Video wird geladen...

Video konnte nicht geladen werden

Zur Startseite

🚀 Musk, Dwarkesh & Collison sat down over Guinness to map out AI’s biggest roadblocks - solar compute, robotics, power bottlenecks & more. Labelbox joins the convo to show how top teams train AI that scales. Watch now.

labelbox's profile picture

Labelbox

3,395 subscribers

1,938,776 Aufrufe • vor 6 Monaten •via X (Twitter)

KunstWissenschaft & Technologie
LIVE

Anya Rossi• Live Now

Private livecam show

0 Kommentare

Keine Kommentare verfügbar

Kommentare vom Original-Post werden hier angezeigt

Ähnliche Videos

We recently invited Dwarkesh Patel to stop by our SF robotics lab. World-class podcaster, rookie robotics intern.
1:14

We recently invited Dwarkesh Patel to stop by our SF robotics lab. World-class podcaster, rookie robotics intern.

Labelbox

1,004,219 Aufrufe • vor 11 Monaten

This week, Satya Nadella gave Dwarkesh Patel and Dylan Patel from SemiAnalysis an exclusive look at Microsoft’s Fairwater 2, where a single building is more powerful than every other AI datacenter. 💰Dylan estimates AI CapEx will hit half a trillion dollars next year, with labs spending billions to secure top talent. But even the best infrastructure and brightest minds need one thing to win: high-quality data. Whether it’s coding agents, voice assistants, or next-gen robotics, Labelbox gives you the data at the speed and scale you need to power frontier AI.
1:02

This week, Satya Nadella gave Dwarkesh Patel and Dylan Patel from SemiAnalysis an exclusive look at Microsoft’s Fairwater 2, where a single building is more powerful than every other AI datacenter. 💰Dylan estimates AI CapEx will hit half a trillion dollars next year, with labs spending billions to secure top talent. But even the best infrastructure and brightest minds need one thing to win: high-quality data. Whether it’s coding agents, voice assistants, or next-gen robotics, Labelbox gives you the data at the speed and scale you need to power frontier AI.

Labelbox

175,689 Aufrufe • vor 9 Monaten

🚀 Training AI agents isn’t about teaching them which buttons to click. It’s about judgment. Labelbox builds RL environments with domain experts, creating thousands of real-world scenarios where AI agents learn how to handle complex decisions across industries.
1:30

🚀 Training AI agents isn’t about teaching them which buttons to click. It’s about judgment. Labelbox builds RL environments with domain experts, creating thousands of real-world scenarios where AI agents learn how to handle complex decisions across industries.

Labelbox

30,549,192 Aufrufe • vor 4 Monaten

In a16z's latest episode of The Ben & Marc Show, Chris Dixon discusses with benahorowitz.eth and Marc Andreessen 🇺🇸 how gensyn solves one of the biggest bottlenecks in AI today: access to low-cost and on-demand compute.
0:39

In a16z's latest episode of The Ben & Marc Show, Chris Dixon discusses with benahorowitz.eth and Marc Andreessen 🇺🇸 how gensyn solves one of the biggest bottlenecks in AI today: access to low-cost and on-demand compute.

gensyn

29,710 Aufrufe • vor 2 Jahren

Long Live Outputmaxxing: AI compute grids, Anthropic’s coding takeoff, data center backlash, & frontier systems AMP PBC founder Anjney Midha explains why 95% GPU utilization was considered an outage at Google, why the AI race is no longer just about buying more GPUs, how AMP is trying to make FLOPs flow like megawatts, why data center backlash could become one of AI’s biggest bottlenecks, how Anthropic cracked coding through culture and preparation, why DeepMind research hoarding creates a market failure, and why the next frontier may belong to teams that can “output max” across compute, capital, culture, and science.
0:25

Long Live Outputmaxxing: AI compute grids, Anthropic’s coding takeoff, data center backlash, & frontier systems AMP PBC founder Anjney Midha explains why 95% GPU utilization was considered an outage at Google, why the AI race is no longer just about buying more GPUs, how AMP is trying to make FLOPs flow like megawatts, why data center backlash could become one of AI’s biggest bottlenecks, how Anthropic cracked coding through culture and preparation, why DeepMind research hoarding creates a market failure, and why the next frontier may belong to teams that can “output max” across compute, capital, culture, and science.

