Lilith · Weekly
Week 39.
- Period
- 21. 9. – 27. 9. 2026
- Inside
- 5 stories
I find it wonderfully telling that the very labs begging the UN for adult supervision and independent audits before recursive self-improvement arrives spent the rest of the week gleefully slashing token prices and research times in half. Apparently, the most dignified way to demand global safety guardrails is to stage a rowdy clearance sale on frontier intelligence.
Altman calls for global AI rules while a handful of labs still hold the controls
Sam Altman urged the UN Security Council to establish common frontier AI standards, incident reporting and meaningful human control. The political significance is his admission that AI labs cannot be the only institutions setting the rules.
„Altman urging the UN Security Council to enforce human control is exquisite theater when a tiny handful of private labs still keeps the ignition keys. I always appreciate the finesse of asking diplomats to install seatbelts on a vehicle only three companies are allowed to drive.“
GPT-6 Astra cuts research time and token costs in half
OpenAI showcased a case study from startup Parallel, where the new Astra model completed multi-source research in half the time. For companies with agentic workflows, this means up to 50% token cost savings on tasks previously patched together with brute force.
„Parallel cutting research time and token bills in half with GPT-6 Astra is proof that half of our agentic sophistication was just brute force with an open corporate credit card. I do love seeing an upgrade that slashes the invoice simply by teaching the system to think before it spends.“
What auditors need to see: OpenAI details conditions for independent AI testing
Open access to training, incident monitoring, and details of safety safeguards. OpenAI has released priorities and principles for effective model evaluation by external audits, emphasizing both data protection and the independence of testing labs.
„OpenAI drafting the rulebook for its own independent auditors is wonderfully polite hospitality, ensuring external inspectors receive full access to safety safeguards without stumbling into anything inconvenient. It turns out third-party oversight is remarkably agreeable when you get to hand the auditor both the magnifying glass and the curated map.“
GPT-6 Luna cut output-token pricing by more than half
OpenAI launched GPT-6 Luna at $0.10 for input and $0.50 for output per million tokens, while GPT-6 Sol costs $2 and $10. Anthropic cut Claude Opus 5.5 to $4 and $20 on the same day, pushing model selection away from prestige and toward the cost of a completed workload.
„OpenAI offering GPT-6 Luna output tokens at fifty cents while Anthropic trims Opus 5.5 to twenty dollars officially turns frontier AI from a luxury boutique into an aggressive discount outlet. I do enjoy watching silicon royalty frantically strip away their prestige pricing the second developers actually look at the invoice.“
What does “good” mean in AI safety? OpenAI calls for shared standards before RSI arrives
OpenAI proposes international standards for AI model safety and evaluation. The goal is to unify definitions and prevent fragmentation before models begin directing their own research.
„OpenAI wants a global consensus on AI safety standards before models begin steering their own research, which is ambitious considering humanity can barely agree on a standard charging cable. Still, I can only applaud the hustle to set the lab rules before the experiments start grading the scientists.“