<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>OpenChamber Blog – AI Coding, Agents, and the Work Around Them</title><description>Field notes from the OpenChamber team on AI coding agents, multi-agent workflows, and the tools around them.</description><link>https://openchamber.dev/</link><language>en</language><item><title>Agent memory: what changed in three years of research</title><link>https://openchamber.dev/blog/agent-memory-research/</link><guid isPermaLink="true">https://openchamber.dev/blog/agent-memory-research/</guid><description>What 3,283 papers from 2023 to 2026 say about memory in LLM agents: what changed, how memory gets attacked, and which approaches hold up when measured.</description><pubDate>Sun, 25 Oct 2026 00:00:00 GMT</pubDate><category>research</category><category>agents</category></item><item><title>OpenChamber 2.0: it&apos;s getting hot reload in here</title><link>https://openchamber.dev/blog/opencode-v2/</link><guid isPermaLink="true">https://openchamber.dev/blog/opencode-v2/</guid><description>OpenChamber 2.0.0 moves to OpenCode 2. Skills, agents, MCP servers and config now apply the moment you save them, plugins can bring whole features, and we used that to ship web search, Code Mode, plugin status and a stats page. Here&apos;s what changed, what&apos;s next, and a thank-you to the OpenCode team.</description><pubDate>Wed, 23 Sep 2026 00:00:00 GMT</pubDate><category>opencode</category><category>plugins</category><category>release</category></item><item><title>A year of AI coding complaints: what stopped hurting and what took its place</title><link>https://openchamber.dev/blog/ai-coding-complaints/</link><guid isPermaLink="true">https://openchamber.dev/blog/ai-coding-complaints/</guid><description>66,320 Reddit complaint posts about AI coding agents, labelled into 62 mechanisms and counted month by month. Wrong code fell from first place to seventh; the cost of a run went from ninth to third.</description><pubDate>Mon, 21 Sep 2026 00:00:00 GMT</pubDate><category>research</category><category>agents</category></item><item><title>Auto mode: an experiment in letting a 250 ms model pick the model for every message</title><link>https://openchamber.dev/blog/jev-model-routing/</link><guid isPermaLink="true">https://openchamber.dev/blog/jev-model-routing/</guid><description>We put Jev in front of every message in OpenChamber so a cheap classifier picks the model. It worked great until I typed &quot;do it&quot;. Here&apos;s what we experimented with, where it pays off, where it doesn&apos;t yet, and what we learned about asking a model the right question.</description><pubDate>Sat, 19 Sep 2026 00:00:00 GMT</pubDate><category>agents</category><category>models</category></item><item><title>Jev by TypeSafe AI: 193× in the headlines, 7× in user reports</title><link>https://openchamber.dev/blog/jev-typesafe-ai/</link><guid isPermaLink="true">https://openchamber.dev/blog/jev-typesafe-ai/</guid><description>We analysed 12,759 tweets about Jev&apos;s launch to separate TypeSafe AI&apos;s speed and cost claims from users&apos; reported results and see where a decision model fits.</description><pubDate>Thu, 17 Sep 2026 00:00:00 GMT</pubDate><category>models</category><category>analysis</category></item><item><title>AI coding agents in 2026: a practical guide to 36 tools</title><link>https://openchamber.dev/blog/ai-coding-agents/</link><guid isPermaLink="true">https://openchamber.dev/blog/ai-coding-agents/</guid><description>Compare 36 coding agents and workspaces by how they work, what they cost, and how much control you keep over your code and models.</description><pubDate>Mon, 14 Sep 2026 00:00:00 GMT</pubDate><category>agents</category><category>comparison</category></item></channel></rss>