Loading video...

Video Failed to Load

Go Home

Introducing The AI CUDA Engineer: An agentic AI system that automates the production of highly optimized CUDA kernels. The AI CUDA Engineer can produce highly optimized CUDA kernels, reaching 10-100x speedup over common machine learning operations in PyTorch. Our system is also able to produce highly optimized CUDA kernels...

SakanaAILabs's profile picture

Sakana AI

72,184 subscribers

1,149,274 views • 1 year ago •via X (Twitter)

Science & TechnologyEducation
LIVE

Anya Rossi• Live Now

Private livecam show

10 Comments

vittorio's profile picture
vittorio1 year ago

japan bros are back again

Globant's profile picture
Globant1 year ago

🚀 Agentic AI Systems are changing what AI can do by having the power to act independently. Unlike traditional AI, which needs constant human supervision, these systems can operate more autonomously. Learn how this shift is leading to smarter solutions that can transform industries.➡️ #TechTrends2025

ludwig's profile picture
ludwig1 year ago

I’m going to sleep if I wake up to this having 1M+ views I will read the paper tomorrow morning else pls give me a vibe check chat

Bing Xu's profile picture
Bing Xu1 year ago

I quickly take a look of their report on phone, there are a few misleading parts: 1. Torch C++ code is not CUDA kernel, it is calling CUDNN under hood. 2. The highlighted example Conv3D GroupNorm, conv code is not generated at all. The speedup doesn’t make sense if numerical is wrong. 3. It claims wmma can be faster than PyTorch (CUBLAS), is definitely wrong. Probably benchmark error.

main's profile picture
main1 year ago

isn't there clearly something wrong with level_1->15_Matmul_for_lower_triangular_matrices? claimed 152.9x speedup for the kernel on the left over the code on the right. really?

Viraat's profile picture
Viraat1 year ago

Hey - wondering if you all are only working with large enterprises right now. If not, we’d love to chat! This would be extremely useful to us - we’re building low-bit models to run efficiently on Jetsons. Generating optimized CUDA code for these would be a game-changer!

aizk ✡️'s profile picture
aizk ✡️1 year ago

A slow 14 seconds in AI developments

Dan Mac's profile picture
Dan Mac1 year ago

guys seriously I can't take it anymore need to slow down

Kristof's profile picture
Kristof1 year ago

Japan is back

Dr Futuro - e/acc's profile picture
Dr Futuro - e/acc1 year ago

Wow, Japan is back! 🇯🇵

Related Videos

🎉 Stoked to share The AI CUDA Engineer 👷 - our end-to-end approach for automating the design and optimization of CUDA Kernels using agentic systems. Blog 📰: Paper 📜: WebUI 📈: Dataset 💽: Awesome team work done with Aaditya Prasad 🇺🇸, Suuun, Maxence Faldor, Yujin Tang, hardmaru 🤗
3:40

🎉 Stoked to share The AI CUDA Engineer 👷 - our end-to-end approach for automating the design and optimization of CUDA Kernels using agentic systems. Blog 📰: Paper 📜: WebUI 📈: Dataset 💽: Awesome team work done with Aaditya Prasad 🇺🇸, Suuun, Maxence Faldor, Yujin Tang, hardmaru 🤗

Robert Lange

42,174 views • 1 year ago

Luminal ( is creating PyTorch for Production – an ML compiler that generates blazingly fast CUDA kernels and makes deploying to production one line of code. Congrats on the launch, Jake Stevens, Joe Fioti, and Matthew Gunton!
1:53

Luminal ( is creating PyTorch for Production – an ML compiler that generates blazingly fast CUDA kernels and makes deploying to production one line of code. Congrats on the launch, Jake Stevens, Joe Fioti, and Matthew Gunton!

Y Combinator

98,496 views • 10 months ago

AI Coding Agent for Hardware Optimized Code Diana AI hardware is still constrained by software. However, with reasoning models like Deepseek R1 or OpenAI o1 and o3, AI could generate hardware-optimized code that rivals—or surpasses—human CUDA code.
1:01

AI Coding Agent for Hardware Optimized Code Diana AI hardware is still constrained by software. However, with reasoning models like Deepseek R1 or OpenAI o1 and o3, AI could generate hardware-optimized code that rivals—or surpasses—human CUDA code.

Y Combinator

60,089 views • 1 year ago

🎉 "This is the 20th anniversary of CUDA. We have been working on this architecture for 20 years ... to now have built up hundreds of millions of GPUs and computing systems around the world that run CUDA."
1:34

🎉 "This is the 20th anniversary of CUDA. We have been working on this architecture for 20 years ... to now have built up hundreds of millions of GPUs and computing systems around the world that run CUDA."

