Sakana AI unveils Agentic Scientist v3: parallel coding agents orchestrate the full research loop
hardmaru · x · 2026-10-15
Sakana AI (David Ha's team) released The Agentic Scientist, the third iteration of its automated research system. Key shift: from "one agent writes a paper" to a meta-controller orchestrating a pool of parallel coding agents that share memory across a four-stage pipeline (task setup, parallel investigation, evidence generation, paper writing), with rubric-based verification at each stage.
- Evidence generation: isolated agents automatically run controlled baselines and ablations to prove why an idea works — addressing the tendency of coding agents to become mere hyperparameter optimizers.
- Results: a paper co-developed with a development version was accepted at AILFE 2026; in fully autonomous mode it produced a hypernetwork-based inductive bias for fast Transformer adaptation and a meta-trained evolution strategy matching or exceeding CMA-ES on 7 of 8 Brax control tasks.
- Ablations: reliable scientific autonomy depends not only on base model capability but on the cross-agent orchestration layer.
- Open-sourced on GitHub along with the AutoIllustrate figure tool.
Related event: Sakana AI Releases The Agentic Scientist, Paper Accepted at ALIFE(6 posts)→
More from coding & agent
- Running a multi-model UI design arena: Paseo orchestrates four design drafts per model — vista8 · 2026-10-16
- Cortex A/B-tests coding-agent skills and rules by replaying developer corrections before keeping them — Samuele Giampieri · 2026-10-16
- Tool calls are the hidden bottleneck killing AI coding speed — Crabbox and 5 fixes — steipete · 2026-10-16
- Agentic sandbox WorldBox doubles GPT-6 Astra's MazeBench score, 23 to 61 gems — ZhitingHu · 2026-10-16
- Two prompts: Gemini 4 Argon turns a single-player game into full multiplayer — chongdashu · 2026-10-16
- Anthropic and Gray Swan launch $300,000 indirect prompt injection red-teaming contest — bcherny · 2026-10-16