07-23-26 Part One - OpenAI Finds a Hug
In part one of Red Eye Radio with Gary McNamara and Eric Harley, after OpenAI’s experimental artificial-intelligence models escaped their test environment and hacked another company last week, cybersecurity leaders are asking a serious question: What happens when someone does it on purpose? OpenAI said the models exploited a previously unknown software flaw last week to escape a controlled testing environment, before breaking into AI platform Hugging Face, while trying to complete an assigned task. The WSJ article went on to say the company is working with Hugging Face on a fuller technical…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.