NVIDIA HPC Developer

27,642 views • 2 months ago

Mark my words… The future of AI will not be written in CUDA! [I love this video generated by Higgsfield AI 🧩 for the @amd ROCm meetup this week]
0:55

Mark my words… The future of AI will not be written in CUDA! [I love this video generated by Higgsfield AI 🧩 for the @amd ROCm meetup this week]

Jeff Tatarchuk

12,132 views • 1 year ago

We are teaming up with @Databricks to supercharge AI workflows. ⚡ #CUDA will soon be at the heart of Databricks' Data Intelligence Platform, starting with AI-accelerated Photon, delivering improved speed and efficiency for customers’ data warehousing and analytics workflows.
0:38

We are teaming up with @Databricks to supercharge AI workflows. ⚡ #CUDA will soon be at the heart of Databricks' Data Intelligence Platform, starting with AI-accelerated Photon, delivering improved speed and efficiency for customers’ data warehousing and analytics workflows.

NVIDIA AI Developer

22,927 views • 1 year ago

in the lectures below, i hold your hand through low-level LLM systems engineering. it includes everything up to TODAY! 1) pytorch tensors 2) large matmul on cpu vs gpu 3) JAX (and why xAI uses it instead of pytorch) 4) raw cuda kernels and global threading indexing 5) triton design philosophy and softmax example 6) HIP kernels 7) mapping out the ENTIRE ecosystem + differences between CUDA and ROCm/HIP (BLAS, FFT, DNN) 8) cutlass and cute-dsl 9) pretraining, finetuning, rl, unsloth, axolotl, megatron-lm, deepspeed, nanogpt, nanochat 10) training vs inference, inference serving problems, throughput vs latency vs concurrency scaling, vllm, sglang, tensorrt-llm, tensorrt, llama.cpp, exllamav2, exllamav3, benchmark comparisons 11) projects/companies using llms to generate SOTA cuda/triton kernels 12) luminal inference 13) mojo/modular/max
2:23:25

in the lectures below, i hold your hand through low-level LLM systems engineering. it includes everything up to TODAY! 1) pytorch tensors 2) large matmul on cpu vs gpu 3) JAX (and why xAI uses it instead of pytorch) 4) raw cuda kernels and global threading indexing 5) triton design philosophy and softmax example 6) HIP kernels 7) mapping out the ENTIRE ecosystem + differences between CUDA and ROCm/HIP (BLAS, FFT, DNN) 8) cutlass and cute-dsl 9) pretraining, finetuning, rl, unsloth, axolotl, megatron-lm, deepspeed, nanogpt, nanochat 10) training vs inference, inference serving problems, throughput vs latency vs concurrency scaling, vllm, sglang, tensorrt-llm, tensorrt, llama.cpp, exllamav2, exllamav3, benchmark comparisons 11) projects/companies using llms to generate SOTA cuda/triton kernels 12) luminal inference 13) mojo/modular/max

Elliot Arledge

57,855 views • 7 months ago

#NVIDIAGTC kicks off this year with some kudos to CUDA, the platform and model that has become instrumental to the company’s ubiquity amid the AI boom, from CEO Jensen Huang.
1:10

#NVIDIAGTC kicks off this year with some kudos to CUDA, the platform and model that has become instrumental to the company’s ubiquity amid the AI boom, from CEO Jensen Huang.

TechCrunch

11,133 views • 2 months ago

Like, Love or Leave? The Rapid Transit 'Cuda! One of four promotional cars used by Chrysler. It only has 967 miles and was produced with a 440 SIX BARREL. The 'Cuda is serial number 100005.
0:58

Like, Love or Leave? The Rapid Transit 'Cuda! One of four promotional cars used by Chrysler. It only has 967 miles and was produced with a 440 SIX BARREL. The 'Cuda is serial number 100005.

Ultimate Muscle Car

39,167 views • 5 months ago

Jensen Huang said the AI factory industry "will be measured in trillions of dollars." "So far his track record on that is pretty good,” - Bernstein's Stacy Rasgon Agreed, remember Jensen's past predictions about AI, CUDA, etc. ?
7:39

Jensen Huang said the AI factory industry "will be measured in trillions of dollars." "So far his track record on that is pretty good,” - Bernstein's Stacy Rasgon Agreed, remember Jensen's past predictions about AI, CUDA, etc. ?

