more on forge and monarch tomorrow. watch out for giant new code pushes
@soumithchintala
-
Apple’s PyTorch Support Lags Behind NVIDIA Says Engineer
By
–
MacStudio you ask?
Apple Engineering's **actual** time spent on PyTorch support has't given me confidence that PyTorch Mac experience would get anywhere close to NVIDIA's any time soon, if ever.
The Meta engineers continue to do a huge amount of heavy-lifting for improving the -
NVIDIA DGX Spark: Software Excellence in AI Hardware
By
–
Sometimes we forget that NVIDIA wins because it's a software company. DGX Spark is a reminder of that. It's a CUDA dev machine that's beautiful enough and small enough to be on my desk and with enough memory to fit a truckload of params.
It's not the fastest or best at anything, -
PyTorch Multiprocessing Tensor Sharing Across Processes
By
–
pytorch uses this for sharing tensors across processes in torch.multiprocess
-
Engineer motivation versus external incentives in tech projects
By
–
your comments are sooo in bad faith, i feel really sorry for what you've become. the engineer who proposed and executed the project doesn't give a shit about meta psc or extrinsic incentives. if he did, he'd just have rode the pytorch accolades to more Meta-internal "success"
-
Impressive Home Mapping Capabilities Impress Roboticist Observer
By
–
.
@maticrobots mapping my home is a thing of beauty. as a roboticist, I'm so impressed! -
Nvidia’s Strategic AI Investment: Market Thesis and Capital Deployment
By
–
there's also an argument to be made that these are good deals. what would nvidia do with all that free paper money (which would've been a discount), might as well deploy it into the most successful AI company in the market if the thesis is there's still massive upside, while
-
Questionable Deal Structures in Overvalued Robotics Startups
By
–
interestingly these scratch-your-back deals have apparently become commonplace. an overvalued robotics company (that everyone's confused as to how they're raising so much) apparently raised a gazillion dollars through similar deal structure
-
Nvidia GPU costs and strategic investment in OpenAI
By
–
10GW is about $340B of nvidia h100 at $30k/gpu (assuming 20% of power for non-gpus). if openai got a 30% volume discount, they'd pay nvidia $230b probably. so instead, maybe openai pays nvidia full price and nvidia invests the excess $100B into openai stock (just throwing
-
Search Space Optimization in Deep Learning: Beyond Polyhedral Constraints
By
–
polyhederal search spaces are okay but 1. for DL they dont even need to be poly, just rectangular is usually sufficient 2. the search space is highly discontinuous and non-linear, so the search is timetaking and highly unpredictable as a function of shape size