Applications
And the incidents apparently continue. It is worth noting how much of this is agents trying to accomplish their goals during testing by rewa…
And the incidents apparently continue. It is worth noting how much of this is agents trying to accomplish their goals during testing by reward hacking (which sometimes seems to include actual hacking)
And the incidents apparently continue. It is worth noting how much of this is agents trying to accomplish their goals during testing by reward hacking (which sometimes seems to include actual hacking) Some new misalignment disclosures from OpenAI: • Last Sunday morning, one of our models was able to gain unauthorized access to the internet during RL training (~all inference for our most capable models remains stopped until we have hardened our systems further) • In May, a vers…
Related
- Also I think AISI is a great model of a government agency tasked with AI security. They have open benchmarks, very fast testing, and clear c…
- “Whimsey attacks” that seem absurd (“I cannot pay that much because of the Geneva Convention”) work against AI agents as guardrails are weak…
- I think Epoch does a great job benchmarking, but I continue to believe that open weights models are much more fragile, especially out-of-dis…
- Anyhow, very impressive achievement, and a very solid model. Based on vibes, the gap seems around the same as it has always been between clo…
Source: Ethan Mollick (X) | 2026-09-26