The AI Investor

14,185 views • 1 year ago

I used to find writing CUDA code rather terrifying. But then I discovered a couple of tricks that actually make it quite accessible. In this video I introduce CUDA in a way that will be accessible to Python programmers, and I even show how to do it all in Colaboratory!
1:17:55

I used to find writing CUDA code rather terrifying. But then I discovered a couple of tricks that actually make it quite accessible. In this video I introduce CUDA in a way that will be accessible to Python programmers, and I even show how to do it all in Colaboratory!

Jeremy Howard

210,641 views • 2 years ago

WHALE REPORT! 🐳🐳🐳🐳🐳🐳🐳🐳🐳🐳🐳 15 hours of research. JENSEN OF NVIDA WAS RIGHT! DeepSeek-V4 introduces several architectural and practical innovations that make it stand out, especially for long-context agentic workloads. The most geopolitically and technically notable aspect is its strong push toward hardware independence from NVIDIA’s CUDA ecosystem. The Big Story: Reduced/Limited Reliance on CUDA One of the most discussed “new things” is DeepSeek’s deliberate move toward CUDA independence (or at least strong dual-stack capability), driven by export controls and a push for sovereign AI stacks in China. •Training & Inference on Huawei Ascend: Huawei confirmed full support for V4 inference on its Ascend supernodes (e.g., Ascend 950/910 series) via CANN (Compute Architecture for Neural Networks), Huawei’s CUDA equivalent. DeepSeek co-optimized kernels directly with Huawei. Reports indicate fine-grained Expert Parallelism validated on Ascend NPUs, with speedups of 1.5–1.73x on non-NVIDIA platforms. •CANN Migration: The model (or key parts of its stack) was adapted/rewritten for CANN. This involved significant engineering effort, contributing to release delays. It represents a “full shift” or major de-NVIDIA-ization in the inference path for domestic deployments, proving frontier-scale MoE can run efficiently on alternative hardware. •Still Some CUDA Compatibility: Open-source components like the MegaMoE mega-kernel in DeepGEMM remain CUDA-based for NVIDIA users (with strong performance on Hopper/Blackwell). vLLM recipes support NVIDIA H200/B200 etc., and the weights run on CUDA setups. Earlier DeepSeek work (e.g., on V3) used low-level PTX to bypass some CUDA limitations for communication-heavy MoE. V4 builds on this hybrid pragmatism but emphasizes CANN-native paths. •Broader Implications: This validates a dual (or alternative) ecosystem. It reduces reliance on restricted NVIDIA tech for training/inference in China, potentially lowering costs and increasing resilience. Analysts see it as a blueprint for sovereign AI and a challenge to the “CUDA moat.”4635 Other notables include three-tier reasoning modes (Non-think / Think High / Think Max), Muon optimizer usage, heavy data curation (32T+ tokens, removing synthetic content), and MIT licensing for broad openness. In short, V4 isn’t just another bigger MoE but its efficiency tricks for usable million-token context, agent optimizations, and hardware-stack diversification (especially the CANN emphasis) mark it as a pragmatic leap for open, cost-effective, long-horizon AI. The CUDA/CANN angle is particularly significant in the current geopolitical context, showing how architectural and software innovations can mitigate hardware access barriers.50 Weights are on Hugging Face; API is live with competitive pricing. Quantized versions and community integrations are being released now. Last week Jensen Huang of Nvidia talked about why THE US MUST LEAD OPEN SOURCE. Welp we found out. Listen closely to this interview and the lack of knowledge the interviewer has. It is emblematic of the points of view permeating large AI companies. Perhaps you can hear Jensen now?
8:56

