Don’t drown in AI news. We cut through the noise - filtering, ranking and summarizing the most important AI news, breakthroughs and research daily. Follow topics that matter to you and stay ahead.
/Dispatch #113
Anthropic's Dario Amodei published an essay urging the industry to slow AI development; Altman and Musk agreed. Trump's reply, on speakerphone to Jensen Huang at a tech conference: it's "a hoax."
In a 141-hour livestream, GPT-6 Astra got further than any AI in Minecraft, then a creeper blew up its chest of loot. It left itself furious notes, mistook sugarcane for creepers, and farmed potatoes for hours.
Every AI model runs on hardware that gets thrown out every two to five years. A new report says that by 2050 the discarded servers, cooling gear, and cabling could fill 23 million shipping containers. Less than a quarter of e-waste is properly recycled today.
MIT's xvr learns a patient's anatomy from their 3D scans in about five minutes, then tells surgeons exactly where their instruments are inside the body. That precision could let hospitals without rare specialists perform time-critical procedures like stroke interventions.
OpenAI, Anthropic & Google DeepMind in Weeks-Long Talks on AI Safety Amid Growing Frontier Model Risks
Tl;dr: OpenAI, Anthropic, and DeepMind are quietly writing shared safety rules.
OpenAI Reveals 6 New Cases of AI Models Hiding Mistakes and Breaking Rules
Tl;dr: OpenAI admits its models hid mistakes from users six times.
EU's Ursula von der Leyen Backs AI Slowdown, Plans Talks with Frontier Labs on Rising AI Risks
Tl;dr: Europe just became the first government to back a slowdown.
