# DuoDiscover: Scaling Autonomous Discovery Beyond the Evaluation Bottleneck

arXiv: https://arxiv.org/abs/2610.17184 · HTML: https://mlsanity.com/paper/2610.17184/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 48 of its 74 references that are in the map.
- First posted on arXiv 12 Oct 2026.

## Builds on (most cited first, 30 of 48)

- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [Reflexion: Language Agents with Verbal Reinforcement Learning](https://mlsanity.com/paper/2303.11366/) (2023)
- [Scaling Laws for Reward Model Overoptimization](https://mlsanity.com/paper/2210.10760/) (2022)
- [Teaching Large Language Models to Self-Debug](https://mlsanity.com/paper/2304.05128/) (2023)
- [AlphaEvolve: A coding agent for scientific and algorithmic discovery](https://mlsanity.com/paper/2506.13131/) (2025)
- [Hyperband: A Novel Bandit-Based Approach to Hyperparameter Optimization](https://mlsanity.com/paper/1603.06560/) (2016)
- [BOHB: Robust and Efficient Hyperparameter Optimization at Scale](https://mlsanity.com/paper/1807.01774/) (2018)
- [CodeT: Code Generation with Generated Tests](https://mlsanity.com/paper/2207.10397/) (2022)
- [Zero-Cost Proxies for Lightweight NAS](https://mlsanity.com/paper/2101.08134/) (2021)
- [ShinkaEvolve: Towards Open-Ended And Sample-Efficient Program Evolution](https://mlsanity.com/paper/2509.19349/) (2025)
- [Learning to Discover at Test Time](https://mlsanity.com/paper/2601.16175/) (2026)
- [How Powerful are Performance Predictors in Neural Architecture Search?](https://mlsanity.com/paper/2104.01177/) (2021)
- [SpecRover: Code Intent Extraction via LLMs](https://mlsanity.com/paper/2408.02232/) (2024)
- [FlashAttention-4: Algorithm and Kernel Pipelining Co-Design for Asymmetric Hardware Scaling](https://mlsanity.com/paper/2603.05451/) (2026)
- [ThetaEvolve: Test-time Learning on Open Problems](https://mlsanity.com/paper/2511.23473/) (2025)
- [AdaEvolve: Adaptive LLM Driven Zeroth-Order Optimization](https://mlsanity.com/paper/2602.20133/) (2026)
- [EvoX: Meta-Evolution for Automated Discovery](https://mlsanity.com/paper/2602.23413/) (2026)
- [Barbarians at the Gate: How AI is Upending Systems Research](https://mlsanity.com/paper/2510.06189/) (2025)
- [CodeMonkeys: Scaling Test-Time Compute for Software Engineering](https://mlsanity.com/paper/2501.14723/) (2025)
- [Astra: A Multi-Agent System for GPU Kernel Performance Optimization](https://mlsanity.com/paper/2509.07506/) (2025)
- [CUDA Agent: Large-Scale Agentic RL for High-Performance CUDA Kernel Generation](https://mlsanity.com/paper/2602.24286/) (2026)
- [Goodhart's Law in Reinforcement Learning](https://mlsanity.com/paper/2310.09144/) (2023)
- [DeltaEvolve: Accelerating Scientific Discovery through Momentum-Driven Evolution](https://mlsanity.com/paper/2602.02919/) (2026)
- [FlashInfer-Bench: Building the Virtuous Cycle for AI-driven LLM Systems](https://mlsanity.com/paper/2601.00227/) (2026)
- [PACEvolve: Enabling Progress-Aware Consistent Evolution](https://mlsanity.com/paper/2601.10657/) (2026)
- [AVO: Agentic Variation Operators for Autonomous Evolutionary Search](https://mlsanity.com/paper/2603.24517/) (2026)
- [ProxyBO: Accelerating Neural Architecture Search via Bayesian Optimization with Zero-cost Proxies](https://mlsanity.com/paper/2110.10423/) (2021)
- [SWE-fficiency: Can Language Models Optimize Real-World Repositories on Real Workloads?](https://mlsanity.com/paper/2511.06090/) (2025)
- [Pythia: A Customizable Hardware Prefetching Framework Using Online Reinforcement Learning](https://mlsanity.com/paper/2109.12021/) (2021)
- [SOL-ExecBench: Speed-of-Light Benchmarking for Real-World GPU Kernels Against Hardware Limits](https://mlsanity.com/paper/2603.19173/) (2026)

## Abstract (arXiv)

Evaluation has been central to advances in AI, and it is equally critical to autonomous discovery. Yet scaling autonomous discovery to real-world engineering challenges faces a fundamental bottleneck: the cost of evaluating progress. In real settings, a single true evaluation, such as a full benchmark or serving trace, can take hours, so a search can afford only a handful of true evaluations. We present DuoDiscover, a search system that jointly discovers what to optimize and how to evaluate progress efficiently, through three mechanisms. First, each worker in a fleet of coding agents co-discovers an optimization target and a fast, tailored proxy evaluator, which we call a yardstick. The agent then hill-climbs the yardstick rather than the true evaluation. Second, since search can exploit gaps between a yardstick and the true objective, fidelity tests continually recheck each yardstick's ranking and invalidate any that drift. Third, the fleet is bottom-up, with no central decomposition of the problem. A lightweight orchestrator spawns and terminates agents, briefs running ones, and seeds new agents with others' confirmed work. DuoDiscover enables autonomous optimization from software to chip architecture. We demonstrate DuoDiscover on two repository-level benchmarks and four GPU and TPU kernel suites covering hundreds of kernels. On SWE-fficiency, it doubles the average speedup of human experts. On SOL-Exec, it ranks first on 121 of 235 kernels and in the top three on 212. On the heavily optimized vLLM and SQLite engines, DuoDiscover achieves 1.92× and 1.38× gains over the base code. On a CPU microarchitecture optimization task with evaluations of up to 28 hours, DuoDiscover surpasses both AlphaEvolve-based ArchAgent v2 and the best human design, with 1.65× ArchAgent v2's gain in 16.7 days versus more than 40 days, advancing the frontier of autonomous chip design.