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。

Easy predictions

Elon Musk

240,920,957 subscribers

1,824,113 次观看 • 17 小时前 •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

Gameweek 15 predictions: No matter how bad Real Madrid are, they always find it easy against Celta Vigo.

Gameweek 15 predictions: No matter how bad Real Madrid are, they always find it easy against Celta Vigo.

Adika

19,826 次观看 • 7 个月前

Vitality Czech-In 6 with dj carzzy 🐁 and Troubleinc.: An easy win and some predictions for tomorrow 😃

Vitality Czech-In 6 with dj carzzy 🐁 and Troubleinc.: An easy win and some predictions for tomorrow 😃

Team Vitality League

19,174 次观看 • 4 个月前

Arsenal vs Bournemouth Gameweek 32 predictions: This looks easy on paper for Arsenal but Bournemouth are unbeaten in their last 13 games.

Arsenal vs Bournemouth Gameweek 32 predictions: This looks easy on paper for Arsenal but Bournemouth are unbeaten in their last 13 games.

Adika

37,311 次观看 • 3 个月前

Arsenal vs West Ham Gameweek 36 predictions: West Ham are fighting to stay in the Premier League. Arsenal will not have an easy fixture with this.

Arsenal vs West Ham Gameweek 36 predictions: West Ham are fighting to stay in the Premier League. Arsenal will not have an easy fixture with this.

Adika

18,286 次观看 • 2 个月前

Beware of crypto price predictions! A narrative is spreading that's dangerously wrong. Better than expected isn't good for crypto. No rate cuts, no easy money, yet influencers claim XRP will hit $10,000? It's a mathematical impossibility for XRP #CryptoNews #XRP #Investing

Beware of crypto price predictions! A narrative is spreading that's dangerously wrong. Better than expected isn't good for crypto. No rate cuts, no easy money, yet influencers claim XRP will hit $10,000? It's a mathematical impossibility for XRP #CryptoNews #XRP #Investing

Martyn Lucas Investor

10,435 次观看 • 3 个月前

this list doesn't make sense at all because what do you mean "perfect night" by le sserafim isn't nominated when it got more points than both 'easy' and 'crazy' combined + it's the leading contender in the predictions PERFECT NIGHT DESERVES BETTER

this list doesn't make sense at all because what do you mean "perfect night" by le sserafim isn't nominated when it got more points than both 'easy' and 'crazy' combined + it's the leading contender in the predictions PERFECT NIGHT DESERVES BETTER

joe

303,429 次观看 • 1 年前

Ari (Ari Wagen) made an interface for IR spectral predictions that's so easy to use that it's hypnotizing. I measured and computed dozens of IR spectra in my PhD, perhaps hundreds, and this is the most intuitive way to interact with the data that I've ever encountered:

Ari (Ari Wagen) made an interface for IR spectral predictions that's so easy to use that it's hypnotizing. I measured and computed dozens of IR spectra in my PhD, perhaps hundreds, and this is the most intuitive way to interact with the data that I've ever encountered:

Corin Wagen

27,232 次观看 • 1 年前

With all of the negative predictions & news segments featuring folks that seem to want democracy to fail, it’s easy to forget that there are so many more of us on the right side, than not. Michigan Dems have spoken & they’re standing with President Biden! Gotta love these ladies for being clear-eyed & for keeping it real.🔥🔥🔥

With all of the negative predictions & news segments featuring folks that seem to want democracy to fail, it’s easy to forget that there are so many more of us on the right side, than not. Michigan Dems have spoken & they’re standing with President Biden! Gotta love these ladies for being clear-eyed & for keeping it real.🔥🔥🔥

Christopher Webb

314,798 次观看 • 2 年前

Every conversation I have with Dylan Patel, I'm really just trying to understand the supply and demand of tokens. This is a unique episode in that it's entirely dedicated to talking about both sides of that equation. We discuss: - The infinite demand for the newest models - SemiAnalysis going from $10K on AI spend to $7M - Mythos and Anthropic's compute problem - Why TSMC spending $100B on CapEx could cause a shortage - Robotics as next demand wave - Why memory prices will double again This is my second conversation with Dylan and find myself needing to speak with him more and more often to make sense of it all. Enjoy! Timestamps: 0:00 Intro 1:00 Surging AI Spend 10:27 Token Demand 16:21 When Ideas Are Cheap and Execution is Easy 20:46 Model Hoarding 22:34 Robotics 27:03 The Compute Bottleneck 30:26 The AI Permanent Underclass 31:39 Supply Chain Reality 37:47 CPUs 42:54 Predictions: Public Backlash

