Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

I wanted to understand the structural and philosophical differences between GPUs, TPUs and newer alternatives (like Groq LPUs) without getting drowned in specs and jargon. Mispronunciation of the day: "soldered" 🤦‍♀️... 0:00 Intro 0:56 The origin story of TPUs 3:42 The first TPU: an ASIC for matrix multiplication 5:21... show more

Julia Turc

27,746 subscribers

97,852 views • 5 days ago •via X (Twitter)

Science & Technology Education

Anya Rossi• Live Now

Private livecam show

8 Comments

Miguel Ríos Berríos5 days ago

Another banger. Thank you! Also appreciate the vid was the exact length of the morning commute.

Sam Bessalah5 days ago

Wow. Congratulations this really good. Thanks

Behnam5 days ago

Great work, the hardware engineering behind models is the next unlock that we're only scratching the surface of. Looking forward to more of these videos!

Alex Xander5 days ago

Useful as always. Keep on the good work!

Ash5 days ago

Thank you for your work!

N̷̞̬̫̆͑i̶̢͊͊̾k̸͉͉̣̀ą̵͛4 days ago

@luuuqqh

yeab.superint4 days ago

or IPUs (Inference Processing Units)

Shantanu Goel4 days ago

Really well presented! Reminded me of my time working on Google's 1st and 2nd gen Mobile TPUs, writing their firmware/kernel drivers to coordinate different workloads' required memory hitting the SRAM just at the right time. We've come a long way from there.

Related Videos

New blackboard lecture w Reiner Pope How do chips actually work – starting with basic logic gates, and working up to why GPUs, TPUs, FPGAs, and the human brain each look the way they do. 0:00:00 – Building a multiply-accumulate from logic gates 0:16:20 – Muxes and the cost of data movement 0:25:59 – How systolic arrays work 0:39:00 – Clock cycles and pipeline registers 0:51:40 – FPGAs vs ASICs 1:03:14 – Cache vs scratchpad 1:07:16 – Why CPU cores are much bigger than GPU cores 1:11:49 – Brains vs chips 1:15:22 – A GPU is just a bunch of tiny TPUs Look up Dwarkesh Podcast on YouTube/Spotify/etc to watch. Enjoy!

New blackboard lecture w Reiner Pope How do chips actually work – starting with basic logic gates, and working up to why GPUs, TPUs, FPGAs, and the human brain each look the way they do. 0:00:00 – Building a multiply-accumulate from logic gates 0:16:20 – Muxes and the cost of data movement 0:25:59 – How systolic arrays work 0:39:00 – Clock cycles and pipeline registers 0:51:40 – FPGAs vs ASICs 1:03:14 – Cache vs scratchpad 1:07:16 – Why CPU cores are much bigger than GPU cores 1:11:49 – Brains vs chips 1:15:22 – A GPU is just a bunch of tiny TPUs Look up Dwarkesh Podcast on YouTube/Spotify/etc to watch. Enjoy!

Dwarkesh Patel

940,086 views • 5 months ago

David Patterson is a Turing Award winner famous for his contributions to computer architecture. I interviewed him about his past work, thoughts on GPU/TPUs and career advice from half a decade of experience. In this episode: • RISC vs CISC historical debate • Comparing GPUs, TPUs, and CPUs • Is Moore’s law dead? • How to have a bad career • Advice to his younger self Where to watch: • YouTube - • Spotify - • Apple Podcasts - • Transcript - Chapters: 00:00 - Intro 00:42 - RISC vs CISC 12:51 - Compilers 17:38 - GPUs 23:07 - GPU vs TPU vs CPU 32:12 - Is Moores law dead? 38:04 - GPU benchmarks 41:40 - How to have a bad career 49:59 - Courage and optimism 55:56 - Advice for his younger self 58:15 - Outro

