Auto-Arch Tournament vs Superpowers

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.

Superpowers is an open-source collection of composable "skills" — structured workflow files — that guide coding agents like Claude Code and Cursor through disciplined software development. Where most agentic coding setups let the model improvise, Superpowers enforces a mandatory sequence: clarify requirements, design, plan into 2-5 minute tasks, execute with TDD, review. Skills are "mandatory workflows, not suggestions." With over 152,000 GitHub stars and climbing fast, Superpowers has become a reference implementation for the growing "how do you keep your agent from going off the rails" problem. The framework implements RED-GREEN-REFACTOR test cycles, forces complexity reduction at each step, and builds in checkpoints where the human reviews before the agent continues. The result is agents that can work autonomously for hours without drifting. The timing is right: as Claude Code, Codex CLI, and Cursor all become more powerful, the bottleneck is shifting from "can the model write code" to "can I trust it to work autonomously without blowing up my codebase." Superpowers is a direct answer to that, and the star count suggests developers are starving for it.