https://hackerpaper.kaipfister.de/p/2602.22661/Collected HN comments and submissions linking this paper.enMon, 02 Nov 2026 11:27:26 GMThttps://news.ycombinator.com/item?id=47336498urn:hackerpapers:submission:2602.22661:47336498Wed, 11 Mar 2026 15:00:02 GMT<p>Diffusion LLM may make most of the AI engineering stack obsolete</p><p>I&#x27;ve been deep-diving into diffusion language models this week and I think this is the most underrated direction in AI right now. The core issue with autoregressive LLMs: Every major model today (GPT, Claude, Gemini) generates one token at a time, left to right. Each token depends on the previous one. This single architectural constraint has shaped the entire AI industry: - Models can&#x27;t revise what they already wrote → we build chain-of-thought, reflection, and multi-pass reasoning to force them to &quot;think before committing&quot; - One forward pass per token → we invest heavily in speculative…</p><p>3 points; 0 discussion replies on HN. <a href="https://hackerpaper.kaipfister.de/p/2602.22661/">Paper and source history</a>. A submission is not an endorsement.</p>cs.CL