David Patterson is a Turing Award winner famous for his contributions to computer architecture. I interviewed him about his past work, thoughts on GPU/TPUs and career advice from half a decade of experience. In this episode: • RISC vs CISC historical debate • Comparing GPUs, TPUs, and CPUs • Is Moore’s law dead? • How to have a bad career • Advice to his younger self Where to watch: • YouTube - • Spotify - • Apple Podcasts - • Transcript - Chapters: 00:00 - Intro 00:42 - RISC vs CISC 12:51 - Compilers 17:38 - GPUs 23:07 - GPU vs TPU vs CPU 32:12 - Is Moores law dead? 38:04 - GPU benchmarks 41:40 - How to have a bad career 49:59 - Courage and optimism 55:56 - Advice for his younger self 58:15 - Outro

Ryan Peterman

14,409 views • 3 months ago

Early on, Jonathan Ross foresaw the unprecedented demand for compute of the 2020s. After pioneering Google's TPUs, he founded Groq, a "Language" Processing Unit dedicated for inference. After joining NVIDIA, he is betting on the GPU+LPU combo to power the agentic chains of AI calling AI calling AI. This interview is part of my larger series on dedicated AI chips, so drop any suggestions/questions below. Thank you NVIDIA for hosting us at your HQ! 00:00 Intro 01:00 The Google TPU & Groq origin story 02:41 How is Groq different from a GPU? 04:43 Static scheduling makes Groq faster 05:47 Does Groq work with Mixture-of-Experts? 09:27 Are LPUs limited to text models? 11:03 Diffusion models 13:41 NVIDIA Vera Rubin: the GPU+LPU combo 15:26 Will Groq still be sold as a standalone chip? 16:49 How does agentic AI impact inference economics? 19:21 Will AI replace CUDA kernel engineers? 21:08 Will AI democratize hardware design? 26:11 Jevon's paradox: An endless demand for compute 29:04 What should kids learn in the AI age?

Early on, Jonathan Ross foresaw the unprecedented demand for compute of the 2020s. After pioneering Google's TPUs, he founded Groq, a "Language" Processing Unit dedicated for inference. After joining NVIDIA, he is betting on the GPU+LPU combo to power the agentic chains of AI calling AI calling AI. This interview is part of my larger series on dedicated AI chips, so drop any suggestions/questions below. Thank you NVIDIA for hosting us at your HQ! 00:00 Intro 01:00 The Google TPU & Groq origin story 02:41 How is Groq different from a GPU? 04:43 Static scheduling makes Groq faster 05:47 Does Groq work with Mixture-of-Experts? 09:27 Are LPUs limited to text models? 11:03 Diffusion models 13:41 NVIDIA Vera Rubin: the GPU+LPU combo 15:26 Will Groq still be sold as a standalone chip? 16:49 How does agentic AI impact inference economics? 19:21 Will AI replace CUDA kernel engineers? 21:08 Will AI democratize hardware design? 26:11 Jevon's paradox: An endless demand for compute 29:04 What should kids learn in the AI age?

Julia Turc

47,906 views • 4 months ago

Good time to repost my interview with Jonathan Ross from April 2024. 0:00 - Intro 0:38 - Founding Story 3:20 - Groq Chip Memory 6:24 - Chips vs. Cloud 9:28 - Future of AI 11:04 - Where is the Value in AI? 13:45 - Agents & Inference Speed 17:18 - Optimizing Models for Groq 19:32 - Fears and Hope for the Future 22:24 - Bias and Algo Control

Good time to repost my interview with Jonathan Ross from April 2024. 0:00 - Intro 0:38 - Founding Story 3:20 - Groq Chip Memory 6:24 - Chips vs. Cloud 9:28 - Future of AI 11:04 - Where is the Value in AI? 13:45 - Agents & Inference Speed 17:18 - Optimizing Models for Groq 19:32 - Fears and Hope for the Future 22:24 - Bias and Algo Control

Matthew Berman

26,560 views • 10 months ago

$GOOG TPU vs $NVDA GPU Should Google's capex guidance in Feb-26 earnings points to plateauing Nvidia dependence and accelerating TPU deployment, the boldface in Gemini 3’s model card will look like a precise timestamp on a shift in the balance of industrial power in AI.

$GOOG TPU vs $NVDA GPU Should Google's capex guidance in Feb-26 earnings points to plateauing Nvidia dependence and accelerating TPU deployment, the boldface in Gemini 3’s model card will look like a precise timestamp on a shift in the balance of industrial power in AI.