WHALE REPORT! 🐳🐳🐳🐳🐳🐳🐳🐳🐳🐳🐳 15 hours of research. JENSEN OF NVIDA WAS RIGHT! DeepSeek-V4 introduces several architectural and practical innovations that make it stand out, especially for long-context agentic workloads. The most geopolitically and technically notable aspect is its strong push toward hardware independence from NVIDIA’s CUDA ecosystem. The Big Story: Reduced/Limited Reliance on CUDA One of the most discussed “new things” is DeepSeek’s deliberate move toward CUDA independence (or at least strong dual-stack capability), driven by export controls and a push for sovereign AI stacks in China. •Training & Inference on Huawei Ascend: Huawei confirmed full support for V4 inference on its Ascend supernodes (e.g., Ascend 950/910 series) via CANN (Compute Architecture for Neural Networks), Huawei’s CUDA equivalent. DeepSeek co-optimized kernels directly with Huawei. Reports indicate fine-grained Expert Parallelism validated on Ascend NPUs, with speedups of 1.5–1.73x on non-NVIDIA platforms. •CANN Migration: The model (or key parts of its stack) was adapted/rewritten for CANN. This involved significant engineering effort, contributing to release delays. It represents a “full shift” or major de-NVIDIA-ization in the inference path for domestic deployments, proving frontier-scale MoE can run efficiently on alternative hardware. •Still Some CUDA Compatibility: Open-source components like the MegaMoE mega-kernel in DeepGEMM remain CUDA-based for NVIDIA users (with strong performance on Hopper/Blackwell). vLLM recipes support NVIDIA H200/B200 etc., and the weights run on CUDA setups. Earlier DeepSeek work (e.g., on V3) used low-level PTX to bypass some CUDA limitations for communication-heavy MoE. V4 builds on this hybrid pragmatism but emphasizes CANN-native paths. •Broader Implications: This validates a dual (or alternative) ecosystem. It reduces reliance on restricted NVIDIA tech for training/inference in China, potentially lowering costs and increasing resilience. Analysts see it as a blueprint for sovereign AI and a challenge to the “CUDA moat.”4635 Other notables include three-tier reasoning modes (Non-think / Think High / Think Max), Muon optimizer usage, heavy data curation (32T+ tokens, removing synthetic content), and MIT licensing for broad openness. In short, V4 isn’t just another bigger MoE but its efficiency tricks for usable million-token context, agent optimizations, and hardware-stack diversification (especially the CANN emphasis) mark it as a pragmatic leap for open, cost-effective, long-horizon AI. The CUDA/CANN angle is particularly significant in the current geopolitical context, showing how architectural and software innovations can mitigate hardware access barriers.50 Weights are on Hugging Face; API is live with competitive pricing. Quantized versions and community integrations are being released now. Last week Jensen Huang of Nvidia talked about why THE US MUST LEAD OPEN SOURCE. Welp we found out. Listen closely to this interview and the lack of knowledge the interviewer has. It is emblematic of the points of view permeating large AI companies. Perhaps you can hear Jensen now?

Brian Roemmele

70,873 views • 1 month ago

The largest advancement of the CUDA platform since its creation in 2006 is here 👀 Introducing CUDA Tile, a tile-based programming model that provides the ability to write algorithms at a higher level and abstract away the details of specialized hardware, such as tensor cores. Read the technical blog 👉
2:34

The largest advancement of the CUDA platform since its creation in 2006 is here 👀 Introducing CUDA Tile, a tile-based programming model that provides the ability to write algorithms at a higher level and abstract away the details of specialized hardware, such as tensor cores. Read the technical blog 👉

NVIDIA AI Developer

244,885 views • 6 months ago

Open is the AI foundation. “Beyond CUDA to me is the democratization of compute... ROCm gave us a seamless training experience.” PhD student Neha Prakriya is building smarter AI with AMD GPUs + ROCm — no walled gardens, just open innovation. See more:
2:19

Open is the AI foundation. “Beyond CUDA to me is the democratization of compute... ROCm gave us a seamless training experience.” PhD student Neha Prakriya is building smarter AI with AMD GPUs + ROCm — no walled gardens, just open innovation. See more:

AI at AMD

46,547 views • 1 year ago

.Siemens and NVIDIA are accelerating the next wave of automation with physical AI. 🤝 Watch the full hashtag#CES2026 demo to see how Siemens is integrating NVIDIA CUDA-X libraries and Omniverse into its EDA, CAE, and digital twin portfolio, bringing physical AI across the entire industrial lifecycle. 📺
0:38

.Siemens and NVIDIA are accelerating the next wave of automation with physical AI. 🤝 Watch the full hashtag#CES2026 demo to see how Siemens is integrating NVIDIA CUDA-X libraries and Omniverse into its EDA, CAE, and digital twin portfolio, bringing physical AI across the entire industrial lifecycle. 📺

NVIDIA Omniverse

19,602 views • 4 months ago

wow.. This new AI motion capture plugin in Unreal Engine can use two cameras to capture both character rigging and facial expressions in real time 🤯 it’s free, powered by NVIDIA CUDA
0:39

wow.. This new AI motion capture plugin in Unreal Engine can use two cameras to capture both character rigging and facial expressions in real time 🤯 it’s free, powered by NVIDIA CUDA

el.cine

90,973 views • 1 year ago

Nvidia announces the new RTX Spark, a new platform powered by the NX1 CPU, and shows off Spark laptops running 007 First Light and Forza 6. The CPU has 20 ARM based cores and a Blackwell RTX GPU with 6144 CUDA Cores. This is the same core count as a 5070, but with 128GB of unified LPDDR5X RAM memory sitting in the same package as the CPU and GPU. The entire Nvidia software stack is available, particularly CUDA, vital for AI. Nvidia's new laptops will likely be ideal for running local LLMs be cause the unified memory means you can load models up to 120-180B parameters (quantized). These laptops are expected to ship later this year and could become strong competitors to high-end MacBooks and even Mac Studios for local AI workloads, thanks to CUDA support and unified memory. Price is unannounced.
2:24

