An optimistic take on neural networks for programming: https://medium.com/@karpathy/software-2-0-a64152b37c35
It makes some good points about predictable runtime performance, ability to trade CPU for accuracy, and the ease of hardware acceleration.
miniblog.
Related Posts
1
I've now worked on several systems where the JSON was so big that avoiding pretty-printing was a performance win!
3
Difftastic 0.71 is out! This is a polish release, still well worth upgrading:
* A ton of performance improvements, especially for nasty corner cases.
* Improved C++, Dockerfile, Haskell, JavaScript, Makefile, Perl, Ruby, Rust, Scala and TypeScript.
3
I've been experimenting with LLM autoresearch. I prompted Sol to make difftastic faster without changing output on the test suite, and log everything it tried:
https://github.com/Wilfred/difftastic/blob/c6e9c6fed4c71276a8b0e377c92695c3d929dc7f/PERF_RESEARCH_LOG.md
It found some interesting performance tweaks, although it's too tolerant of complexity.