News for hackers. Stuff that matters.
Re:Tinygrad will be the next Linux and LLVM by TimSchumann on Monday September 23, 2024 @03:02AM (#41623138)
> Without CPUs, we can be freed from the tyranny of the halting problem.

Can someone please explain to me what this even means in this context?

Serious question.

Re:Tinygrad will be the next Linux and LLVM by WithinReason on Monday September 23, 2024 @03:54AM (#41623485)
A neural network is perfectly deterministic, the runtime is predictable before you run it. Which I don't think is going to be true much longer.

https://news.ycombinator.com/item?id=41623474

Re:Tinygrad will be the next Linux and LLVM by orra on Monday September 23, 2024 @03:40AM (#41623394)
He also claims that the cardinality of the reals is the same as the integers.

https://news.ycombinator.com/item?id=36074287

You could say he had a history of using big words to talk shit.

Re:Tinygrad will be the next Linux and LLVM by mikewarot on Monday September 23, 2024 @04:10AM (#41623587)
Think of it as unwinding a program all the way until it's just a list of instructions. You can know exactly how long that program will take, and it will always take that same time.
Re:Tinygrad will be the next Linux and LLVM by krisoft on Monday September 23, 2024 @04:25AM (#41623702)
But will it always solve the task? Because without that it it is trivially easy to “solve” the halting problem by just declaring that the turing machine halts after X steps.
Re:Tinygrad will be the next Linux and LLVM by TimSchumann on Monday September 23, 2024 @10:57AM (#41626850)
Wouldn’t this also imply a lack of Turing completeness, and thus not be good for general purpose computing?
Re:Tinygrad will be the next Linux and LLVM by benzible on Monday September 23, 2024 @03:05AM (#41623158)
It's gibberish. For one thing... https://arxiv.org/abs/1901.03429