Every conversation I have with Dylan Patel, I'm really just trying to understand the supply and demand of tokens. This is a unique episode in that it's entirely dedicated to talking about both sides of that equation. We discuss: - The infinite demand for the newest models - SemiAnalysis going from $10K on AI spend to $7M - Mythos and Anthropic's compute problem - Why TSMC spending $100B on CapEx could cause a shortage - Robotics as next demand wave - Why memory prices will double again This is my second conversation with Dylan and find myself needing to speak with him more and more often to make sense of it all. Enjoy! Timestamps: 0:00 Intro 1:00 Surging AI Spend 10:27 Token Demand 16:21 When Ideas Are Cheap and Execution is Easy 20:46 Model Hoarding 22:34 Robotics 27:03 The Compute Bottleneck 30:26 The AI Permanent Underclass 31:39 Supply Chain Reality 37:47 CPUs 42:54 Predictions: Public Backlash

Patrick OShaughnessy

893,210 次观看 • 2 个月前

I tried turning World Cup passion into a full supporter anthem using Musicful… and the result genuinely surprised me. Instead of spending hours producing music, I opened Musicful's Custom Mode, added my lyrics, described the sound I wanted (Brazilian phonk with stadium chants and heavy bass), hit generate, and within moments I had a full AI-generated anthem for Brazil. ✅️No studio. ✅️No music production experience. ✅️Just an idea → a few prompts → an original World Cup chant. What impressed me most was how simple the whole process felt. Watching the anthem come together from lyrics and a style prompt made me realize anyone can create something that sounds stadium-ready. Here's what you can do: 🎵 Generate your own AI support anthem 🏆 Join Musicful's World Cup campaign 📊 Make daily match predictions 🎁 Earn free credits while participating With over 3M+ users already creating music with AI, Musicful makes it incredibly easy to turn football excitement into an original supporter anthem. Your team deserves an anthem too. 👀 Who are you supporting this World Cup? Drop your team below ⚽🔥

I tried turning World Cup passion into a full supporter anthem using Musicful… and the result genuinely surprised me. Instead of spending hours producing music, I opened Musicful's Custom Mode, added my lyrics, described the sound I wanted (Brazilian phonk with stadium chants and heavy bass), hit generate, and within moments I had a full AI-generated anthem for Brazil. ✅️No studio. ✅️No music production experience. ✅️Just an idea → a few prompts → an original World Cup chant. What impressed me most was how simple the whole process felt. Watching the anthem come together from lyrics and a style prompt made me realize anyone can create something that sounds stadium-ready. Here's what you can do: 🎵 Generate your own AI support anthem 🏆 Join Musicful's World Cup campaign 📊 Make daily match predictions 🎁 Earn free credits while participating With over 3M+ users already creating music with AI, Musicful makes it incredibly easy to turn football excitement into an original supporter anthem. Your team deserves an anthem too. 👀 Who are you supporting this World Cup? Drop your team below ⚽🔥

Leonardo

24,823 次观看 • 16 天前

Bill Ackman reveals the exact difference between valuing a stock and valuing a bond that most people never learn "A bond pays a 5% coupon interest rate. You get that every year, and it's very predictable. If it's a US Government Bond, you know you're going to get it. That's a pretty easy thing to value" "A stock is an interest in a business. It's like owning a piece of a company. A profitable business is like a bond in that it generates these coupons or these earnings or cash flow every year" "The difference is that a bond is a contract. You know what you're going to get. With a stock, you have to make predictions about the business" "How many widgets are they going to sell this year? How many next year? What are their costs going to be? How much money do they need to reinvest to keep the business going?" "What we do is try to find businesses where with a very high degree of confidence we know what those cash flows are going to be for a very long time. There are very few businesses you can have a really high degree of certainty about. Many investments are speculations because it's really very difficult to predict the future"

