131–140 of 167 posts
The risk for Tinygrad is that PyTorch will create a new backend for Inductor, plug in their AMD codegen stuff and walala, PyTorch still king. I mean, they could have easily just taken that route themselves instead of bothering with a new ML framework and AD engine. 99% of the work is just the AMD codegen part of the compiler. Either way, super cool project and I wish them the best.
> walala Is this like voila or something else?
Earlier quoted context omitted.
That's comma.ai's policy since they make hardware and solve physical problems. The tiny corp has been hybrid (remote-first) since day 1 because it primarily writes open source software, and there's a long track record of success with remote for this kind of task. We have a few whole-team meetups in Hong Kong each year for 2-4 weeks, and there's a San Diego or Hong Kong office that anyone can work from as they choose.…
Have you run into problems with contributors who can't enter the PRC?
>People get hired by contributing to the repo. It’s a very self directed job, with one meeting a week and a goal of making tinygrad better I find this organizational structure compelling, probably the closest to reaching 100% productivity in a week as you can get.
It’s hard to argue against hiring contributors, but a bounty system that pays pennies vs. market value for skilled developers shouldn’t be the only interview path, it’s borderline exploitative.
Dude is an absolute nut, I'll stick with PyTorch https://geohot.github.io/blog/jekyll/update/2025/04/22/a-way...
Aside from the edgy Unabomber quote and calling both established US political parties racists, everything there is reasonable analysis. What exactly got your goat?
Earlier quoted context omitted.
You’re getting some pushback but one look at pm_reduce_collapse and I’m glad I don’t have to work in this codebase.
I agree. It is definitely readable but clearly formatting decisions were made to keep line count down over the flow of code. It isn't even just the variable names, things like having a function with a single return statement on one line, while legal, I cannot imagine why someone would do it in a non one off script code.
I agree. Seems to be a personal preference, but still feels odd.
Earlier quoted context omitted.
This idea that you're supposed to accept worse pay because you believe in the idea doesn't apply to George. If his companies succeed, he'll be rich. Of course, there's nothing wrong with even working for free if that's what you like, just don't make a moral principle out of it.
I don’t mean to imply it’s to get employees to accept lower pay. I mean that there is some implicit screening against people who are solely optimizing for high pay. Lots of reasons why you would want to simultaneously pay well and not attract people who are optimizing for pay.
Dude is an absolute nut, I'll stick with PyTorch https://geohot.github.io/blog/jekyll/update/2025/04/22/a-way...
The blog seems like a regular political opinion. We can't just call people we disagree with "absolute nuts" all the time.
https://geohot.github.io/blog/jekyll/update/2023/04/11/wireh...
https://geohot.github.io/blog/jekyll/update/2025/10/19/solut...
Dude is an absolute nut, I'll stick with PyTorch https://geohot.github.io/blog/jekyll/update/2025/04/22/a-way...
Aside from the edgy Unabomber quote and calling both established US political parties racists, everything there is reasonable analysis. What exactly got your goat?
Lots of words and weird analogies to say basically nothing. What is the status of the project? What can it do? What has it achieved in 5 years? But no, let's highlight how we follow the "Elon process". As a side note, whenever someone incessantly focuses on lines of code as a metric (in either direction), I immediately start to take them less seriously.
Did we read the same article? The third sentence is "The codebase is 18,935 lines not including tests." You can go to the repo and see what it does. Look at issues, PRs, etc. He wasn't saying it to brag about its size, but the opposite. The mission is to commoditize the petaflop. Basically allow LLMs to be trained efficiently on commodity non-NVIDIA GPUs. Would you prefer some bullshit mission like Meta of "connectin…
I've looked at the "only 18,935 lines of code" python code and it made me want to poke my eyes out. Not sure what's the point of this extreme code-golfing.
yes, it's really crazy, if people think you're exaggerating look at this: https://github.com/tinygrad/tinygrad/blob/master/tinygrad/co... One of the cases why I think obsession with lines of code is one of the most counterproductive metrics, it always produces code like this.
It helps that a single person can read all related bits on one screen.