Jev Meets the Trolley Problem
This article explores the emergence of 'decision models'—AI systems designed to make choices rather than generate text—by testing them against the classic ethical Trolley Problem. The author compares the performance of new models like Jev against established LLMs.
Why it matters
The shift toward AI models that output probabilistic choices rather than just text represents a significant evolution in how AI might be integrated into real-world decision-making systems.
If you’re active in AI or software development circles, you’ve probably heard about Jev in the last few weeks. It seems like everyone, myself included, is excited by the possibilities of Jev and the Jev-like models: fast, cheap, general decision making. Sometimes dozens or even hundreds of times cheaper and faster than an LLM like Claude, purportedly with the intelligence of a (lower-tier) frontier model. These models don’t output words, they output choices, with meaningful probabilities attached so you can understand how confident the model is in its choice. People are calling this new breed of model ‘decision models’ . So naturally, I decided to ask Jev and its ilk to decide the Trolley Problem.
Get smarter about the news
Sign up free for a feed built around what you actually care about, Dive Deeper research on any story, and the full text of every article.
Create free accountAlready have an account? Sign in