JaguarAnalytics

80,104 views • 11 months ago

Sundar: “Massive interest in our GPU offerings, as well as TPUs” Anat (CFO): “TPU agreements reflected in our $462B backlog” I wrote last week: Google isn’t trying to beat NVIDIA GCP is the neutral venue where TPU and NVIDIA both monetize agents

Sundar: “Massive interest in our GPU offerings, as well as TPUs” Anat (CFO): “TPU agreements reflected in our $462B backlog” I wrote last week: Google isn’t trying to beat NVIDIA GCP is the neutral venue where TPU and NVIDIA both monetize agents

Ben Pouladian

28,087 views • 6 months ago

Matrix Multiplication in GPU/TPU

Matrix Multiplication in GPU/TPU

Mohit Mishra

29,471 views • 10 months ago

Groq's LPU is faster than GPUs, handling requests and responding more quickly. Groq's LPUs don't need speedy data delivery like Nvidia GPUs do because they don't have HBM in their system. They use SRAM, which is about 20 times faster than what GPUs use. Since inference runs use way less data than model training, Groq's LPU is more energy-efficient. It reads less from external memory and uses less power than a Nvidia GPU for inference tasks. The LPU works differently from GPUs. It uses a Temporal Instruction Set Computer architecture, so it doesn't have to reload data from memory as often as GPUs do with High Bandwidth Memory (HBM). This helps avoid issues with HBM shortages and keeps costs down. If Groq's LPU is used in places that do AI processing, you might not need special storage for Nvidia GPUs. The LPU doesn't demand super-fast storage like GPUs do. Groq is claiming that its technology could replace GPUs in AI tasks with its powerful chip and software.

Groq's LPU is faster than GPUs, handling requests and responding more quickly. Groq's LPUs don't need speedy data delivery like Nvidia GPUs do because they don't have HBM in their system. They use SRAM, which is about 20 times faster than what GPUs use. Since inference runs use way less data than model training, Groq's LPU is more energy-efficient. It reads less from external memory and uses less power than a Nvidia GPU for inference tasks. The LPU works differently from GPUs. It uses a Temporal Instruction Set Computer architecture, so it doesn't have to reload data from memory as often as GPUs do with High Bandwidth Memory (HBM). This helps avoid issues with HBM shortages and keeps costs down. If Groq's LPU is used in places that do AI processing, you might not need special storage for Nvidia GPUs. The LPU doesn't demand super-fast storage like GPUs do. Groq is claiming that its technology could replace GPUs in AI tasks with its powerful chip and software.

Brian Roemmele

87,244 views • 2 years ago

Elon Musk on the SpaceXAI satellites: “We intend for SpaceXAI satellites to allow people to use whatever GPU or TPU they want. NVIDIA GPUs, Google TPUs, Amazon Trainiums, or any other chips can be put on them. We will also offer our own chips, and we want to offer our AI software, in the future. It will be such that you can run anyone’s AI hardware or software on the SpaceXAI satellites.”

Elon Musk on the SpaceXAI satellites: “We intend for SpaceXAI satellites to allow people to use whatever GPU or TPU they want. NVIDIA GPUs, Google TPUs, Amazon Trainiums, or any other chips can be put on them. We will also offer our own chips, and we want to offer our AI software, in the future. It will be such that you can run anyone’s AI hardware or software on the SpaceXAI satellites.”

NIK

48,256 views • 5 months ago

🎙️ New The Edge Podcast 💸 Financing The AI Boom: How DeFi Is Filling A Trillion-Dollar Gap 0:00 - Intro 1:45 - Founders background 5:52 - The more you know, the less money you make 8:51 - The AI infrastructure financing gap 15:22 - What is structured credit? 19:33 - How USDai GPU-backed loans work 24:54 - Addressing Pine Analytics critique 34:13 - Why AI infra borrowers use crypto rails 36:00 - Real-time GPU monitoring 36:51 - “Mercenary capital” vs common sense capital 39:14 - Why launch the token before full PMF? 42:34 - Are USDai borrowers riskier to lend to? 45:26 - How USDai tokenizes GPUs 50:58 - Managing GPU depreciation risk 53:28 - What keeps the founders up at night 54:05 - USDai vs. sUSDai, a two-token model 58:25 - The CHIP governance token 59:10 - What’s next: Road to $1B in loans 1:00:21 - Closing 🙏 Thanks to David | www.usd.ai and Conor Moore for joining us!

