CUDA

Dave Blundin

mentioned

"CUDA is still the dominant mode for AI researchers who are working on new training. All the AI researchers in the world moved to Python and PyTorch, and the only way to just run it at scale is the CUDA layer, which translates your PyTorch right in..."

"Nvidia invented — I think it was in the 2000s — they put out something called CUDA, which is the underlying software library and the way to run software on GPUs. It took them a solid decade or more to make it something where they could do data ana..."

"I don't look for CUDA experience at all. That's actually a huge red herring. CUDA as a concept or GPU as a concept have evolved so much in the last 5 years. There's no point asking for 10 years of experience."

"CUDA's moat is dramatically diminished because the models don't care what they run on — and that's what actually matters, what's built on top of the models. But it still matters. It's still something of a moat."

Tom Bilyeu

mentioned

"CUDA, which is Nvidia's software, has proven that it can improve performance over time, even on already installed chips, which serves to extend that economic life."

"When they started to build a real software stack with CUDA and SIMT as a technology — multi-threading and so on — it just kept getting a little bit better every release and became more robust. All of a sudden the HPC community said, 'Hey, we could..."

Jordi Hays

mentioned

"Is there a world where you get off of classical — much like we're seeing leading AI firms get off of CUDA — but by and large it's the same training paradigm?"

"CUDA was always built to go beyond CUDA and to scale out infrastructure. CUDA is not the moat at Nvidia, their ecosystem is the moat."

"Jensen made CUDA free. He made it so that any GPU, even the crappiest gaming GPU could also run CUDA. And so the accessibility barrier was really low so anyone could get into it."

Invest Like The Best

mentioned

"Something like CUDA that has been a part of NVIDIA's story for a long time now that allows you to do a lot with the underlying actual hardware"

TBPN

mentioned

"CUDA is a programming model that enables GPUs to accelerate demanding workloads by parallelizing computation."

Sebastian Raschka

mentioned

"The moat of NVIDIA is probably not just the GPU. It's more like the CUDA ecosystem, and that has evolved over two decades"