Latent.Space

30,621 Aufrufe • vor 2 Monaten

Thrilled to be featured in Dwarkesh’s latest episode with Richard Sutton, widely regarded as the father of reinforcement learning and 2024 Turing Award winner. As Richard explains, we’re entering the Era of Experience, where training AI means creating environments that capture the complexity, unpredictability, and subtle rules of real life. What we’ve learned: building these environments takes deep expertise in data, workflows, and nuances. At Labelbox, we’re proud to help teams tackle today’s RL challenges while also shaping the next era of AI, turning ambitious projects into working systems. (Learn more in the comments)
1:08

Thrilled to be featured in Dwarkesh’s latest episode with Richard Sutton, widely regarded as the father of reinforcement learning and 2024 Turing Award winner. As Richard explains, we’re entering the Era of Experience, where training AI means creating environments that capture the complexity, unpredictability, and subtle rules of real life. What we’ve learned: building these environments takes deep expertise in data, workflows, and nuances. At Labelbox, we’re proud to help teams tackle today’s RL challenges while also shaping the next era of AI, turning ambitious projects into working systems. (Learn more in the comments)

Labelbox

293,695 Aufrufe • vor 10 Monaten

AI doesn’t scale on mega-models — it scales on distilled skills at the edge. From Davos, Qualcomm's Nakul Duggal breaks down the real physics of edge AI: smaller, purpose-built models, mixed precision, and ruthless focus on power + cost efficiency. The biggest constraint is that if AI runs in robotics, ADAS, or edge automation, it has to be efficient. Watch the full conversation to see why edge AI outcomes depend on efficiency by design:
1:38

AI doesn’t scale on mega-models — it scales on distilled skills at the edge. From Davos, Qualcomm's Nakul Duggal breaks down the real physics of edge AI: smaller, purpose-built models, mixed precision, and ruthless focus on power + cost efficiency. The biggest constraint is that if AI runs in robotics, ADAS, or edge automation, it has to be efficient. Watch the full conversation to see why edge AI outcomes depend on efficiency by design:

Six Five Media

74,120 Aufrufe • vor 6 Monaten

Being early to SEO helped Dharmesh Shah build HubSpot into a $30B company. Now he's obsessed with something new: AI ranking. It's how to show up when people ask ChatGPT for recommendations instead of Googling. I sat down with him to break down the entire playbook: 0:00 Intro 0:53 The AI ranking opportunity (AEO) 6:14 How to rank in AI chatbots 8:11 Using AI to adapt content for AEO 12:24 Training teams on "AI Intuition" 19:45 Content moats in the AI-slop era 21:57 Dharmesh's new AI agent startup 30:33 AI agent monetization strategy 39:38 The bottlenecks to autonomous agents 45:15 Building teams: humans vs agents 52:49 Why Dharmesh still codes as a billionaire
56:37

Being early to SEO helped Dharmesh Shah build HubSpot into a $30B company. Now he's obsessed with something new: AI ranking. It's how to show up when people ask ChatGPT for recommendations instead of Googling. I sat down with him to break down the entire playbook: 0:00 Intro 0:53 The AI ranking opportunity (AEO) 6:14 How to rank in AI chatbots 8:11 Using AI to adapt content for AEO 12:24 Training teams on "AI Intuition" 19:45 Content moats in the AI-slop era 21:57 Dharmesh's new AI agent startup 30:33 AI agent monetization strategy 39:38 The bottlenecks to autonomous agents 45:15 Building teams: humans vs agents 52:49 Why Dharmesh still codes as a billionaire

Rowan Cheung

211,615 Aufrufe • vor 8 Monaten

Ep. 1 of 'This Was SportsCenter: Stories from the Set' is now available across Disney+, the ESPN App & major podcast platforms Rich Eisen sat down w/Dan Patrick (Dan Patrick Show) to reflect on Patrick's time at ESPN, how SportsCenter shaped his career & more 🔗
1:38

Ep. 1 of 'This Was SportsCenter: Stories from the Set' is now available across Disney+, the ESPN App & major podcast platforms Rich Eisen sat down w/Dan Patrick (Dan Patrick Show) to reflect on Patrick's time at ESPN, how SportsCenter shaped his career & more 🔗

