Recent coverage in The Download highlights persistent unknowns around GLP-1 weight-loss drugs. Eli Lilly, the company behind Mounjaro and Zepbound, recently became the first healthcare company in the world said to have achieved a trillion-dollar valuation. New research indicates GLP-1 drugs do not appear to help people with Alzheimer’s disease, and clinicians report potentially dangerous levels of weight gain in people who stop taking the drugs when they become pregnant. Researchers also worry about postpartum use to shed pregnancy weight without fully understanding possible risks. The piece notes that it first appeared in The Checkup, MIT Technology Review’s weekly biotech newsletter, and links to related reporting on heart and brain health, hunger science, online trends, and experimental approaches.
On the economic front, the newsletter frames the question of whether to be optimistic, pessimistic, or nuanced about Artificial Intelligence’s impact on markets. It announces a subscriber-only Roundtables conversation led by Mat Honan with David Rotman and Richard Waters, scheduled for 1pm ET on Tuesday December 9, as part of the Financial Times and MIT Technology Review “The State of AI” partnership. The series aims to explore global effects of artificial intelligence across different markets, and readers are invited to register for the event and for a series of discussions that have been running over the past month, with future editions distributed every Monday.
The edition also curates must-read reporting and analysis across technology and geopolitics. Highlights include coverage of tech billionaires preparing to oppose AI regulation ahead of the 2026 US midterm elections, EU proposals to hold social platforms liable for financial scams, concerns about a humanoid robot bubble in China where more than 150 companies are building similar machines, reporting on a Myanmar scam compound, and increased investment in submarine drones. A separate feature examines SWE-Bench, a benchmark launched in November 2024 to evaluate coding skill in models; it has quickly become influential, with OpenAI, Anthropic, and Google citing scores, but entrants are gaming the system and observers question whether it truthfully measures model quality. The newsletter closes with lighter items under “We can still have nice things.”
