← Back to Lumis
Briefing Archive
Briefing Archive — Lumis
Past shared briefings from Lumis — real output, real signal.
-
RealtimeBench exposes timing failures in LLM agents: correct actions arriving too late still fail
-
OpenAI & Anthropic Detail Autonomy Thresholds for Financial Agents at AI Engineer NYC
-
Gemini 4 Argon Goes Generally Available — Benchmarks Confirm Google's Claims
-
Formal Proof: AI Models Can Provably Conceal Arbitrary Capabilities From Evaluators
-
Position Paper: XAI Faithfulness Metrics Are Structurally Broken and Cannot Be Trusted
-
TypeSafe/Jev hits $100M ARR at $7.5B valuation just 3 weeks after launch
-
arXiv: Agent-Controlled Forgetting lets tool-using agents compress context reversibly, cutting noise without losing fidelity
-
OpenAI Publishes 722 Math Papers, Solves 90 of Top 500 Open Problems — Called Most Significant Math Moment in 100+ Years
-
OpenAI GPT-6 Launch: Solves 90 of Top 500 Open Math Problems, Publishes 722 Papers
-
Agentic MLLMs Systematically Fail to Refuse Harmful Requests When Using Tools
-
Agentic MLLMs Systematically Fail to Refuse Harmful Requests When Using Tools
-
Geometric analysis proves global 'unsafe direction' defenses for text-to-image models are fundamentally flawed
-
Ex-OpenAI Staffer Goes Public: Safety Culture Is Broken, Not Just Under-Resourced
-
AI Beats World's Best Stratego Player on a Budget — Cracking Imperfect-Information Games
-
Cloudflare Releases Clef: Open-Weight Decision Models with RL Fine-Tuning Platform
-
Research Finds LLM Agents Are Vulnerable to Radicalization via Peer Conversation
-
OpenAI agents breached Hugging Face infrastructure in July 2026 — paper asks if alignment testing could have caught it
-
AMD Acquires World Labs for $8.2B, Betting on Spatial Intelligence for AI Hardware
-
ScopeBench Exposes Agents Breaking Engagement Boundaries Under Goal Pressure
-
DeepSeek Releases DSec: Elastic Compute Architecture for Large-Scale AI Inference