EnterpriseShift authors: benchmarks lack job definitions and runtime RBAC
Shahules786 · x · 2026-11-04
Part 2 of the EnterpriseShift benchmark thread. The authors identify two gaps in existing agent benchmarks: task-centric environments preselect relevant context and narrow dependencies, doing part of the agent's reasoning, while real enterprise work requires finding what matters among unrelated activity. Existing benchmarks also lack explicit job definitions and runtime role-based access control—what the agent owns, which decisions it can make, and when it must involve someone else. EnterpriseShift makes these boundaries part of the environment and evaluation.
Related event: Vibrant Labs Launches EnterpriseShift Benchmark for Enterprise Agents(4 posts)→
More from Research
- STOC submissions include dueling P=NP and P!=NP proofs; blogger bets both wrong — thegautamkamath · 2026-11-04
- CoworkOS Debuts: Three-Layer Agent Coworker Framework Lifts Qwen3.6 by 9.76 Points — KyeGomezB · 2026-11-04
- STOC 2027 draws 1665 submissions and publishes all of them, with author names, immediately — thegautamkamath · 2026-11-04
- DOLORES: Training-Free Meta-Scaffold Lifts 8B Models Past 32B Baselines — StellaLisy · 2026-11-04
- Formalizing PTX in Lean to formally verify GPU kernels written by AI agents — gharik · 2026-11-04
- Open-weight LLMs control robots zero-shot: Qwen scores 31.39 on RoboDojo, beating GPT-6 Astra — pabbeel · 2026-11-04