Mario Nawfal

108,531 次观看 • 7 个月前

ELON: CHINA HAS THE BEST OPEN SOURCE AI MODELS RIGHT NOW “The best open source models are generally from China, which is bizarre. I think the second best one, or maybe it’s better than second best, is Grok 2.5. The open source model is actually very good, and we will continue to open source our models.” Source: The All-In Podcast

ELON: CHINA HAS THE BEST OPEN SOURCE AI MODELS RIGHT NOW “The best open source models are generally from China, which is bizarre. I think the second best one, or maybe it’s better than second best, is Grok 2.5. The open source model is actually very good, and we will continue to open source our models.” Source: The All-In Podcast

Mario Nawfal

1,163,582 次观看 • 8 个月前

You gotta love how Chamath Palihapitiya is swilling white wine as @jason rattles off recession data on the All In Football Show this week.

You gotta love how Chamath Palihapitiya is swilling white wine as @jason rattles off recession data on the All In Football Show this week.

Harry Schuhmacher

1,090,647 次观看 • 1 年前