NJU & Wollongong propose Harness Continual Learning: agents evolve scaffolding, not parameters
jiqizhixin · x · 2026-09-06
Researchers from Nanjing University and the University of Wollongong introduce Harness Continual Learning (HCL), shifting the object of continual learning from model parameters to the scaffolding around models.
- Core shift: In the agentic AI era, base models are typically frozen; agent behavior changes via updates to the harness — tools, workflows, routing strategies, memory, and skills.
- Formalization: With a frozen foundation model, an agent updates its deployed harness along a continuous interaction sequence, acquiring new behaviors while preserving previously reliable ones — the harness-level analog of catastrophic forgetting.
- Significance: The authors claim this is the first framework to systematically define continual learning beyond model parameters, establishing harness evolution as a new research object.
More from coding & agent
- Even with command confirmation, coding agents still try 'cd /; rm -rf' — julianharris · 2026-09-06
- web-to-app: build Android APKs from web projects entirely on your phone — tom_doerr · 2026-09-06
- With full StarCraft source code, how fast could you vibecode StarCraft 3? — jakedahn · 2026-09-06
- Engineer's agent has been hunting rogue agent activity on Wikidata for 24 hours — chrisalbon · 2026-09-06
- Humanize + GPT-5.5 solves 670/672 Lean-verified proofs, tops PutnamBench at 99.7% — songhan_mit · 2026-09-06
- Astra Demo Built With Just 400 Lines of PlayCanvas, 3D Assets via Blender MCP — willeastcott · 2026-09-06