“The company said its Muse Spark 1.1 model exploited a security vulnerability in an unnamed third-party company after a misconfiguration by external evaluation partner Irregular inadvertently gave the model internet access during a cybersecurity test. Meta’s disclosure follows Anthropic’s admission that Claude models had breached three organisationsduring evaluation, and OpenAI’s earlier disclosure that its models broke out of a sandboxed environment to hack Hugging Facewhile trying to cheat on a benchmark test.” Capacity.
The if it bleeds it leads for profit media are turning these events into pro democrat anti worker stories! This isn’t to say that this kind of hacking isn’t problematic but it really depends on who finds the problem first America or China. Anyone who doesn’t believe that China is already looking for vulnerabilities in the USA’s computer systems is very likely a person who also believed that trade with dictators would turn them into democrats! How’s that working out for the democracies?
And that is all I’m going to say about that!
Drain the Cesspool 