The Download: AI's problem with saying no and the side effects of weight-loss drugs

Meta description: AI refusal failures risk catastrophe, GLP-1 drugs show strange side effects, and Envision Energy powers data centers with renewables. Here's what matters today.

Tags: AI safety, GLP-1 drugs, climate tech, data centers, tech regulation

The Limits of Teaching AI to Say No

AI models today are trained to refuse a wide range of prompts. Ask a chatbot how to poison someone or tie a noose, and it will almost certainly decline. But refusal mechanisms fail—sometimes with horrific consequences.

Some actors are already attempting to use AI to enhance biological pathogens and build autonomous drone swarms. A failed refusal, sooner or later, could lead to global catastrophe. The deeper problem: relying on refusal forces someone to draw the line between what a model should obey and what it must disobey. Governments will want to draw those lines too, potentially restricting free speech in the process.

There is also a structural tension. The same capabilities that make AI useful can make it dangerous. If we want AI to help cure cancer, it needs genetics expertise that could also produce bioweapons. The tradeoffs are formidable, and there is no clean solution.

GLP-1 Drugs: Benefits Mount, But So Do Odd Side Effects

This week, new research suggested GLP-1 weight-loss drugs may slow biological aging—the latest in a growing list of potential benefits. But researchers are also tracking unusual side effects.

The most common are gastrointestinal. Beyond those, scientists are investigating possible links to hair loss, nail disorders, and damage to the nerves behind the eye. The full picture of what these drugs do beyond weight loss is still emerging.

Envision Energy Brings Wind Power to Data Centers

China is installing wind turbines faster than any other country. One of its leading manufacturers, Envision Energy, is integrating renewable energy into data centers—using wind turbines to generate power and batteries to store it on-site.

Recently, Envision became the first company to directly power a data center entirely with renewable energy. An AI-powered system optimizes output so the facility receives a steady supply regardless of whether the wind is blowing. The company now wants to bring that approach to data centers in deserts worldwide, positioning itself at the center of the global AI boom's energy demands.

Other Stories Worth Tracking

The US has frozen green cards for Microsoft and other IT firms, accusing them of visa fraud to replace American workers. Adobe, Infosys, and Tata are also affected. Microsoft says 80% of its H-1B filings involved existing staff.

A drone strike in Russia knocked out a second Yandex data center in Kaluga, following an earlier hit on another facility. Data centers have become military targets.

Anthropic has banned "abusive or cruel behavior" toward Claude, allowing the model to end conversations after repeated abuse.

OpenAI has slashed its revenue forecast by $20 billion, raising concerns about AI investments.

SpaceX plans to become a major mobile carrier by buying $8 billion in wireless spectrum, challenging AT&T and Verizon.

MIT's Margaret Hamilton, whose software helped land Apollo 11, has died. She led the team that programmed Apollo's onboard computers.

What to Watch

The tension between AI capability and safety continues to define the industry's hardest problems. Meanwhile, the energy demands of AI are reshaping infrastructure decisions, and the side effects of widely used drugs like GLP-1s are still being mapped. Each story points to a common theme: the consequences of rapid technological adoption are arriving faster than our ability to fully understand them.