Tag
The paper presents Task Model Induction, a method that decomposes screen recordings of real computer work into separate tasks and goal trees to better train AI agents, improving performance by 30% on unseen tasks compared to current summarizers.
AutoMedBench is a workflow-aware benchmark for autonomous medical-AI research, evaluating agents across five stages on diverse medical imaging tasks. Stage-level scoring reveals validation as the weakest stage, highlighting the need for reliable verification in agentic workflows.
The article argues that companies are collections of algorithms and AI will soon optimize every component, leading to a wave of consulting-led transparency and efficiency.