vova_hn2 4 days ago

This article seems to use the word "energy" to call both the energy function and its value, which makes it very confusing and hard to read.

At first I thought that it might just be a convention of this field that I am not familiar with, but another related recent article [0][1] (discussed on HN [2]) doesn't have this problem.

Also, I should note that these two articles are weirdly similar (similar approaches, both use sudoku as an example) while being published by two seemingly unrelated groups of authors in quick succession. I wonder why.

[0] https://energy-based-reasoners.github.io/

[1] https://arxiv.org/pdf/2611.08977

[2] https://news.ycombinator.com/item?id=50543989

  • bikvfeee 2 days ago

    Perhaps it's either something that clause saw before or close enough to something Claude saw before.