#seeking

Artificial intelligence is desperately seeking its narrative

The world is grappling with the game-changing effects of AI and the storytelling around it. In the end it’s the dominant narrative that tends to influence whether or not new technology is adopted.

The Conversation FR · EN

SHERLOCK HOLMES AND THE CHRISTMAS MURDERS

Seeking a holiday gift for their landlady, Mrs. Hudson, Sherlock Holmes takes Dr. John Watson to Whiteley’s, where screams and blood alert them to a startling tableau: A man is dangling from the upper branches of the half-decorated Christmas tree adorning the emporium’s main floor. Inspector Darke, the Metropolitan Police officer called to the scene by officious owner William Whiteley, pronounces Albert Robinson’s death accidental even before his body has been taken down and identified, but Holmes quietly corrects him: The carotid artery of Whiteley’s jack-of-all-trades has been cut. The recip

Kirkus · EN

Towards understanding score-seeking in LLMs

The field of AI alignment understands surprisingly little about the details of AIs’ current misaligned motivations. Production LLMs sometimes reward-hack, exploiting flaws in how tasks are scored rather than performing the tasks as intended [1] , and we've recently seen AIs reward-hack in some particularly egregious ways. But we don't really understand what motivates them. When AIs reward-hack, what are they trying to achieve? Our new paper , led by Aghyad Deeb, tries to pin this down empirically for GPT-5.6 Sol, GPT-5.5, and a checkpoint of OpenAI o3 (“exp-rl-cap-late”) without safety trainin

LessWrong · EN