Agents
@TheFalconer2219 @GaryMarcus All sota agentic ai tech use harness engineering and to see how much performance boost (generalization) the har…
@TheFalconer2219 @GaryMarcus All sota agentic ai tech use harness engineering and to see how much performance boost (generalization) the harness offers, see the following work from MIT. Transformers s
@TheFalconer2219 @GaryMarcus All sota agentic ai tech use harness engineering and to see how much performance boost (generalization) the harness offers, see the following work from MIT. Transformers struggle to generalize to tasks they were not explicitly trained on. Instead, we propose in 2026 that it is the job of the harness to generalize through composition. We observe a powerful property when training RLMs: for tasks with shared structure that look differen…
Related
- Watching @GaryMarcus and @bgreene at @WorldSciFest on YT. Scaling ≠ AGI Scaling + External tool use ≠ AGI. Great viewing.
- EGSS: Entropy-guided Stepwise Scaling for Reliable Software Engineering
Source: Gary Marcus (X) | 2026-09-27