https://hackerpaper.kaipfister.de/p/2605.04325/Collected HN comments and submissions linking this paper.enMon, 02 Nov 2026 11:16:22 GMThttps://news.ycombinator.com/item?id=49878456urn:hackerpapers:mention:2605.04325:49878456Mon, 28 Sep 2026 14:21:58 GMT<p>Several weeks ago, we published a more substantive article using this very same analogy: https://www.linkedin.com/pulse/announcing-universal-mathemat... The article describes a unifying mathematical framework for representing neural network architectures. It can be used to describe existing ones, like transformers and convolution, but can also be used to generate entirely novel ones. As part of our research, we generated several thousand new architectures, some of which show remarkable properties. The corresponding research paper is up on arXiv: https://arxiv.org/abs/2605.04325</p><p><a href="https://hackerpaper.kaipfister.de/p/2605.04325/">Paper and discussion history</a>. A mention is not an endorsement.</p>cs.LGhttps://news.ycombinator.com/item?id=48562956urn:hackerpapers:submission:2605.04325:48562956Tue, 16 Jun 2026 22:15:39 GMT<p>On The Architectural Complexity of Neural Networks</p><p></p><p>5 points; 0 discussion replies on HN. <a href="https://hackerpaper.kaipfister.de/p/2605.04325/">Paper and source history</a>. A submission is not an endorsement.</p>cs.LG