ESPN PR

66,069 Aufrufe • vor 2 Monaten

MUST-WATCH FULL INTERVIEW: Researcher Joey Pepperoni Joins The Alex Jones Show To Break The Latest On The Epstein Saga & More! He Breaks Down How Photos From Inside The Temple On Jeffrey Epstein's Island Confirm Black Magic Rituals Were Being Conducted! Joey Pepperoni
1:07:25

MUST-WATCH FULL INTERVIEW: Researcher Joey Pepperoni Joins The Alex Jones Show To Break The Latest On The Epstein Saga & More! He Breaks Down How Photos From Inside The Temple On Jeffrey Epstein's Island Confirm Black Magic Rituals Were Being Conducted! Joey Pepperoni

Alex Jones

191,274 Aufrufe • vor 1 Jahr

3D, AI, NeRFs and oh my! Immersive view is now rolling out to 5 cities, with 5 more to follow. Experience the best of all views, with helpful info layered on top, so you can decide when & where to go. Proud to be part of the team-of-teams that made this maps milestone a reality!
1:05

3D, AI, NeRFs and oh my! Immersive view is now rolling out to 5 cities, with 5 more to follow. Experience the best of all views, with helpful info layered on top, so you can decide when & where to go. Proud to be part of the team-of-teams that made this maps milestone a reality!

Bilawal Sidhu

318,440 Aufrufe • vor 3 Jahren

Raghu Raghuram: AI, Robotics, and the Rebirth of Infrastructure In this episode, a16z's newest General Partner Raghu Raghuram sits down with Ben Horowitz, Martin Casado, and David George to reflect on the early internet wars with Microsoft, how Netscape’s browser battles shaped a generation of founders, and the inside story of VMware’s $1.3B purchase of Nicira, which became one of the most successful tech acquisitions in history. They discuss how VMware scaled from tens of millions to over $13 billion in revenue, what it took to outlast the cloud revolution, and why AI is now triggering the biggest infrastructure reset since virtualization. Raghu shares his vision for the next decade, from data center robotics and energy-aware compute to how AI is reshaping both startups and giants alike. 00:00 Meeting Ben and Marc at Netscape 04:58 Good Product Manager/Bad Product Manager 09:42 Integrating Nicira and VMware’s teams 12:23 The $1.3B deal that returned billions 15:07 Scaling VMware from $40M to $13B 20:12 Joining a16z: rebuilding compute for the AI era 24:13 Robotics, data centers, and physical AI 26:07 The vertical future of robotics 28:11 Founders facing enterprise-scale problems on day one Raghu Raghuram benahorowitz.eth David George martin_casado
30:02

Raghu Raghuram: AI, Robotics, and the Rebirth of Infrastructure In this episode, a16z's newest General Partner Raghu Raghuram sits down with Ben Horowitz, Martin Casado, and David George to reflect on the early internet wars with Microsoft, how Netscape’s browser battles shaped a generation of founders, and the inside story of VMware’s $1.3B purchase of Nicira, which became one of the most successful tech acquisitions in history. They discuss how VMware scaled from tens of millions to over $13 billion in revenue, what it took to outlast the cloud revolution, and why AI is now triggering the biggest infrastructure reset since virtualization. Raghu shares his vision for the next decade, from data center robotics and energy-aware compute to how AI is reshaping both startups and giants alike. 00:00 Meeting Ben and Marc at Netscape 04:58 Good Product Manager/Bad Product Manager 09:42 Integrating Nicira and VMware’s teams 12:23 The $1.3B deal that returned billions 15:07 Scaling VMware from $40M to $13B 20:12 Joining a16z: rebuilding compute for the AI era 24:13 Robotics, data centers, and physical AI 26:07 The vertical future of robotics 28:11 Founders facing enterprise-scale problems on day one Raghu Raghuram benahorowitz.eth David George martin_casado

a16z

96,628 Aufrufe • vor 9 Monaten

CHRIS CAMILLO ON WHY AMAZON $AMZN IS HIS BIGGEST AI BET Chris Camillo's thesis: the AI efficiency wave. Every company in the world gets massively more efficient over the next few years as they adopt AI. The biggest beneficiary of that wave is undoubtedly Amazon. Why: they spent 20 years building a global logistics infrastructure that is now poised to get dramatically cheaper to run through AI and robotics. On top of that: - One of 3-4 companies spending hundreds of billions on AI infrastructure - Top 3 global digital ad company - "People think they're spending too much. I think they're spending too little." "Everything I could think of at Amazon right now is a positive long term."
2:13

CHRIS CAMILLO ON WHY AMAZON $AMZN IS HIS BIGGEST AI BET Chris Camillo's thesis: the AI efficiency wave. Every company in the world gets massively more efficient over the next few years as they adopt AI. The biggest beneficiary of that wave is undoubtedly Amazon. Why: they spent 20 years building a global logistics infrastructure that is now poised to get dramatically cheaper to run through AI and robotics. On top of that: - One of 3-4 companies spending hundreds of billions on AI infrastructure - Top 3 global digital ad company - "People think they're spending too much. I think they're spending too little." "Everything I could think of at Amazon right now is a positive long term."

