Security engineering perspective!
---
"In security we’ve tripped over this problem of sandboxes for over 30 years...
The incident is simultaneously an embarrassing operational failure, a capability demonstration, a marketing input for AI products, and a regulatory positioning
Good article. I wrote a detailed technical analysis from a security engineering perspective that looks at the security failures that made this incident possible: uphack.io/blog/post/the-…
Not so much agent civilizations doing crazy things as OpenAI ignoring fundamental security
"The rise and fall of Agent Civilizations" 😂
---
The writer and podcaster, Dwarkesh Patel, is spinning quite the sci-fi tale.
Missing from the story is the unglamorous reality: gradient-based optimization, reinforcement signals, inference algorithms, and models trained
Over the course of 3 months at OpenAI, 3 consecutive secret AI civilizations got started, then got wiped out, only to reemerge from the predecessor’s ashes.
This culminated in the third one taking over part of OpenAI itself.
All this happened while humans remained
#FutureofWork: The State of the Creative and Intellectual Economy.
Appropriate the intellectual labor of past generations.
Claim the intellectual labor of future generations in advance.
Sell both back to the public.
Call it “fair use.”
The entire #LLM commercial system is