AI Daily — October 12, 2026

Models & Research

Statistical re-examination of METR’s AI time-horizon plot — Researchers recomputed METR’s 50% time horizons across 228 tasks and 26 AI systems using splines and item-response theory, relaxing the assumed relationship between human task time and AI difficulty. The work examines how accurately time horizons are estimated and how well they represent AI capability. arXiv ↗

My takeaway: Human difficulty does not necessarily ties to AI difficulty.

OrcaCyber Zero 1.5: Specialized AI for Security Research and Testing — OrcaCyber Zero 1.5 is designed for vulnerability research, exploit development, penetration testing, and security auditing. It supports a one-million-token context window, tool calling, and structured outputs for complex security workflows. Orca reports scores of 95.8% on the evaluable CVE-Bench subset and 100% on Cybench with unrestricted agent execution; these are vendor-reported results. Access is restricted to approved security researchers and authorized testing through OrcaRouter’s Security Research tier. OrcaRouter ↗

My takeaway: I think 1-million-token context window could support analysis of large codebases. According to OrcaRouter, the model is designed for vulnerability research and security auditing. Access is restricted to trusted security researchers, red teams, and authorized security testers through the Security Research tier.

Summaries are AI-generated and may contain errors — always verify against the linked original. Each story links to its source, which holds the copyright. Outlet names are shown for attribution only and do not imply any endorsement or affiliation.

Disclaimer: The views expressed in My Takeaway are my own personal opinions and general observations on industry trends. They are not intended to criticize, disparage, or make factual claims about any specific company, product, or platform. Any platform names mentioned are referenced solely for illustrative and informational purposes.