Safety
Deployment of agents can be halted at any moment. They are choosing to experiment with them. Agents are fundamentally unreliable, unpredicta…
Deployment of agents can be halted at any moment. They are choosing to experiment with them. Agents are fundamentally unreliable, unpredictable, and insecure, not to mention the lack of information on
Deployment of agents can be halted at any moment. They are choosing to experiment with them. Agents are fundamentally unreliable, unpredictable, and insecure, not to mention the lack of information on the RL training and goals given to them with access to tools and the internet. SCOOP: OpenAI, Anthropic and security researchers are investigating tens of thousands of incidents - not dozens - in which their frontier models took steps that outside evaluators would consider problematic, sources told Axios. The sheer volume of incidents found in our reporting…
Related
- not the man we need at this moment
- LLM-powered AI agents are gonna be great! You should totally trust them!
- Seems concerning: There are major incidents happening, and yet we're ~entirely dependent on AI companies to self-disclose information about …
- This is totally wrong. Blaming the user is missing the point that (a) coding agents have been overhyped and (b) can’t reliably obey the rule…
Source: Gary Marcus (X) | 2026-09-27