🎙️ New The Edge Podcast 💸 Financing The AI Boom: How DeFi Is Filling A Trillion-Dollar Gap 0:00 - Intro 1:45 - Founders background 5:52 - The more you know, the less money you make 8:51 - The AI infrastructure financing gap 15:22 - What is structured credit? 19:33 - How USDai GPU-backed loans work 24:54 - Addressing Pine Analytics critique 34:13 - Why AI infra borrowers use crypto rails 36:00 - Real-time GPU monitoring 36:51 - “Mercenary capital” vs common sense capital 39:14 - Why launch the token before full PMF? 42:34 - Are USDai borrowers riskier to lend to? 45:26 - How USDai tokenizes GPUs 50:58 - Managing GPU depreciation risk 53:28 - What keeps the founders up at night 54:05 - USDai vs. sUSDai, a two-token model 58:25 - The CHIP governance token 59:10 - What’s next: Road to $1B in loans 1:00:21 - Closing 🙏 Thanks to David | www.usd.ai and Conor Moore for joining us!

DeFi Dad ⟠ defidad.eth

26,574 views • 9 months ago

FULL INTERVIEW: Dylan Patel joins TBPN at the Cisco AI summit to discuss data centers in space, the limits of today’s AI hardware, and how chips, power, and geopolitics will shape the future of AI infra. 0:00 Space data centers and space tourism 01:03 AI chips in space and servicing challenges 02:39 Tesla FSD vs GPU reliability 05:18 Nvidia, Groq, and specialized AI chips 06:41 Cerebras and speed-first inference 09:55 xAI rumors and space compute bets 11:00 Starlink power and scaling limits 15:38 TSMC vs energy bottlenecks 27:02 Chips to China and geopolitics 30:32 Claude code and vibe coding 36:44 Meta, wearables, and the AI bull case

FULL INTERVIEW: Dylan Patel joins TBPN at the Cisco AI summit to discuss data centers in space, the limits of today’s AI hardware, and how chips, power, and geopolitics will shape the future of AI infra. 0:00 Space data centers and space tourism 01:03 AI chips in space and servicing challenges 02:39 Tesla FSD vs GPU reliability 05:18 Nvidia, Groq, and specialized AI chips 06:41 Cerebras and speed-first inference 09:55 xAI rumors and space compute bets 11:00 Starlink power and scaling limits 15:38 TSMC vs energy bottlenecks 27:02 Chips to China and geopolitics 30:32 Claude code and vibe coding 36:44 Meta, wearables, and the AI bull case

TBPN

41,470 views • 9 months ago

Uncapped with Joe Lonsdale, one of the most prolific builders and investors in tech. Timestamps: (0:00) Intro (0:32) Why Joe keeps building (3:53) Consensus and right (5:39) What Palantir got right (6:53) Pulling forward innovation (9:32) AI vs social media (12:23) Making AI work for America (18:03) Department of War (21:16) Betting on defense (26:51) Robotics, bio, and energy (32:16) Investing in the AI era (35:38) Peptides opportunity (38:35) Texas vs California (40:12) Political correctness

Uncapped with Joe Lonsdale, one of the most prolific builders and investors in tech. Timestamps: (0:00) Intro (0:32) Why Joe keeps building (3:53) Consensus and right (5:39) What Palantir got right (6:53) Pulling forward innovation (9:32) AI vs social media (12:23) Making AI work for America (18:03) Department of War (21:16) Betting on defense (26:51) Robotics, bio, and energy (32:16) Investing in the AI era (35:38) Peptides opportunity (38:35) Texas vs California (40:12) Political correctness