Bill Ackman reveals the exact difference between valuing a stock and valuing a bond that most people never learn "A bond pays a 5% coupon interest rate. You get that every year, and it's very predictable. If it's a US Government Bond, you know you're going to get it. That's a pretty easy thing to value" "A stock is an interest in a business. It's like owning a piece of a company. A profitable business is like a bond in that it generates these coupons or these earnings or cash flow every year" "The difference is that a bond is a contract. You know what you're going to get. With a stock, you have to make predictions about the business" "How many widgets are they going to sell this year? How many next year? What are their costs going to be? How much money do they need to reinvest to keep the business going?" "What we do is try to find businesses where with a very high degree of confidence we know what those cash flows are going to be for a very long time. There are very few businesses you can have a really high degree of certainty about. Many investments are speculations because it's really very difficult to predict the future"

Jaynit

157,758 次观看 • 2 天前

Super excited to share 🧠MLGym 🦾 – the first Gym environment for AI Research Agents 🤖🔬 We introduce MLGym and MLGym-Bench, a new framework and benchmark for evaluating and developing LLM agents on AI research tasks. The key contributions of our work are: 🕹️ Enables the exploration of different training algorithms for AI Research Agents such as RL 🛠️ Provides a flexible evaluation framework that can accommodate different artifacts such as models, algorithms, or predictions 🤖 Allows researchers to evaluate any model without the need to develop a custom agentic harness 🎯 Introduces 13 diverse open-ended AI Research tasks for evaluating AI Research Agents on a wide range of domains such as computer vision, natural language processing, reinforcement learning, game theory, and logical reasoning. 📈 Proposes a new evaluation metric for AI Research Agents MLGym makes it easy to: 1) Add new tasks 2) Evaluate new models 3) Integrate new agents Check out a video of the MLGym Agent to see how it performs the full pipeline of idea generation💡, implementation 👩‍💻, experimentation 👩‍🔬, and iteration 🔄 to improve on ML tasks. Huge thanks to the exceptionally talented Deepak Nathani who led this work and to all the other amazing collaborators who made this possible 🙏🫶🚀

Super excited to share 🧠MLGym 🦾 – the first Gym environment for AI Research Agents 🤖🔬 We introduce MLGym and MLGym-Bench, a new framework and benchmark for evaluating and developing LLM agents on AI research tasks. The key contributions of our work are: 🕹️ Enables the exploration of different training algorithms for AI Research Agents such as RL 🛠️ Provides a flexible evaluation framework that can accommodate different artifacts such as models, algorithms, or predictions 🤖 Allows researchers to evaluate any model without the need to develop a custom agentic harness 🎯 Introduces 13 diverse open-ended AI Research tasks for evaluating AI Research Agents on a wide range of domains such as computer vision, natural language processing, reinforcement learning, game theory, and logical reasoning. 📈 Proposes a new evaluation metric for AI Research Agents MLGym makes it easy to: 1) Add new tasks 2) Evaluate new models 3) Integrate new agents Check out a video of the MLGym Agent to see how it performs the full pipeline of idea generation💡, implementation 👩‍💻, experimentation 👩‍🔬, and iteration 🔄 to improve on ML tasks. Huge thanks to the exceptionally talented Deepak Nathani who led this work and to all the other amazing collaborators who made this possible 🙏🫶🚀

Roberta Raileanu

105,030 次观看 • 1 年前

💡 Whats the upgrade that our game-changing Trading 🐦 is going to get: Our upgraded trading tools will be built on a foundation of advanced AI technologies and blockchain integrations to deliver a seamless, smarter trading experience. Here’s a glimpse of the tech behind this upgraded trading agent: 1️⃣ Multi-Layer Attention (MLA) - This is the backbone of our AI system, enabling multiple AI agents to work in sync. - It allows the agents to collaborate on tasks like analyzing market trends, identifying token opportunities, and optimizing strategies in real time. - MLA ensures parallel processing of data for better decision-making and faster 2️⃣ Learning and Evolution System - Our AI agents are powered by a self-learning framework that constantly evolves based on market conditions and user behavior. - With every interaction, the system adapts and gets smarter, improving the accuracy of its predictions and strategies. 3️⃣ On-Chain Data Analysis - The AI bots pull data directly from Ethereum and other blockchain networks, giving them real-time access to liquidity pools, token prices, and market activity. - This deep integration ensures precise and timely execution of tasks like token purchases, profit analysis, and cross-chain swaps. 4️⃣ Natural Language Processing (NLP) - NLP models power the bot’s ability to understand your tweets and translate them into complex trading actions. - This ensures an easy-to-use, human-friendly interface that connects your social interactions to advanced trading strategies. 5️⃣ Cloud-Hosted Infrastructure - The AI operates on scalable cloud infrastructure, ensuring 24/7 uptime, fast processing, and the ability to handle large volumes of trades simultaneously.

