Today for AI
HOT RADAR
agentHEAT 0.0°

AI CURATED HOT STORY · 10/21/2026

OpenAI Uses Nearly 2,000 AI Agents to Map Whole Brain in Minutes

5 reports archived0 independent sourcesupdated 10/22/2026, 18:01:53
Synthesis & Latest Updates

OpenAI, in collaboration with Scripps Research and PerturbAI, deployed nearly 2,000 GPT-Rosalind-based AI agents to analyze a whole-brain mouse CRISPR map in just 32 minutes, compressing months of manual work into minutes. This breakthrough demonstrates the significant efficiency gains of multi-agent systems in accelerating foundational biomedical research.

Using a "one-gene-one-agent" parallel architecture, the system increased citation density by approximately 700 times, strengthening evidence support. Human experts preferred the multi-agent reports over single-agent outputs in 75% of blind tests. The method successfully reproduced 48 out of 55 published findings and identified potential links between cholesterol regulation and neurodegeneration, reasoning independently of post-training data cutoffs.

While the current results focus on data interpretation and hypothesis generation requiring further experimental validation, this study proves the scalability of distributed agents for high-dimensional biological data. It offers new mechanistic clues for Alzheimer's and Parkinson's diseases, establishing an efficient paradigm for knowledge extraction and drug discovery guidance.

LATEST/The latest developments show that the multi-agent system generates "reasoning maps" linking genes, cellular responses, and literature to propose testable hypotheses. Results are open-sourced and published, validating potential in discovering new mechanisms for Alzheimer's disease, though conclusions still require experimental verification.

Hourly heat curve

168 fully observed hours
Now
0.0
Peak
49.519:00
24h change
–

The line compares only sources observed throughout; gap hours are interpolated to keep the trend continuous.

Coverage timeline

Total 5 reports · Latest first
  2. IT之家 · IT业界T2·68 pts

    OpenAI and PerturbAI Deploy ~2000 Agents to Accelerate Drug Discovery

    • Nearly 2,000 AI agents worked in parallel, reducing whole-brain genomic map analysis time from an estimated months to 32 minutes.
    • The system generates 'inference maps' linking genes, cellular responses, and literature to help scientists propose testable hypotheses.
    • Experts preferred multi-agent generated explanations in 75% of blind reviews, reflecting report quality preference rather than conclusion accuracy.
  3. Rohan Paul (@rohanpaul_ai)T2·88 pts

    OpenAI Deploys ~2K Agents to Parallel-Process Whole-Brain CRISPR Atlas

    • The 'one gene per agent' strategy decomposes massive analysis tasks into nearly 2,000 parallel sub-tasks, compressing months of work into minutes.
    • Compared to single-agent processing, the parallel architecture increases citation volume by ~700x, significantly strengthening evidence support.
    • In blind tests, human experts preferred AI agent-generated reports 75% of the time, demonstrating reasoning capabilities independent of post-cutoff papers.