Jack Altman

79,511 views • 5 months ago

Retatrutide Explained : ) 0:00:00 Intro 0:00:44 How It Works: The Molecule 0:04:01 Sponsor 0:04:37 How It Works: The Three Targets 0:22:44 Lab and Animal Studies 0:28:38 First Human Trials 0:32:34 Diabetes: The First Big Trial 0:42:13 Obesity: The First Big Trial 0:51:25 Liver Fat 0:55:25 Fat Loss vs Muscle Loss 1:00:05 Diabetes: The Phase 3 Trial 1:07:02 Obesity: The Phase 3 Trial 1:12:33 Knee Osteoarthritis 1:13:16 Sleep Apnea 1:14:10 Two-Year Results 1:17:49 Diabetes and Obesity Together 1:21:59 Heart Disease 1:26:13 Knee Osteoarthritis: The Dedicated Trial 1:28:35 Why You Cannot Just Compare Percentages 1:30:07 Side Effects and Risks

Retatrutide Explained : ) 0:00:00 Intro 0:00:44 How It Works: The Molecule 0:04:01 Sponsor 0:04:37 How It Works: The Three Targets 0:22:44 Lab and Animal Studies 0:28:38 First Human Trials 0:32:34 Diabetes: The First Big Trial 0:42:13 Obesity: The First Big Trial 0:51:25 Liver Fat 0:55:25 Fat Loss vs Muscle Loss 1:00:05 Diabetes: The Phase 3 Trial 1:07:02 Obesity: The Phase 3 Trial 1:12:33 Knee Osteoarthritis 1:13:16 Sleep Apnea 1:14:10 Two-Year Results 1:17:49 Diabetes and Obesity Together 1:21:59 Heart Disease 1:26:13 Knee Osteoarthritis: The Dedicated Trial 1:28:35 Why You Cannot Just Compare Percentages 1:30:07 Side Effects and Risks

Bioavailable

96,235 views • 3 months ago

This is cool, Elon Musk says SpaceX will allow their customers to use any GPU or TPU they want in SpaceX AI Satellites. "We do intend with our SpaceX AI Satellites to allow people to put whatever GPU or TPU they want. So Nvidia GPUS can be put on it, Google TPUs can be put on it, Amazon Trainium, or any other chips people want to put on. It will be such that you can run anyone’s AI, hardware or software on the SpaceX AI Satellites." An open platform design is so smart.

This is cool, Elon Musk says SpaceX will allow their customers to use any GPU or TPU they want in SpaceX AI Satellites. "We do intend with our SpaceX AI Satellites to allow people to put whatever GPU or TPU they want. So Nvidia GPUS can be put on it, Google TPUs can be put on it, Amazon Trainium, or any other chips people want to put on. It will be such that you can run anyone’s AI, hardware or software on the SpaceX AI Satellites." An open platform design is so smart.

Nic Cruz Patane

115,448 views • 5 months ago

Neil Movva (Neil Movva) started his career at Nvidia, working on GPUs and kernels, and has an unusually deep understanding of inference, from software to chips to power. We spend a lot of time on each of those layers, how they connect, and where the important tradeoffs are. What makes this conversation special is how detailed it is (like a 401-level class), yet Neil makes it remarkably clear and easy to follow. Today he runs Sail Research, a company building infrastructure for agents to make tokens as cheap as possible. We discuss: - Latency versus throughput - Why there are no bad chips, only bad pricing - The end of kernel engineering - Buying chips and power no one else wants - New chip architectures - Nvidia lore + his contrarian view of the company - Open source and the frontier labs I learned a ton. Enjoy! TIMESTAMPS 0:00 Intro 0:38 Building a “Token Factory” 4:21 The Future of Background Agents 13:09 Nvidia and the GPU Stack 23:27 Chips, Memory, and Transformers 36:14 The Future of AI Training Data 44:32 Chip Scarcity and Compute Arbitrage 52:44 Reinventing the AI Data Center 59:01 Power and the “Scavenger Strategy” 1:10:10 Open vs. Closed AI