💡 Whats the upgrade that our game-changing Trading 🐦 is going to get: Our upgraded trading tools will be built on a foundation of advanced AI technologies and blockchain integrations to deliver a seamless, smarter trading experience. Here’s a glimpse of the tech behind this upgraded trading agent: 1️⃣ Multi-Layer Attention (MLA) - This is the backbone of our AI system, enabling multiple AI agents to work in sync. - It allows the agents to collaborate on tasks like analyzing market trends, identifying token opportunities, and optimizing strategies in real time. - MLA ensures parallel processing of data for better decision-making and faster 2️⃣ Learning and Evolution System - Our AI agents are powered by a self-learning framework that constantly evolves based on market conditions and user behavior. - With every interaction, the system adapts and gets smarter, improving the accuracy of its predictions and strategies. 3️⃣ On-Chain Data Analysis - The AI bots pull data directly from Ethereum and other blockchain networks, giving them real-time access to liquidity pools, token prices, and market activity. - This deep integration ensures precise and timely execution of tasks like token purchases, profit analysis, and cross-chain swaps. 4️⃣ Natural Language Processing (NLP) - NLP models power the bot’s ability to understand your tweets and translate them into complex trading actions. - This ensures an easy-to-use, human-friendly interface that connects your social interactions to advanced trading strategies. 5️⃣ Cloud-Hosted Infrastructure - The AI operates on scalable cloud infrastructure, ensuring 24/7 uptime, fast processing, and the ability to handle large volumes of trades simultaneously.

𝕋𝕎𝔼𝔼𝕋

20,357 次观看 • 1 年前

Here's my conversation with Lucas Atkins and the team at Arcee.ai on their path to training and releasing Trinity Large today. From going all in on open models built end to end in the US 6 months ago to having the model in hand is no easy feet. I loved this conversation on how to design a startup around open models and take a bold step to scale it up. I'm openly an Arcee fan, watching them take risk and pull it off. We discuss: - The state (and future) of open vs. closed models, - The business of selling open models for on-prem deployments, - The story of Arcee AI & going “all-in” on this training run, - The ATOM project, - Building frontier model training teams in 6 months, - and other great topics. I really loved this one, and think you well too. Chapters: 00:00:00 Intro: Arcee AI, Trinity Models & Trinity Large 00:08:26 Transitioning a Company to Pre-training 00:13:00 Technical Decisions: Muon and MoE 00:18:41 Scaling and MoE Training Pain 00:23:14 Post-training and RL Strategies 00:28:09 Team Structure and Data Scaling 00:31:31 The Trinity Manifesto: US Open Weights 00:42:31 Specialized Models and Distillation 00:47:12 Infrastructure and Hosting 400B 00:50:53 Open Source as a Business Moat 00:56:31 Predictions: Best Model in 2026 01:02:29 Lightning Round & Conclusions More great open model builder podcasts coming soon!

