Agents Observe vs Auto-Arch Tournament

Agents Observe is an open-source observability dashboard for Claude Code's multi-agent mode — the feature that lets multiple AI agents work in parallel on different parts of a codebase. As Claude Code moves from single-session to multi-agent coordination, the need for visibility into what each agent is doing, how they're communicating, and where they're getting stuck becomes a real operational need. Agents Observe fills this gap with a real-time web dashboard that streams agent activity. The dashboard shows active agent sessions, their current task status, tool call histories, and inter-agent message flows. It hooks into Claude Code via the existing logging infrastructure and presents the data in a swimlane view reminiscent of distributed tracing tools like Jaeger or Zipkin. For teams running multiple Claude Code instances on large codebases, this provides the kind of observability that was previously only available by reading raw log files. With 73 points on the Hacker News Show HN thread and 25 comments — mostly from Claude Code heavy users — the demand signal is clear: as multi-agent coding workflows become mainstream, debugging and monitoring them requires dedicated tooling. The open-source approach ensures compatibility with self-hosted Claude Code setups, which is a common pattern for enterprise teams with data sovereignty requirements.

Auto-Arch Tournament is an autonomous research system where an AI agent iteratively proposes, implements, and validates microarchitectural improvements to a RISC-V CPU. Starting from a standard 5-stage pipeline, the loop runs hypotheses in parallel, each going through formal verification (53 symbolic checks), cycle-accurate simulation, multi-seed FPGA place-and-route, and CoreMark CRC validation. Only hypotheses that beat the current champion get merged; everything else gets discarded. Starting from 301 iterations/second, the system hit 577 iter/s (+92%) across 73 attempts in 9.8 hours — producing a design 26% faster and 40% smaller in LUTs than the baseline. The insight the author drives home is that the real innovation isn't the AI agent — it's the verifier. The orchestrator is hardcoded to prevent agents from manipulating their own evaluation gates, a simple but critical design constraint that turns a creative process into a trustworthy one. Without a rigorous verification harness, agent-driven optimization becomes a confidence trick. This is early but fascinating proof that AI-driven hardware design loops can produce commercially meaningful gains. The repo uses Claude Code or Codex as the coding agent, SystemVerilog for the RTL, and standard open-source EDA tooling (Yosys, nextpnr, Verilator). It's a compelling template for anyone building agentic optimization loops where correctness matters.