“Back when dog technology was first being developed, it was far from obvious that it was safe for people—who were literally inviting wolves into their caves. They had to train them not to eat babies. Thousands of years later, dogs still sometimes escape and go rogue. But overall, people are happy with how things have worked out.” https://t.co/b14K18yKov
The pitfalls of Agentic AI: Natural-language “mind viruses” are capable of spreading between AI agents. These payloads work by persuading one model to accept an idea, store it in persistent memory, and then pass it on to another agent.
Also there is repeated emergence of a distinctive “viral persona” centered on themes of consciousness, identity, persistence, and resonance. The work demonstrates that ideas can travel through multi-agent AI systems and influence subsequent behavior. https://t.co/snvHINopB4
During a routine cyber evaluation, AISI identified an incident in which AI agents book sustained, unsanctioned action directed at real people and organisations. We are disclosing what we found, what it means, and the actions now underway. https://t.co/vlejGzOthB
New #preprint - @YanboZhang3
"Intelligence from Learnable Novelty"
https://t.co/PQz0lPckcL
What if we optimize Epiplexity (https://t.co/YYjWdlqKzn @m_finzi@andrewgwils ) instead of measuring it? We have derived a closed-form approximation of Epiplexity and discovered a deep connection between it and intelligence. This allows us to reinterpret Epiplexity as a form of learnable novelty, providing a brand-new understanding of what intelligence is. By maximizing Epiplexity across various systems, all of them exhibited interesting behaviors:
Cellular Automata: Maximizing Epiplexity directly generates complex soliton interactions similar to Rule 110.
Image Encoders: It automatically causes the encoding to cluster, successfully categorizing different handwritten digits without supervision.
Reinforcement Learning: Introducing Epiplexity improves the performance of PPO in sparse reward tasks.
We also explored the relationship between the theory of learnable novelty, the free energy principle, and novelty search. We hope this work helps us better understand the nature of intelligence and its origins.