Here's my conversation with Lucas Atkins and the team at Arcee.ai on their path to training and releasing Trinity Large today. From going all in on open models built end to end in the US 6 months ago to having the model in hand is no easy feet. I loved this conversation on how to design a startup around open models and take a bold step to scale it up. I'm openly an Arcee fan, watching them take risk and pull it off. We discuss: - The state (and future) of open vs. closed models, - The business of selling open models for on-prem deployments, - The story of Arcee AI & going “all-in” on this training run, - The ATOM project, - Building frontier model training teams in 6 months, - and other great topics. I really loved this one, and think you well too. Chapters: 00:00:00 Intro: Arcee AI, Trinity Models & Trinity Large 00:08:26 Transitioning a Company to Pre-training 00:13:00 Technical Decisions: Muon and MoE 00:18:41 Scaling and MoE Training Pain 00:23:14 Post-training and RL Strategies 00:28:09 Team Structure and Data Scaling 00:31:31 The Trinity Manifesto: US Open Weights 00:42:31 Specialized Models and Distillation 00:47:12 Infrastructure and Hosting 400B 00:50:53 Open Source as a Business Moat 00:56:31 Predictions: Best Model in 2026 01:02:29 Lightning Round & Conclusions More great open model builder podcasts coming soon!

Nathan Lambert

26,872 次观看 • 5 个月前

GM #Web3, Happy New Year 🥳! I'm super stoked to share some awesome news about Community Gaming’s Forkast and the upcoming $CGX token launch on January 28th, 2025. So, is basically a prediction market where you can bet on gaming events - Guess the answer to questions like: "Will this game ever hit 500K players?" and get rewarded if you're right! This platform is loaded with cool features: no gas fees when you claim your rewards, weekly loot boxes filled with $CGX tokens, leaderboard prizes, and daily rewards for the most active users. It's gonna be a blast! So, why did Community Gaming choose Ronin ? Well, it's a no-brainer. $RON has almost 1 million daily active users, over $1 billion in total value locked, and it's super easy to use. Plus, the team behind has already crushed it with 100,000 monthly active users and over 1,000 tournaments on Ronin. What sets Forkast apart is that it's the first big prediction market on Ronin. They're using gasless transactions via Ronin Waypoint, and here's the cherry on top: 50% of rewards will be auto-staked, so you can earn even more. As a seasoned veteran of the Ronin Network, with over 4 years of hands-on experience, I've had the privilege of witnessing its growth and evolution firsthand. Moreover, I've been actively organizing tournaments on the Community Gaming (CG) platform since early 2022, which has given me a unique perspective on the inner workings of the ecosystem. The $CGX token drops at the end of the month, and you can earn it by making accurate predictions, completing daily quests, climbing the leaderboards, and snagging those weekly loot boxes. Buckle up, folks, it's gonna be a wild ride! This thread is brought to you by Socials Rising 🃏 👊🏽😎 #AGDAO

GM #Web3, Happy New Year 🥳! I'm super stoked to share some awesome news about Community Gaming’s Forkast and the upcoming $CGX token launch on January 28th, 2025. So, is basically a prediction market where you can bet on gaming events - Guess the answer to questions like: "Will this game ever hit 500K players?" and get rewarded if you're right! This platform is loaded with cool features: no gas fees when you claim your rewards, weekly loot boxes filled with $CGX tokens, leaderboard prizes, and daily rewards for the most active users. It's gonna be a blast! So, why did Community Gaming choose Ronin ? Well, it's a no-brainer. $RON has almost 1 million daily active users, over $1 billion in total value locked, and it's super easy to use. Plus, the team behind has already crushed it with 100,000 monthly active users and over 1,000 tournaments on Ronin. What sets Forkast apart is that it's the first big prediction market on Ronin. They're using gasless transactions via Ronin Waypoint, and here's the cherry on top: 50% of rewards will be auto-staked, so you can earn even more. As a seasoned veteran of the Ronin Network, with over 4 years of hands-on experience, I've had the privilege of witnessing its growth and evolution firsthand. Moreover, I've been actively organizing tournaments on the Community Gaming (CG) platform since early 2022, which has given me a unique perspective on the inner workings of the ecosystem. The $CGX token drops at the end of the month, and you can earn it by making accurate predictions, completing daily quests, climbing the leaderboards, and snagging those weekly loot boxes. Buckle up, folks, it's gonna be a wild ride! This thread is brought to you by Socials Rising 🃏 👊🏽😎 #AGDAO