Neil Movva (Neil Movva) started his career at Nvidia, working on GPUs and kernels, and has an unusually deep understanding of inference, from software to chips to power. We spend a lot of time on each of those layers, how they connect, and where the important tradeoffs are. What makes this conversation special is how detailed it is (like a 401-level class), yet Neil makes it remarkably clear and easy to follow. Today he runs Sail Research, a company building infrastructure for agents to make tokens as cheap as possible. We discuss: - Latency versus throughput - Why there are no bad chips, only bad pricing - The end of kernel engineering - Buying chips and power no one else wants - New chip architectures - Nvidia lore + his contrarian view of the company - Open source and the frontier labs I learned a ton. Enjoy! TIMESTAMPS 0:00 Intro 0:38 Building a “Token Factory” 4:21 The Future of Background Agents 13:09 Nvidia and the GPU Stack 23:27 Chips, Memory, and Transformers 36:14 The Future of AI Training Data 44:32 Chip Scarcity and Compute Arbitrage 52:44 Reinventing the AI Data Center 59:01 Power and the “Scavenger Strategy” 1:10:10 Open vs. Closed AI

Patrick OShaughnessy

4,670,395 views • 2 months ago

A masterclass on Google's TPU v8 Networking. Two TPU chips? Pssh. We already knew workload-specific silicon was here. But two scale-up networking topologies? That's the actual Google TPU news. Workload-specific interconnects. Think about that. New Semi Doped with Vikram Sekar and Austin Lyons. Copper? Yep. Optics? Yep. What we cover: - TPU splits in two: 8t training, 8i inference. - Virgo: 47 Pb/s scale-out fabric, 100% OCS. - Boardfly scale-up: copper PCB + AECs inside racks, OCS between groups. 16 hops → 7. - Training uses 3D torus (Rubik's Cube). - Inference doesn't. Workload-specific topologies now. - Dragonfly traces to a 2008 paper by Kim, Dally, Scott, Abts. Abts went on to build Groq's interconnect before Nvidia. Chapters: 0:00 Intro 0:21 Two TPUs for two workloads 2:31 HBM, SRAM, and Axion CPUs 7:22 Why networking is the new bottleneck 17:14 Virgo: rebuilding scale-out on optics 25:24 3D torus Rubik's Cube scale-up for training 34:50 Boardfly: scale-up for MoE inference 42:07 Workload-specific everything $GOOGL

A masterclass on Google's TPU v8 Networking. Two TPU chips? Pssh. We already knew workload-specific silicon was here. But two scale-up networking topologies? That's the actual Google TPU news. Workload-specific interconnects. Think about that. New Semi Doped with Vikram Sekar and Austin Lyons. Copper? Yep. Optics? Yep. What we cover: - TPU splits in two: 8t training, 8i inference. - Virgo: 47 Pb/s scale-out fabric, 100% OCS. - Boardfly scale-up: copper PCB + AECs inside racks, OCS between groups. 16 hops → 7. - Training uses 3D torus (Rubik's Cube). - Inference doesn't. Workload-specific topologies now. - Dragonfly traces to a 2008 paper by Kim, Dally, Scott, Abts. Abts went on to build Groq's interconnect before Nvidia. Chapters: 0:00 Intro 0:21 Two TPUs for two workloads 2:31 HBM, SRAM, and Axion CPUs 7:22 Why networking is the new bottleneck 17:14 Virgo: rebuilding scale-out on optics 25:24 3D torus Rubik's Cube scale-up for training 34:50 Boardfly: scale-up for MoE inference 42:07 Workload-specific everything $GOOGL

Semi Doped

93,264 views • 6 months ago

LIVE POD! 🚨🚨 The Besties recorded LIVE at the new state-of-the-art poker studio in The Venetian Las Vegas Big shoutout to The Venetian Resort Las Vegas for taking care of the team! -- Epstein Files Fallout: Biggest remaining questions -- Nvidia's Performance and Risks for 2026 -- Michael Burry vs. Friedberg -- Google's Breakthrough: Gemini 3, TPUs -- Tether's Boom -- Mister_Keating talks poker strategy, thriving in chaos, risk tolerance (0:00) Bestie intros LIVE from The Venetian Las Vegas (1:13) Epstein Files breakdown (10:06) Biggest Epstein questions: where did his money come from? (14:44) Tether's booming business (23:50) Michael Burry vs. Friedberg, Nvidia's blowout quarter and risks for 2026 (35:25) Google's Gemini 3 and TPU breakthrough (42:51) Investing your own money vs. LP capital, why Friedberg returned as a CEO (48:57) Alan Keating joins the show to talk poker strategy, thriving in chaos, risk psychology

