We're sunsetting PodQuest on 2025-07-28. Thank you for your support!
Export Podcast Subscriptions
back
“We should try to automate AI safety work asap” by Marius Hobbhahn
31:13
Share
2025/4/26
LessWrong (30+ Karma)
AI Chapters
Transcribe
Chapters
Why Should We Start Automating AI Safety and Security Work Now?
Is Automating AI Safety Work Inevitable?
Why Simply Asking AI to Do Alignment Research Isn't Enough?
How Crucial Is the Short Timeframe for AI Development?
What Takes Time to Build in AI Safety?
How Gradual Increases in AI Capabilities Impact Automation?
Does the Order of Safety Technique Development Matter?
High-Level Comments on Preparing for AI Automation
Maintaining a Lead for Defense in AI Safety
Building the Safety Pipeline: What Can We Do Today?
Evaluations, Red-Teaming, and Monitoring: Key Areas of Preparation
Interpretability, Scalable Oversight, and Computer Security: Future Directions?
Conclusion
Shownotes
Transcript
No transcript made for this episode yet, you may request it for free.