A clearer view of a changing world.
Loading the edition…
Section
How machine intelligence is built, governed and put to work.
Large employers describe moving software agents from pilot projects into routine back-office tasks. The harder problems are oversight, accountability and what happens to the people who used to do the work.
The Daily Edition Illustrative Desk · 2 min read · Complete coverage
An agreement among national safety institutes would exchange evaluations of advanced models before release. Two of the largest AI-producing countries have not signed.
2 min read
The number of signatories rose to forty-two by evening in this illustrative scenario. The two largest non-signatories remain outside.
Evening Update
The assistant that knows your calendar, your inbox and your preferences is technically close. Whether people want one, and who it works for, are the questions that will decide what gets built.
Artificial Intelligence
Humanoid robots walk convincingly in videos and fold laundry in laboratories. The gap between demonstration and deployment is wider than the videos suggest, and the reasons are instructive.
Future Desk
Output per hour rose in sectors with heavy adoption of AI tools, according to figures described in this illustrative scenario. The same figures are consistent with several explanations that have nothing to do with AI.
2 min read
A model that answers is a reference tool. A model that acts is a colleague, a contractor or a liability. The difference between the two is smaller in code than it is in consequence.
3 min read
Districts in this illustrative scenario are replacing take-home essays with in-class writing and oral defense, while adopting AI tutors for practice. Teachers describe more work, not less.
2 min read
Coding was the first job that AI agents did well, because code is the rare kind of work that tells you immediately when you are wrong. The consequences for who programs, and how, are still arriving.
2 min read
Models that reason step by step get more right. They also produce confident, well-structured arguments for wrong answers. Reliability, it turns out, is not the same as capability.
2 min read