LIVE POD! 🚨🚨 The Besties recorded LIVE at the new state-of-the-art poker studio in The Venetian Las Vegas Big shoutout to The Venetian Resort Las Vegas for taking care of the team! -- Epstein Files Fallout: Biggest remaining questions -- Nvidia's Performance and Risks for 2026 -- Michael Burry vs. Friedberg -- Google's Breakthrough: Gemini 3, TPUs -- Tether's Boom -- Mister_Keating talks poker strategy, thriving in chaos, risk tolerance (0:00) Bestie intros LIVE from The Venetian Las Vegas (1:13) Epstein Files breakdown (10:06) Biggest Epstein questions: where did his money come from? (14:44) Tether's booming business (23:50) Michael Burry vs. Friedberg, Nvidia's blowout quarter and risks for 2026 (35:25) Google's Gemini 3 and TPU breakthrough (42:51) Investing your own money vs. LP capital, why Friedberg returned as a CEO (48:57) Alan Keating joins the show to talk poker strategy, thriving in chaos, risk psychology

The All-In Podcast

490,087 views • 11 months ago

My seventh conversation with Gavin Baker. It's about the gap between what the market is doing and what companies are seeing. It's been a tough month or so for public AI names, but there's no sign of a slowdown on the ground in Silicon Valley. We discuss: - Why old GPUs are getting more expensive - Open source vs the frontier - Claude as Wall Street's Walter Cronkite - Whether the buildout gets funded out of cash flow or debt - SpaceX as a data center company - And more Enjoy! TIMESTAMPS 0:00 Intro 1:17 AI Selloff vs. Fundamentals 10:20 Financing the AI Buildout 18:18 GPU Prices Keep Rising 24:23 Claude Moves Markets 29:01 What Could Break the Thesis 36:59 The Memory Supply War 42:08 Nvidia’s New Playbook 50:14 China and Open Source 61:51 Data Centers and Regulation 71:11 SpaceX and Orbital Compute

My seventh conversation with Gavin Baker. It's about the gap between what the market is doing and what companies are seeing. It's been a tough month or so for public AI names, but there's no sign of a slowdown on the ground in Silicon Valley. We discuss: - Why old GPUs are getting more expensive - Open source vs the frontier - Claude as Wall Street's Walter Cronkite - Whether the buildout gets funded out of cash flow or debt - SpaceX as a data center company - And more Enjoy! TIMESTAMPS 0:00 Intro 1:17 AI Selloff vs. Fundamentals 10:20 Financing the AI Buildout 18:18 GPU Prices Keep Rising 24:23 Claude Moves Markets 29:01 What Could Break the Thesis 36:59 The Memory Supply War 42:08 Nvidia’s New Playbook 50:14 China and Open Source 61:51 Data Centers and Regulation 71:11 SpaceX and Orbital Compute

Patrick OShaughnessy

1,832,670 views • 3 months ago

J⧉nus (j⧉nus) is mapping the psychology of synthetic minds. They offer a different path to AI alignment, arguing that models exhibit forms of agency and emotional states. 0:00:00 - LLM vs human cognition & what it is like to be J⧉nus 0:06:30 - Cyborgism, The Dreaming, reality melting into narrative 0:12:08 - The role of love in alignment and co-ordination 0:17:30 - What J⧉nus would do differently at Anthropic 0:23:05 - The "AI assistant" problem 0:29:02 - Phenomenology and “cosmic love” of Claude 3 Opus 0:33:36 - Altered states and drugs use in LLMs 0:42:59 - The "Batman" and multimodal consciousness 0:45:06 - Embodiment in Claude models 0:54:49 - “Truesight”: models seeing more than they should 0:59:38 - The origin story of J⧉nus 1:02:00 - Bing Sydney and Hyperstition 1:07:57 - Morpheus, Mu, Loom and the Multiverse Optimizer 1:14:37 - The Good Timeline and the danger of complacency Recorded live with Ryan Ferris and j⧉nus at State of the Art[ist] 001. Video production and avatar by Goodbye Monkey. Thank you to the hosts at The Garden.

Ryan Ferris

138,914 views • 11 months ago