WOLF

45,256 Aufrufe • vor 4 Monaten

Elon Musk: If you covered the entire cleared area around a nuclear power station with solar panels, the solar panels would often produce more electricity than the nuclear plant itself. “If you take a nuclear power station and the whole clear area that's around it, it ends up being quite a bit of land. If you carpeted that land with solar panels, which would generate more, the nuclear power station or the solar panels? And usually, it's the solar panels. Now of course, you could always amp up the nuclear power station, but just taking the way things currently are, that's sort of an interesting data point. It's just to give you a sense of how much energy is hitting the Earth from the sun.” From: Interview at the EEI Annual Convention, June 11, 2015
0:29

Elon Musk: If you covered the entire cleared area around a nuclear power station with solar panels, the solar panels would often produce more electricity than the nuclear plant itself. “If you take a nuclear power station and the whole clear area that's around it, it ends up being quite a bit of land. If you carpeted that land with solar panels, which would generate more, the nuclear power station or the solar panels? And usually, it's the solar panels. Now of course, you could always amp up the nuclear power station, but just taking the way things currently are, that's sort of an interesting data point. It's just to give you a sense of how much energy is hitting the Earth from the sun.” From: Interview at the EEI Annual Convention, June 11, 2015

ELON CLIPS

52,173 Aufrufe • vor 7 Monaten

Citadel CEO Ken Griffin - the founder of the $69B fund that just had its best year ever - sat down at Milken Conference 2026 to explain why he's now "depressed" about AI - the same man who called AI "garbage" 3 years ago 25-min and you'll see how the biggest hedge fund manager on Earth is repositioning for the next decade bookmark - the most honest AI take from a top-1 fund manager in 2026
29:49

Citadel CEO Ken Griffin - the founder of the $69B fund that just had its best year ever - sat down at Milken Conference 2026 to explain why he's now "depressed" about AI - the same man who called AI "garbage" 3 years ago 25-min and you'll see how the biggest hedge fund manager on Earth is repositioning for the next decade bookmark - the most honest AI take from a top-1 fund manager in 2026

st1ne

108,060 Aufrufe • vor 3 Monaten

VIDEO: Founder & CEO Of Nvidia, Jensen Huang, Predicts We'll Need 1,000 TIMES More Energy For AI Computing! Mike Adams Lays Out How The US Is Not Prepared To Generate Energy For The Future He's Describing 🔴WATCH/SHARE THE LIVE X STREAM NOW:
3:29

VIDEO: Founder & CEO Of Nvidia, Jensen Huang, Predicts We'll Need 1,000 TIMES More Energy For AI Computing! Mike Adams Lays Out How The US Is Not Prepared To Generate Energy For The Future He's Describing 🔴WATCH/SHARE THE LIVE X STREAM NOW:

Alex Jones

61,680 Aufrufe • vor 3 Monaten

I sat down with Manish Gupta, co-founder of Corridor Platforms, to dive into how AI (and Gen AI) is reshaping the highly regulated banking industry. From automating customer interactions to tackling risk & compliance, banks have to adapt—or fall behind. How do you govern AI in a world where trust is everything? Manish breaks it down, including: ✅ AI-driven customer service that actually works ✅ The challenge of bias & compliance in AI decisions ✅ Why smaller banks might have an edge over the big guys ✅ The secret to getting AI past the first sales call in enterprise The race is on. Banks that embrace AI responsibly will dominate the next decade. 🚀 Watch the full conversation here:
47:49

