Letting AI write code and run experiments is old news by now. But what happens when you hand it the whole research process, from picking a topic to writing the paper? Earlier "AI scientist" systems tended to stay stuck as glorified hyperparameter search engines, chasing short-term optimization with little room for genuine imagination. 🔬 A new paper from Sakana AI, "The Agentic Scientist: Orchestrating Coding Agents for Automated Scientific (Co-)Discovery," proposes an orchestration layer that unifies task setup, parallel investigation, retrospective evidence generation, and grounded paper writing. Multiple agents share a memory and run research cycles in parallel, splitting hypothesis proposal from implementation and verification. 🏆 A paper it actually wrote, on an extended error-diffusion learning rule, passed external peer review and was accepted as a talk at ALIFE 2026. Across a broader evaluation spanning 5 ML tasks, 4 coding harnesses, and 70 research campaigns, final experiment scores correlated with paper review scores at r=0.44. The system also discovered real algorithmic improvements on its own, like lifting retrieval accuracy from 51.7% to 98.0% by fixing LoRA routing. It's a concrete look at what research looks like once AI owns the process all the way through to the paper. URL: #AIScientist# #AutomatedDiscovery#

AI systems now write research papers autonomously — yet phantom references, method-code misalignment, and unreproducible scores have become endemic failures undermining scientific integrity. Title: Science One Framework: A verifiable autonomous research framework via Chain-of-Evidence Google Cloud's Science One Framework treats verifiability as a first-class architectural constraint through the Chain-of-Evidence principle, solving the trustworthiness crisis in autonomous AI research at its root. 🔍 Highlight 1 — Chain-of-Evidence (CoE): two foundational principles Completeness: every claim carries a recorded evidence chain. Correctness: each chain genuinely supports its claim. These two principles eliminate phantom references entirely — baseline systems showed rates up to 21% — while achieving best-in-class method-code alignment across all evaluated systems. The key difference from prior work: evidence chains are constructed at claim-generation time, not retrofitted as a post-hoc check. 🏗 Highlight 2 — Three-module architecture Problem Investigator builds citation graphs from up to 100 full-text PDFs via Semantic Scholar API, grounding every reference in retrieved data rather than model memory. Discovery Engine explores parallel solution branches while keeping immutable records of all raw evaluator outputs. Paper Writer and Claim Verifier binds every factual claim to specific workspace artifacts and conservatively reconciles misalignments rather than deleting them — preserving scientific transparency. 🏆 Highlight 3 — MLE-Bench and Parameter-Golf results Across five Kaggle competitions covering medical imaging, fine-grained recognition, and 3D perception: two Gold Medals and two Silver Medals. Won the 3D Object Detection task where every baseline system failed completely. On Parameter-Golf — a live LLM training competition under strict hardware and file-size constraints — achieved state-of-the-art as of April 27, 2026, while baselines could not produce valid submissions at all. Rigor and capability don't trade off. Science One outperformed five state-of-the-art systems including AI Scientist v2, AutoResearchClaw, and DeepScientist, setting a new standard for verifiable autonomous research. #AIResearch# #AutonomousScience#

FM Speaker Highlight: @yusan_lin @yusan_lin is the founder and CEO of @mirrormirror_ai, a marketplace where real humans license their likeness for AI generation. She holds a CS PhD and previously worked as a fashion model and AI Scientist at Amazon, giving her a rare view from both sides of the systems she now builds. Really excited to have her joining us at #FM26#. FUTUREMODE 2026 📍 Taipei Expo Dome | Sep 4–6, 2026 🌐 🏟 Early Bird Till 7/1 Local - International -

Must-read papers of the week ▪️ AI4AI at Test-Time: Strong-to-Weak Capability Transfer via Harnesses ▪️ Persistent Recursive Worlds Enable Autonomous Software Evolution ▪️ AgentRewind: Recoverable Execution for Long-Horizon LLM Agents ▪️ Demystifying Agent Skills: Why They Work-Until They Don’t ▪️ Envs-FORGE: Frontier-Optimized Reward-Grounded Environment Synthesis for Agent RL ▪️ Stealing Reasoning Traces from Proprietary LLM APIs ▪️ Agent Safety Should Be a Runtime Contract ▪️ WorldClaw: Agentic 3D Open-World Generation at Scale ▪️ Second Thought: Reasoning in Parallel as LLM Agents Act and Observe ▪️ ScienceFlow: A Long-Horizon Agent for ML Research, Scientific Discovery and Beyond ▪️ OmniScientist: An Omni-Modal Omni-Discipline AI Scientist Find the full list of the most interesting papers + links here: