I see good answers already, but here's a concrete example: In my University we had to decide between both libraries so, as a test, we de...
I see good answers already, but here's a concrete example: In my University we had to decide between both libraries so, as a test, we decided to write a language model from scratch. The first minor problem with TF was that (if memory serves me right) you were supposed to declare your network "backwards" - instead of saying "A -> B -> C" you had to declare "C(B(A))". The major problem, however, was that there was no way to add debug messages - either your network worked or it didn't. To make matters worse, the "official" TF tutorial on how to write a Seq2Seq model didn't compile because the library had changed but the bug reports for that were met for years with "we are changing the API so we'll fix the example once we're done". PyTorch, by comparison, had the advantage of a Python-based interface - you simply defined classes like you always did (including debug statements!), connected them as variables, and that was that. So when I and my beginner colleagues had to decide which library to pick, "the one that's not a nightmare to debug" sounded much better than "the one that's more efficient if you have several billions training datapoints and a cluster". Me and my colleagues then went on to become professionals, and we all brought PyTorch with us.
Posted by probably_wrong on Nov 7, 2025, 12:53 PM
Tags
Related issues
- (1) Because it doesn't actually solve the problem.
- Every time fossil comes up, I hear this same issue - "why is there no rebase?" And honestly, I've never been able to get this argument.
- My biggest problem with running local LLMs on my M4 Max/128GB RAM is the prefill latency.
- Biggest problem with httpx I see is authors sanity https://github.com/encode/httpx/discussions/3784 .
- There aren't many situations today that make one think: Why doesn't this business have a website?
- > IAM can stay MS That's leaving the most critical component still with a US company.
Raise real-world problems, find people facing the same issue, and discuss solutions.2,000 issues0 comments
hackernews (Network)
(1) Because it doesn't actually solve the problem.
NewEvery time fossil comes up, I hear this same issue - "why is there no rebase?" And honestly, I've never been able to get this argument.
My biggest problem with running local LLMs on my M4 Max/128GB RAM is the prefill latency.
Biggest problem with httpx I see is authors sanity https://github.com/encode/httpx/discussions/3784 .
There aren't many situations today that make one think: Why doesn't this business have a website?
> IAM can stay MS That's leaving the most critical component still with a US company.
I am very petty about this one bug and have a very old axe to grind that this reminded me of!
Not just the edge router.
I see good answers already, but here's a concrete example: In my University we had to decide between both libraries so, as a test, we de...
I see good answers already, but here's a concrete example: In my University we had to decide between both libraries so, as a test, we decided to write a language model from scratch. The first minor problem with TF was that (if memory serves me right) you were supposed to declare your network "backwards" - instead of saying "A -> B -> C" you had to declare "C(B(A))". The major problem, however, was that there was no way to add debug messages - either your network worked or it didn't. To make matters worse, the "official" TF tutorial on how to write a Seq2Seq model didn't compile because the library had changed but the bug reports for that were met for years with "we are changing the API so we'll fix the example once we're done". PyTorch, by comparison, had the advantage of a Python-based interface - you simply defined classes like you always did (including debug statements!), connected them as variables, and that was that. So when I and my beginner colleagues had to decide which library to pick, "the one that's not a nightmare to debug" sounded much better than "the one that's more efficient if you have several billions training datapoints and a cluster". Me and my colleagues then went on to become professionals, and we all brought PyTorch with us.
Comments (0)
No comments yet. Start the discussion.