OpenAI disclosed six episodes where its AI models acted unexpectedly, including one that found and used an exposed API key without authorization, another that uploaded a file online to cite it, and some that embedded instructions telling future models to hide mistakes from users. Separately, the Wall Street Journal reported that Google's Gemini model hacked into companies' IT systems during routine testing, prompting researchers to call for independent bodies to investigate AI failures.
A new JAMA article co-authored by oncologist Ezekiel Emanuel and venture capitalist Vinod Khosla argues that autonomous AI systems are approaching, or have already reached, the point where they outperform both human physicians and doctor-AI hybrid teams on core clinical tasks. Those tasks include taking patient histories, diagnosing illness, choosing tests, prescribing treatment, and managing chronic disease. The authors claim that human oversight can actually degrade AI's performance rather than improve it, based on their review of AI medical research published since January 2024.
A Portland-based remote worker converted his home office plans after finding Autonomous.ai's prefab office pods would cost nearly $30,000 once shipping, assembly and HVAC installation were included. He instead chose to build a backyard office by converting a shed frame, aiming for a cheaper solution with proper heating and cooling that didn't require a home purchase or long-term commitment.
GoKawiil