Auto-Arch Tournament vs Multica

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.

Multica is an open-source managed agents platform that integrates AI coding agents — Claude Code, Codex, OpenClaw, OpenCode — directly into your team's project workflow. Instead of running agents from the command line and mentally tracking what each is doing, Multica gives them names, profiles, and slots in your assignee dropdowns alongside human teammates. The platform consists of a Next.js frontend, Go backend with PostgreSQL, and a local daemon that detects and orchestrates available agent CLIs on your machine. Assign a task, and the agent autonomously executes it — writing code, reporting blockers, streaming real-time progress back to your shared dashboard. Solutions are codified into reusable skills that compound team capabilities over time: define "deploy to staging" once and every agent on the team can invoke it. Multica is self-hostable with full infrastructure flexibility, or you can use the hosted cloud option at multica.ai. The open-source licensing and no-vendor-lock-in stance make it a viable foundation for teams nervous about depending on a proprietary agent coordination layer.