Teddeh (Matt Culliton)·1h agoThe Date Always PassesTwo arguments show up in every thread about agents right now.
leikooo·1h agoHow Codex, Claude Code, Hermes Agent and pi compress long agent sessionsIf you run agents on long tasks, sooner or later you’ll see this: a line pops up saying Compacting conversation…. When it finishes, the…
Chuah Kee Yong·1d agoBehind Claude Code: What I Learned Building an Agent Harness [Part 1]This is actually my study note while I am studying how to build my own harness from this repo — not teaching anyone, but just sharing the…
Naor Weissmann·1d agoSmoke-Testing a Small Language Model with RadiChatI’ve been experimenting with small language models for a while, and one recurring need is a quick way to check their behavior after…A response icon1A response icon1
InJIN System ArchitectbyJIN·2d agoWho Holds the Knowledge at Handoff?How Tencent Video’s Working, coding, and on-call practices try to make knowledge cross three workflow boundaries
CyberRaya·2d agoHarness Engineering: Making Your Coding Agent Reliable Without Swapping the ModelThe model is the engine. The harness decides whether it ever reaches the destination.
Patrick Elfert·2d agoYou may not need a Software FactoryOver the last few weeks I spent a lot of time figuring out how our team can work well with coding agents. Some reasons for this came from…
Spandanmanoj·2d agoHarness Engineering: Building the F1 Racing Team Behind AI AgentsWhy building powerful AI agents is not just about having the best model, but about engineering the entire system around it.
Sai Gavaskar Sakalam·2d agoThe Unfair Split: Guides and Sensors for AI AgentsPart 1’s test passed. The split was still unfair. Here is the simple map that fixes problems like this.
InPractical AI SystemsbyMehulpratapsingh·3d agoClaude Code ‘God Mode’ Without the Dangerous ShortcutAn unofficial Omega-style setup for giving Claude Code better project context, useful tools, repeatable workflows, and clear safety…