Auto-Arch Tournament vs Kelet

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.

Kelet is a root-cause analysis agent for LLM applications that goes beyond trace visualization. Where most observability tools stop at showing you what happened, Kelet automatically reads your traces, cross-references failure patterns across thousands of sessions — thumbs-down ratings, abandoned conversations, LLM-judge flags — generates root cause hypotheses, and produces targeted prompt patches to address them. The workflow is: connect your traces (LangSmith, Langfuse, or direct API), let Kelet ingest your failure signals, and receive a prioritized list of failure clusters with explanations and draft prompt fixes. SOC 2 Type II certified, read-only access to traces — nothing is mutated. The indie team positions it as the missing "closing of the loop" in LLM observability: most teams can detect failures but have no systematic path from detection to fix. The HN thread surfaced a real pain point: teams know their chatbot is failing somewhere, but diagnosing which prompts, tools, or routing decisions are responsible requires manual trace archaeology. Kelet automates that archaeology and produces actionable output, not just dashboards.