<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>feynpath — AI, software, and computing research</title><description>An open notebook exploring AI, software, and computing research through experiments, technical notes, and evolving ideas.</description><link>https://feynpath.dev/</link><language>en-US</language><item><title>AI Models Are Challenging Our Defenses</title><link>https://feynpath.dev/research/the-model-is-not-the-moat/</link><guid isPermaLink="true">https://feynpath.dev/research/the-model-is-not-the-moat/</guid><description>Microsoft’s MDASH reached 88.45% on CyberGym. The next eleven weeks showed why cybersecurity’s durable AI advantage will live in open agents, tools, benchmarks, and proof, not in any single model.</description><pubDate>Wed, 29 Jul 2026 00:00:00 GMT</pubDate><category>ai</category><category>cybersecurity</category><category>open source</category><category>agents</category><category>benchmarks</category></item><item><title>AI Learns to Calibrate a Quantum Computer</title><link>https://feynpath.dev/research/ai-learns-to-calibrate-a-quantum-computer/</link><guid isPermaLink="true">https://feynpath.dev/research/ai-learns-to-calibrate-a-quantum-computer/</guid><description>Google Quantum AI taught a reinforcement learning agent to tune a quantum processor from its own error detection events, without stopping the computation. It scales because the problem is local.</description><pubDate>Mon, 27 Jul 2026 00:00:00 GMT</pubDate><category>quantum</category><category>error correction</category><category>reinforcement learning</category><category>scalability</category><category>research</category></item><item><title>When Everything Becomes Easy, Raise the Difficulty</title><link>https://feynpath.dev/research/reflections/when-everything-becomes-easy-raise-the-difficulty/</link><guid isPermaLink="true">https://feynpath.dev/research/reflections/when-everything-becomes-easy-raise-the-difficulty/</guid><description>AI is lowering the barriers to creation, and the research says it is also lowering how much we think while we create. The answer is not to refuse the tools. It is to design harder problems.</description><pubDate>Mon, 27 Jul 2026 00:00:00 GMT</pubDate><category>reflection</category><category>ai</category><category>programming</category><category>games</category><category>learning</category></item><item><title>Slow down, be human again</title><link>https://feynpath.dev/research/reflections/slow-down-be-human-again/</link><guid isPermaLink="true">https://feynpath.dev/research/reflections/slow-down-be-human-again/</guid><description>A reflection on the need to focus on doing things well instead of doing them fast, especially in a time that pushes us to produce without stopping.</description><pubDate>Mon, 20 Jul 2026 00:00:00 GMT</pubDate><category>reflection</category><category>ai</category><category>work</category><category>humanity</category></item><item><title>The Last Machine We Need to Build</title><link>https://feynpath.dev/research/reflections/the-key-to-humanitys-expansion/</link><guid isPermaLink="true">https://feynpath.dev/research/reflections/the-key-to-humanitys-expansion/</guid><description>What would it mean for humanity to build a physical system capable of constructing any possible machine including itself?</description><pubDate>Mon, 20 Jul 2026 00:00:00 GMT</pubDate><category>reflection</category><category>von neumann</category><category>self replication</category><category>constructor theory</category><category>artificial life</category></item><item><title>Journalism’s Missing Safety Stack</title><link>https://feynpath.dev/research/journalisms-missing-safety-stack/</link><guid isPermaLink="true">https://feynpath.dev/research/journalisms-missing-safety-stack/</guid><description>Journalists working under violence, surveillance, and network failure need a safety stack designed for the entire reporting mission.</description><pubDate>Tue, 14 Jul 2026 00:00:00 GMT</pubDate><category>journalism</category><category>safety</category><category>technology</category><category>public-interest-tech</category><category>resilience</category></item><item><title>Reflection on Simulating Time With Square Root Space</title><link>https://feynpath.dev/research/reflections/williams-square-root-space/</link><guid isPermaLink="true">https://feynpath.dev/research/reflections/williams-square-root-space/</guid><description>A reflection on Ryan Williams’s square root space simulation and what it suggests for post-von-Neumann computing.</description><pubDate>Fri, 01 May 2026 00:00:00 GMT</pubDate><category>reflection</category><category>complexity</category><category>computing</category><category>post-von-neumann</category></item><item><title>Bayesian Probabilistic Layer for AI agents</title><link>https://feynpath.dev/research/bayesian-probabilistic-layer/</link><guid isPermaLink="true">https://feynpath.dev/research/bayesian-probabilistic-layer/</guid><description>A probabilistic reasoning layer for AI agents, grounded in Bayesian inference.</description><pubDate>Sun, 19 Apr 2026 00:00:00 GMT</pubDate><category>ai</category><category>bayesian</category><category>agents</category><category>research</category></item></channel></rss>