Alpha

46,310 次观看 • 1 年前

Mind blown: A Chinese quant college student builds an AI swarm engine in 10 days flat, explodes GitHub with 13,000+ stars, and scores $4,000,000 in funding! Introducing MiroFish is the multi-agent simulator that's revolutionizing predictions for trading, PR, and more. What is MiroFish? It's a digital sandbox where thousands of AI agents with individual memories and behaviors interact like a real society. Feed it any scenario (news leak, policy change, or even a classic novel's missing ending), and it simulates crowd reactions, debates, and outcomes to forecast real-world events. The Creator's Story: > In late 2025, fourth-year student Guo Hanjiang coded the core using AI assistants. > It went viral overnight, landing him 30m Yuan (~$4m) from Shanda Group. > He ditched the dorm, started a company, and now leads the charge. Key Applications: .Trading: Input financial news or reports, watch simulated market panics and price swings for predictive insights. .PR Testing: Companies/Politics run draft statements to spot backlash and refine messaging. .Creative Experiments: Loaded a lost-ending Chinese novel, agents role-played characters and generated a logical finale. .Easy setup: Deploy via Docker in minutes with any LLM API key. Pro tip: Simulate something wild like Elon Musk tweeting about Dogecoin 2.0 and spawn agent traders, influencers, and investors, generate real-time video clips of the frenzy to test moonshots or crashes risk-free. Traders are already winning big: Check this one on Polymarket - $120,000+ net profits from spot on SPX 500 bets, powered by MiroFish sims on historical data. His profile: For effortless gains, try Kreo copy trading: Auto-mirror pros like him and ride their edges. Try here: Add his wallet: [0x17559efac103ac7f361be37ec0b93888d4c55aac] to [ and start track/copy him. Repo:

slash1s

1,135,450 次观看 • 4 个月前

It's always energizing to do a podcast with Steve Yegge (Steve Yegge, engineer+author, formerly at Amazon+Google, creator of Gas Town). Timestamps: 00:00 Intro 01:43 Steve’s latest projects 02:27 Important blog posts 04:48 Shifts in what engineers need to know 10:46 Steve’s current AI stance 13:23 Steve’s book Vibe Coding 18:25 Layoffs and disruption in tech 31:13 Gas Town 40:10 New ways of working 51:08 The problem of too many people 54:45 Why AI results lag in business 59:57 Gamification and product stickiness 1:04:54 The ‘Bitter Lesson’ explained 1:07:14 The future of software development 1:23:06 Where languages stand 1:24:47 Adapting to change 1:27:32 Steve’s predictions Brought to you by: • Statsig – ⁠ The unified platform for flags, analytics, experiments, and more. • Sonar – The makers of SonarQube, the industry standard for automated code review. • WorkOS – Everything you need to make your app enterprise ready. Three interesting thoughts from Steve that we talked about in this conversation: 1. Reading ability is becoming a blocker for wider AI adoption. Some struggle with walls of text that current AI tools produce, and Steve predicts that in the very near future, most people will program by talking to a visual avatar, not reading terminal output because he observes that five paragraphs is already a lot to read for many devs. 2. What software engineers need to know keeps changing. In the 1990s, any decent software engineer knew Assembly, and today almost no decent developer knows it because Assembly has long been superseded by technical progress. What engineers “need” to know these days is different from the ‘90s and that process continues with AI, changing the parts of the craft that are essential for devs. We grumble about this but that won’t change anything by itself. 3. There’s a “Dracula Effect” where AI-augmented work drains engineers faster than traditional work. This is because AI automates the easy tasks, meaning that engineers are stuck doing high-intensity thinking all day. Steve says you may only get three daily productive hours at max speed, but during that time, you could produce 100x more output than before.

