Waleed Atallah from Mako joins us to discuss automating GPU kernel generation with AI.
03:07 – Why GPU kernels bottleneck progress
04:46 – Three current ways to get kernels
06:02 – LLMs: can they write fast kernels?
07:48 – Mako’s “kernel agent” in action
09:12 – Early adopters & best-fit users
12:05 – Data scarcity limits fine-tuning
16:18 – AI-written kernels by 2030?
20:27 – Top technical risks ahead
#BeyondCUDA #GPUkernels #mako
🌐 Connect with us:
https://www.linkedin.com/in/jtatarchuk
https://www.linkedin.com/in/waleedatallah
Keith Strier gives his take on the GPU divide, his move from NVIDIA to AMD and what it means for AI. He breaks down sovereign AI, CUDA’s role and why having more than one vendor matters.
02:10 – Quitting 17-yr consulting for NVIDIA
04:55 – NVIDIA culture shock: flat & fast
07:32 – EY’s first AI practice (2016)
10:08 – Estonia & Malta AI strategies
12:56 – Launching NVIDIA’s “AI Nations”
14:30 – ChatGPT hits; Jensen pivots
17:05 – GPU land-grab by governments
19:40 – Real bottleneck: data centers
23:55 – Move to AMD to scale AI infra
#BeyondCUDA #AIInfrastructure #AMD
🌐 Connect with us:
https://www.linkedin.com/in/jtatarchuk
https://www.linkedin.com/in/keithstrier/
Learn how Spectral’s Scale toolkit breaks CUDA’s vendor lock-in by letting you run the same code on AMD GPUs with zero rewrites. This episode shows how to future-proof your stack without starting from scratch.
#GPUCloud #BeyondCUDA #TensorWave
01:05 - How did you get into GPU development?
03:39 - What is Scale?
05:35 - What are the technical hurdles?
06:34 - What are the opportunities?
07:46 - What is Spectral Focused On?
09:14 - How to break down Scale?
🌐 Connect with us:
https://www.linkedin.com/in/jtatarchuk
https://dk.linkedin.com/in/michaelsondergaard
Listen to the story of ZLUDA directly from Andrzej Janik! From Intel beginnings to AMD, hear the untold story of running CUDA apps on non-Nvidia GPUs. Learn about the journey, challenges, and future of multi-vendor AI compute.
#ZLUDA #AMD #CUDA
02:47 - What drew you into GPU software?
06:33 - Why start the ZLUDA project?
08:16 - When did you realize CUDA was viable?
09:56 - What happened at Intel?
11:45 - How did you transition to AMD?
14:58 - What are the challenges?
17:13 - Why did AMD stop funding?
19:59 - Where is ZLUDA today?"
🌐 Connect with us:
https://www.linkedin.com/in/jtatarchuk
https://github.com/vosen
How we cuts cloud GPU costs 3-7x! Andrey Chepstov - CEO @ dstack (https://dstack.ai) chats with Jeff Tatarchuk on open source AI & redefining infrastructure.
#AI #OpenSource #GPUs
00:29 - Intro and journey at JetBrains.
09:02 - How was dstack born?
13:14 - What are the challenges with Kubernetes?
18:36 - What problems does dstack solve?
26:21 - dstack roadmap and features.
30:25 - Open source journey and community building.
42:03 - What does Beyond CUDA mean?
🌐 Connect with us:
https://www.linkedin.com/in/jtatarchuk/
https://www.linkedin.com/in/andrey-cheptsov
Quentin Anthony, model training lead at Zyphradiscussing building a multimodal agent system for enterprises. He covers optimizing models, hardware, and reducing training costs.
02:45 - Oak Ridge National Lab Work
05:53 - How to make models personalizable offline?
08:13 - How do hybrid models differ?
09:38 - How to reduce training costs?
11:56 - What is Quentin excited about?
13:09 - What are the biggest challenges?
16:14 - What's next for Zyphra?
#TensorWave #AMDInstinct #AICloud
🌐 Connect with us:
https://www.linkedin.com/in/jtatarchuk
https://www.linkedin.com/in/quentin-anthony
Ex-Meta tech lead, on curing cancer with molecule programming!
#AI #Biotech #GPUs
00:23 - What Sparked Kush's AI Journey?
00:33 - How AI Models are Built
00:43 - Can AI Cure Cancer?
01:16 - What Does Beyond Cuda Mean?
00:51 - What is Molecule Programming?
00:13 - US vs China: Who's Ahead in AI?
00:17 - What is AI Beyond LLMs?
00:22 - How Did Kush Start His Company?
🌐 Connect with us:
https://www.linkedin.com/in/jtatarchuk/
https://www.linkedin.com/in/kooshiar/
Jay Dawani from Lemurian Labs talks about GPU performance and how they're uniting the stack for accessible AI development.
#AIdevelopment #BeyondCUDA #LemurianLabs
00:49 - Why Lemurian?
02:56 - Challenges in AI and Robotics.
06:13 - Reinventing Software for AI.
08:33 - Future of AI Hardware.
0:12:18 - Optimizing AI Performance.
0:16:38 - Prioritize to Make Things Easier?
0:22:23 - AI's Role in Software Development.
🌐 Connect with us:
https://www.linkedin.com/in/jtatarchuk
https://ca.linkedin.com/in/jaydawani
History of CUDA with Greg Diamos, an early pioneer talk about CUDA's challenges, successes, and its impact on AI. Is the CUDA moat shrinking?
00:00 - Origins of CUDA
02:15 - Greg's Early Interest
03:30 - Joining Nvidia
04:45 - Build a Faster Computer
06:30 - CUDA Dominated AI Research
08:50 - What is Nvidia's Moat?
13:45 - AMD's Accelerators
17:35 - More People Using AMD
#CUDA #AMD #AI
🌐 Connect with us:
https://www.linkedin.com/in/jtatarchuk
https://www.linkedin.com/in/gregory-diamos-1a8b9083/