Nvidia announces the new RTX Spark, a new platform powered by the NX1 CPU, and shows off Spark laptops running 007 First Light and Forza 6. The CPU has 20 ARM based cores and a Blackwell RTX GPU with 6144 CUDA Cores. This is the same core count as a 5070, but with 128GB of unified LPDDR5X RAM memory sitting in the same package as the CPU and GPU. The entire Nvidia software stack is available, particularly CUDA, vital for AI. Nvidia's new laptops will likely be ideal for running local LLMs be cause the unified memory means you can load models up to 120-180B parameters (quantized). These laptops are expected to ship later this year and could become strong competitors to high-end MacBooks and even Mac Studios for local AI workloads, thanks to CUDA support and unified memory. Price is unannounced.

Grummz

28,671 views • 2 days ago

The CUDA moat is real, but probably not for long, says CEO of AI infrastructure platform Modal Erik Bernhardsson. He says he's bullish on alternative accelerators over the 2-3 year timeline, even though there's currently zero demand from his customers for TPUs, etc. "The cost today of rewriting your software to run on those stacks is very high... But the cost is going to go down." "You're going to have software that basically lets you take CUDA-compatible stuff and run it on alternative accelerators."
1:33

The CUDA moat is real, but probably not for long, says CEO of AI infrastructure platform Modal Erik Bernhardsson. He says he's bullish on alternative accelerators over the 2-3 year timeline, even though there's currently zero demand from his customers for TPUs, etc. "The cost today of rewriting your software to run on those stacks is very high... But the cost is going to go down." "You're going to have software that basically lets you take CUDA-compatible stuff and run it on alternative accelerators."

TBPN

34,932 views • 12 days ago

Chamath: US AI Startups Can Learn A Lot from DeepSeek "This is a case where necessity was the mother of invention." On E213, Chamath Palihapitiya explained how, even if the $6M number is inaccurate, DeepSeek still had some impressive breakthroughs: GRPO > PPO for reinforcement learning "These guys were like, 'Well, how am I going to do this reinforcement learning thing?' They invented a totally different algorithm." "It uses a lot less computer memory and it's highly performant." PTX > CUDA to build the model "And then the second thing that was crazy is everybody is used to building models and compiling through CUDA, which is NVIDIA's proprietary language." "And these guys worked totally around CUDA and they did something called PTX, which goes right to the bare metal." What US AI startups can learn from this: "We, meaning the West, with all the money that we've had, didn't come up with these ideas." "And I think part of why we didn't come up is not that we're not smart enough to do it, but we weren't forced to because the constraints didn't exist." "And so I just wonder how we make sure we learn this principle, meaning, when the AI company wakes up and rolls out of bed and some VC gives them $200M, maybe that's not the right answer for a Series A or a Seed." "And maybe the right answer is $2M so that they do these DeepSeek-like innovations."
0:55

Chamath: US AI Startups Can Learn A Lot from DeepSeek "This is a case where necessity was the mother of invention." On E213, Chamath Palihapitiya explained how, even if the $6M number is inaccurate, DeepSeek still had some impressive breakthroughs: GRPO > PPO for reinforcement learning "These guys were like, 'Well, how am I going to do this reinforcement learning thing?' They invented a totally different algorithm." "It uses a lot less computer memory and it's highly performant." PTX > CUDA to build the model "And then the second thing that was crazy is everybody is used to building models and compiling through CUDA, which is NVIDIA's proprietary language." "And these guys worked totally around CUDA and they did something called PTX, which goes right to the bare metal." What US AI startups can learn from this: "We, meaning the West, with all the money that we've had, didn't come up with these ideas." "And I think part of why we didn't come up is not that we're not smart enough to do it, but we weren't forced to because the constraints didn't exist." "And so I just wonder how we make sure we learn this principle, meaning, when the AI company wakes up and rolls out of bed and some VC gives them $200M, maybe that's not the right answer for a Series A or a Seed." "And maybe the right answer is $2M so that they do these DeepSeek-like innovations."

The All-In Podcast

108,789 views • 1 year ago

Wracają kontrole Ruch Obrony Granic to wraca Straż Graniczna ! Cuda cuda! Świetna robota Ruch Obrony Granic - dziekujemy!
1:20

Wracają kontrole Ruch Obrony Granic to wraca Straż Graniczna ! Cuda cuda! Świetna robota Ruch Obrony Granic - dziekujemy!

Aleksandra Fedorska

24,180 views • 1 month ago