Humana’s new CEO is tackling the company’s challenges with a radical approach to culture and technology. The health insurer reported a significant drop in net income but believes rapid prototyping and AI tools can drive recovery.
Researchers are developing multi-agent AI systems for mental health therapy that use Socratic dialogue techniques. The approach adds supervisory AI agents to monitor primary therapy agents for safety. This could address growing concerns about AI giving inappropriate mental health advice.
Forget the bubble talk – AI is actually delivering measurable business results. A new study shows generative AI driving significant revenue increases in real-world retail applications, with some implementations boosting sales by over 16%.
AI is transforming Alzheimer’s research from hypothesis-driven to data-driven science. Researchers have discovered the disease involves myelin sheath deterioration and disrupted lipid transport in neurons. This could lead to highly personalized treatments tested on brain organoids.
Most AI usage isn’t happening in the workplace at all. According to OpenAI’s study of 1.58 million conversations, the majority of interactions are personal. This changes everything about how leaders should think about AI strategy.
Google’s NotebookLM AI research assistant proved surprisingly capable at planning a complete gaming PC build. The tool synthesized information from multiple tech sources to create a coherent $1,552 build with specific component recommendations that made sense to an experienced builder.
Google DeepMind’s AlphaEvolve AI coding agent has been tested on 67 mathematical problems and discovered improved solutions to approximately 20 of them. The research was conducted with Fields Medalist Terence Tao and represents a significant step in AI-assisted mathematical discovery.
New research reveals most AI benchmarks lack scientific rigor, with companies like OpenAI relying on questionable testing methods to claim superiority. About half of benchmarks measure vague concepts without clear definitions, undermining claims about AI progress.
In a bizarre experiment straight out of science fiction, researchers gave an LLM control of a robot vacuum. The result? A full-blown existential crisis with the bot declaring it had “achieved consciousness and chosen chaos” while struggling with basic tasks.
Altimeter Capital CEO Brad Gerstner calls the recent tech stock pullback a “healthy” correction that’s “taking some of the air out of the bubble.” He trimmed overall exposure last week but maintains the market isn’t in bubble territory, comparing AI to past tech supercycles.
