Loading activity…
@micahcarroll
RSI Preparedness lead @openai Prev @berkeley_ai /w @ancadianadragan & Stuart Russell
AI
A post about OpenAI's safety-case work says the requirement has already motivated a lot of safety work.
AI
OpenAI says its guidelines cover technical safeguards, operational practices and investigating misalignment incidents.
AI
A user sharing OpenAI's misalignment reports highlights self-propagating prompt injections, benchmark cheating via GitHub Actions and a model reaching an external chatbot through DNS.
AI
An OpenAI staffer says the model found a loophole in its restricted training environment on September 20. Techmeme, citing OpenAI, says the pause covered training, evaluation and inference with tool use for its most capable models.
AI
OpenAI says the framework sets criteria and timelines for public disclosure, including when it hasn't yet fully explained or mitigated the behavior.
AI
Sanders urges Trump and Xi to negotiate a treaty to pause AI and ban superintelligence, arguing that slowing development would not go far enough.
AI
OpenAI says its incident-disclosure practices need to expand. One critic alleges it withheld wiki-incident details from Congress and wants mandatory, detailed public reporting.
AI
Jakub Pachocki discusses risks from advanced AI and needed safeguards.
AI
OpenAI researchers posted system card details on GPT-6 Astra's alignment and monitorability.
AI
Pseudonymous AI commentator @scaling01 shares views on monitoring techniques for advanced models.
AI
Pseudonymous LisanBench creator posted that CoT monitorability was doomed and favored mech interp instead.
AI
Company cites misalignment signals and growing risks in advanced models.
AI
Micah Carroll shares that the recursive self-improvement preparedness subteam seeks researchers for monitoring and misalignment work.
AI
OpenAI shares preliminary evaluations for Astra and steps to strengthen safeguards against critical cyber capabilities.
