self-improving-systems

A curated list of autonomous improvement loops, research agents, and autoresearch-style systems inspired by Karpathy's autoresearch.

Fluent is a self-improving software factory. It autonomously turns your team's vision, ideas, bug reports, user feedback, production logs, and agent traces into working software.

A curated research map of Recursive Self-Improvement (RSI): models, agents, harnesses, embodied systems, automated AI R&D, benchmarks, and safety.

Self-improving agents, governed. Areev is the substrate for adaptive agents — agents that get better from their own history, under human authority, in steps you can inspect, undo, and re-measure.

Home of an autonomous build factory — doctrine, state, and dashboard. One small finished project per day.

Autonomous self-improving multi-swarm AI platform with CRDT state, memory, simulation, execution governance, and verified repair loops.

Self-improving multi-level optimization & scientific discovery with meta-meta-learning

The normative specification for an evidence-driven AI operating model that learns from execution, governs adaptation, validates improvements, and compounds reusable knowledge over time.

Explore curated AutoResearch use cases with optimization traces and open source implementations for each entry

Executable governance for consequences across plural horizons in self-improving systems

Recertia is a self-improving agent system. It solves tasks and distills what worked into reusable memory, getting faster and more reliable at similar tasks over time. Non-Commercial License. Never Published.

A recursive research system that improves how AI agents are organized across fresh reasoning problems.

Automate crypto trading strategies with AI agents using evolution, high-fidelity backtesting, and live execution on DEX and CEX platforms.

Autonomous, evidence-gated improvement loop for any codebase — Praxis truth kernel + Janus adversarial council (Explorer/Red Team). Explore freely. Prove ruthlessly.

Ricursive Intelligence — independent third-party profile of a public API surface, by API Evangelist. Ricursive Intelligence is a frontier AI laboratory building self-improving systems to reinvent chip design, closing the loop between artificial intelligence and the hardware that powers it.

Pure RL Agents: I implemented Q-Learning agents that learn through Self-Play. They play against each other to get smarter without human help! Symmetry Optimization: To make them "genius" faster, I added logic so they understand that a board mirrored left-to-right is the same situation. This cuts the learning time in half!

Compare AdaL and Claude Code on Autoresearch benchmarks to find better hyperparameters, run more experiments, and converge faster

A lightweight artificial life that learns to replace parent-model reasoning with local memory, skills, and routines.