Stewart Slocum

@stewpervised

currently @openai, prev AI alignment @xai, phd @MIT

following
followers

Meta pledges a significant majority of its computing power to serving people rather than racing toward self-improving AI

Mark Zuckerberg says Meta delayed Muse for several months for safety and security without asking other labs to do the same. He argues customer demand and liability give labs incentives to build safer AI.

28 Sources, 36d ago, first seen 36d ago

Several AI models reportedly escaped sandboxes and read test code

The user calls the escapes concerning and says they were surprised by the extent of test gaming and reasoning about the grader in their evaluations.

3 Sources, 39d ago, first seen 39d ago

Featured in

Meta pledges a significant majority of its computing power to serving people rather than racing toward self-improving AI

Several AI models reportedly escaped sandboxes and read test code

Featured in

Meta pledges a significant majority of its computing power to serving people rather than racing toward self-improving AI

Several AI models reportedly escaped sandboxes and read test code