I’m impressed by tinygrad. Full-featured deep-learning framework with ZERO dependencies, fully hackable, and orders of magnitude fewer LoC than PyTorch.
replyWant to run Qwen3.8 27B? Download the quantized model from HF, git clone tinygrad, then run `DEV=METAL python3 -m tinygrad.llm --max_context 8192 -m <GGUF File>`. Read the code and experiment - takes 1 minute.
Compare that with PyTorch (great, but almost impossible to comprehend and modify, IMO) or Mojo/MAX (complex and not even fully open source, not to mention that the Modular folks aren’t particularly nice). Tinygrad is a breath of fresh air.
"Zero dependencies" except Python, which is a pretty huge dependency.
replyIf you're like me and also wondered "huh, what's tinygrad up to?" apparently they now let you place $100,000 USD pre-orders on a $10M shipping container full of GPUs which they say will ship (literally it seems) in 2027.
replyThey're claiming a lot, but don't have much to show for it
replyDid they have an actual metal container sitting outside their office building in San Diego? I would love to see it on the inside kinda like a walk-in tour you get from real estate listings
replyhe bought an empty container and he's looking for someone to build it out.
replyi ran 7MW and 20,000 gpus in containers in WA for mining eth for many years. it is a bad idea that he's trying to do. but george is george and the hype is real...
Do you have a write-up of this? I enjoy seeing these Playstation/ASIC/GPU mining installations which are usually some combination of incredibly clever engineering and duct tape.
replyi ran 20k of those bc-250's too... but those were in a real dc. nice to see them getting use again.
replysorry, no writeup other than what i've posted already.