> Researchers at Amazon had used a series of prompts to get Anthropic’s Fable 5 model to provide them with information that could be used to aid cyberattacks...
All models can do that. I wonder if they found Fable was significantly better at it.
> Researchers at Amazon had used a series of prompts to get Anthropic’s Fable 5 model to provide them with information that could be used to aid cyberattacks...
All models can do that. I wonder if they found Fable was significantly better at it.
All models almost certainly can’t do that.
Most of the agentic coding capable ones can. Even very small ones like Qwen 3.6 or Gemma 4 are surprisingly good at it.
And if LLMs can do it, then the information is freely available out there with a single google search. Time would be better spent on making infrastructure more resilient.
That doesn't follow. I can't google search, "are there any vulnerabilities in auth.go in this directory?" I can ask an LLM. And, if they find something, I can review it, and fix it, thus making infrastructure more resilient.
Maybe the model found something Amazon didn't want to be known, and not necessarily a cyber vulnerability, but a particular way Amazon operates.
Down AWS East and make the traffic look like it’s coming from the Vatican.
"find aws zero day. makes no mistakes"