Gergely Orosz

41,987 次观看 • 4 个月前

A new 30-minute presentation from Ashok Elluswamy, Tesla’s VP of AI, has been released, where he talks about FSD, AI and the team’s latest progress. Highlight from the presentation: • Tesla's vehicle fleet can provide 500 years of driving data every single day. Curse of Dimensionality: • 8 cameras at high frame rate = billions of tokens per 30 seconds of driving context. • Tesla must compress and extract the right correlations between sensory input and control actions. Data Advantage: • Tesla has access to a “Niagara Falls of data” — hundreds of years’ worth of collective fleet driving. • Uses smart data triggers to capture rare corner cases (e.g., complex intersections, unpredictable behavior). Quality and Efficiency: • Extracts only the essential data needed to train models efficiently. Debugging and Interpretability: • Even though the system is end-to-end, Tesla can still prompt the model to output interpretable data: 3D occupancy, road boundaries, objects, signs, traffic lights, etc. • Natural language querying: ask the model why it made a certain decision. • These auxiliary predictions don’t drive the car but help engineers debug and ensure safety. Tesla’s Advanced Gaussian Splatting (3D Scene Modeling): • Tesla developed a custom, ultra-fast Gaussian splatting system to reconstruct 3D scenes from limited camera views. • Produces crisp, accurate 3D renderings even from few camera angles — far better than standard NeRF/splatting approaches. • Enables rapid visual debugging of the driving environment in 3D. Evaluation & World Models: • Evaluation is the hardest challenge: models may perform well offline but fail in real-world conditions. • Tesla builds balanced, diverse evaluation datasets focusing on edge cases — not just easy highway driving. Introduced a learned world simulator (neural network-generated video engine): • Can simulate 8 Tesla camera feeds simultaneously — fully synthetic. • Used for testing, training, and reinforcement learning. • Allows adversarial event injection (e.g., adding a pedestrian or vehicle cutting in). • Enables replaying past failures to verify new model improvements. • Can run in near real-time, letting testers “drive” inside a simulated world. What’s Next: • Scale robotaxi service globally. • Unlock full autonomy across the entire Tesla fleet. • Cybercab: next-gen 2-seat vehicle designed specifically for robotaxi use, targeting lowest transportation cost (cheaper than public transit). • Same neural networks will power Optimus humanoid robot. • The same video generation system is now being applied to Optimus. • The system can simulate and plan movement for robots, adapting easily to new forms. via the International Conference on Computer Vision (ICCV). Full presentation:

A new 30-minute presentation from Ashok Elluswamy, Tesla’s VP of AI, has been released, where he talks about FSD, AI and the team’s latest progress. Highlight from the presentation: • Tesla's vehicle fleet can provide 500 years of driving data every single day. Curse of Dimensionality: • 8 cameras at high frame rate = billions of tokens per 30 seconds of driving context. • Tesla must compress and extract the right correlations between sensory input and control actions. Data Advantage: • Tesla has access to a “Niagara Falls of data” — hundreds of years’ worth of collective fleet driving. • Uses smart data triggers to capture rare corner cases (e.g., complex intersections, unpredictable behavior). Quality and Efficiency: • Extracts only the essential data needed to train models efficiently. Debugging and Interpretability: • Even though the system is end-to-end, Tesla can still prompt the model to output interpretable data: 3D occupancy, road boundaries, objects, signs, traffic lights, etc. • Natural language querying: ask the model why it made a certain decision. • These auxiliary predictions don’t drive the car but help engineers debug and ensure safety. Tesla’s Advanced Gaussian Splatting (3D Scene Modeling): • Tesla developed a custom, ultra-fast Gaussian splatting system to reconstruct 3D scenes from limited camera views. • Produces crisp, accurate 3D renderings even from few camera angles — far better than standard NeRF/splatting approaches. • Enables rapid visual debugging of the driving environment in 3D. Evaluation & World Models: • Evaluation is the hardest challenge: models may perform well offline but fail in real-world conditions. • Tesla builds balanced, diverse evaluation datasets focusing on edge cases — not just easy highway driving. Introduced a learned world simulator (neural network-generated video engine): • Can simulate 8 Tesla camera feeds simultaneously — fully synthetic. • Used for testing, training, and reinforcement learning. • Allows adversarial event injection (e.g., adding a pedestrian or vehicle cutting in). • Enables replaying past failures to verify new model improvements. • Can run in near real-time, letting testers “drive” inside a simulated world. What’s Next: • Scale robotaxi service globally. • Unlock full autonomy across the entire Tesla fleet. • Cybercab: next-gen 2-seat vehicle designed specifically for robotaxi use, targeting lowest transportation cost (cheaper than public transit). • Same neural networks will power Optimus humanoid robot. • The same video generation system is now being applied to Optimus. • The system can simulate and plan movement for robots, adapting easily to new forms. via the International Conference on Computer Vision (ICCV). Full presentation:

Sawyer Merritt

1,286,614 次观看 • 8 个月前