Inside SOTA Deep Research: Native Model Training and 150 Sub-Agents
SimonShaoleiDu · x · 2026-07-30
In a recent interview, Simon Shaolei Du shared the architectural secrets behind his team's SOTA performance in deep research and prediction:
- Native Model Training: They focused on training the model itself, rather than just building external scaffolding.
- Massive Multi-Agent Concurrency: Up to 150 sub-agents can run on top of the model to collaboratively process tasks.
- Independent Verification Layer: Verification is treated as its own distinct layer within the system, rather than a simple final prompt step.
Related event: Simon Du Reveals SOTA Deep Research via Model Training(2 posts)→
More from coding & agent
- Codex Autoresearch: Autonomous Code Iteration Until Targets Met, 2k Stars on GitHub — tom_doerr · 2026-07-30
- Deep Agents v0.7 Released: Slashes Default Context Tokens by 65% — hwchase17 · 2026-07-30
- Developer Builds Fully Procedural 3D RPG Game Using Claude — chrisfirst · 2026-07-30
- jasonkneen Launches Browser Agent with 3D Voxel Avatars — jasonkneen · 2026-07-30
- Open Source Sticky Notes Plugin for Hermes Desktop Manages Fleeting Ideas — Teknium · 2026-07-30
- Agensis Launches Shared Workspace for Humans and AI Agents — jasonkneen · 2026-07-30