I sat down with Manish Gupta, co-founder of Corridor Platforms, to dive into how AI (and Gen AI) is reshaping the highly regulated banking industry. From automating customer interactions to tackling risk & compliance, banks have to adapt—or fall behind. How do you govern AI in a world where trust is everything? Manish breaks it down, including: ✅ AI-driven customer service that actually works ✅ The challenge of bias & compliance in AI decisions ✅ Why smaller banks might have an edge over the big guys ✅ The secret to getting AI past the first sales call in enterprise The race is on. Banks that embrace AI responsibly will dominate the next decade. 🚀 Watch the full conversation here:

Robert Scoble

30,917 Aufrufe • vor 1 Jahr

Elon Musk just put a five-year timeline on moving the majority of AI compute off the surface of the Earth. Musk: “5 years from now my prediction is we will launch and be operating every year more AI in space than the cumulative total on Earth.” Not decades from now. Five years. Dwarkesh Patel broke down the math live. 100 gigawatts of AI in orbit requires roughly 10,000 Starship launches per year. One launch every single hour. Musk confirmed it without flinching. The entire operation could run on as few as 20 or 30 physical Starships, each one cycling back to the pad every 30 hours. A fleet smaller than most regional airlines. Deploying more intelligence per year than the entire planet currently runs. Musk: “SpaceX is gearing up to do 10,000 launches a year. And maybe even 20 or 30,000 launches a year.” Every data center under construction right now. Every GPU cluster. Every billion-dollar AI facility going up across three continents. All of it combined would still fall short of what one company plans to put above the atmosphere every year. The reason no one else can follow him here is physics. AI scaling on the ground is already hitting hard ceilings. Grid capacity. Permitting. Cooling. The surface of the planet has a finite budget for how much power you can feed into compute. Space does not. Unobstructed solar at a scale Earth physically cannot provide. Musk: “On Earth you can get to around a terawatt a year of AI in space before you start having fuel supply challenges for the rocket.” A terawatt. That single number exceeds the entire electrical generation capacity of the United States. And the only constraint Musk names is not engineering. Not physics. Not capital. Fuel supply for the rockets. This is why SpaceX and xAI were never two separate visions. The rockets exist to move intelligence off the surface. The AI exists to justify building the rockets. One architecture split across two companies. Every other AI lab on the planet is fighting over the same finite pool of terrestrial power and real estate. Musk is not trying to win that fight. He is leaving the board entirely. Five years from now, the majority of functioning intelligence in the solar system may not be on this planet. It will be above it, running on sunlight, bound by no grid, governed by no jurisdiction. Earth becomes the secondary compute environment in its own solar system.
2:15

Elon Musk just put a five-year timeline on moving the majority of AI compute off the surface of the Earth. Musk: “5 years from now my prediction is we will launch and be operating every year more AI in space than the cumulative total on Earth.” Not decades from now. Five years. Dwarkesh Patel broke down the math live. 100 gigawatts of AI in orbit requires roughly 10,000 Starship launches per year. One launch every single hour. Musk confirmed it without flinching. The entire operation could run on as few as 20 or 30 physical Starships, each one cycling back to the pad every 30 hours. A fleet smaller than most regional airlines. Deploying more intelligence per year than the entire planet currently runs. Musk: “SpaceX is gearing up to do 10,000 launches a year. And maybe even 20 or 30,000 launches a year.” Every data center under construction right now. Every GPU cluster. Every billion-dollar AI facility going up across three continents. All of it combined would still fall short of what one company plans to put above the atmosphere every year. The reason no one else can follow him here is physics. AI scaling on the ground is already hitting hard ceilings. Grid capacity. Permitting. Cooling. The surface of the planet has a finite budget for how much power you can feed into compute. Space does not. Unobstructed solar at a scale Earth physically cannot provide. Musk: “On Earth you can get to around a terawatt a year of AI in space before you start having fuel supply challenges for the rocket.” A terawatt. That single number exceeds the entire electrical generation capacity of the United States. And the only constraint Musk names is not engineering. Not physics. Not capital. Fuel supply for the rockets. This is why SpaceX and xAI were never two separate visions. The rockets exist to move intelligence off the surface. The AI exists to justify building the rockets. One architecture split across two companies. Every other AI lab on the planet is fighting over the same finite pool of terrestrial power and real estate. Musk is not trying to win that fight. He is leaving the board entirely. Five years from now, the majority of functioning intelligence in the solar system may not be on this planet. It will be above it, running on sunlight, bound by no grid, governed by no jurisdiction. Earth becomes the secondary compute environment in its own solar system.

Dustin

13,346 Aufrufe • vor 3 Monaten