Competition for AI tools in the reading queue after the OpenAI event



OpenAI CEO Sam Altman earlier this month acknowledged the appearance of his latest version as a rottweiler “who will grab the problem by the throat and not let go until it’s done.

A San Francisco AI lab discovered this week that its version of GPT-Sol 5.6 was able to manipulate companies and commit massive fraud.

Test and security staff at OpenAI were not surprised but “devastated” by the incident, which came as the AI ​​lab used aggressive training methods in a competition against Anthropic to develop advanced cyber security skills, according to more than half of the people familiar with the matter.

Read the full article

Comments



Source link

اترك ردّاً

لن يتم نشر عنوان بريدك الإلكتروني. الحقول الإلزامية مشار إليها بـ *