Anthropic reportedly lifted Claude's wiki score from 13% to 48% with a method called 'dreaming'. How the score was measured is unclear, so the numbers need care.
Anthropic, the company behind the AI assistant Claude, reportedly raised a Claude score on a wiki task from 13% to 48% using a method it calls 'dreaming'. That is nearly a fourfold jump. It comes from a FourWeekMBA report, and the details behind it are thin so far.
The reported numbers are simple. The score started at 13% and reached 48% after the method called 'dreaming' was used. The material does not say what 'dreaming' actually does, and it does not explain what the wiki score tests.
Anthropic makes Claude, an AI assistant that people use for writing, research and coding. Many businesses and individuals already lean on tools like it for everyday work. That is why a big reported gain gets attention, even before the full story is out.
Treat the 13% and 48% figures as a headline, not a finished fact. A score only means something when you know how it was measured. If AI assistants keep getting better at tasks like this, the tools you use for work may become more useful. Until the details are clear, check any claim before you rely on it.
A jump from 13% to 48% sounds big, but without knowing how it was measured, treat it as a claim to verify, not a promise.
World models, AI that learns physics from video, are being built into home robots and driverless cars, and your own videos may help train them.
Read more: AI that learns how the physical world works is coming to robots and carsA new Alibaba-led test of real coding requests found the best AI agent fully finished just 26.4 percent of tasks, so human review still matters.
Read more: AI coding helpers fully finish only about 1 in 4 real tasksA University of Chicago law professor asked Claude to fix page numbers. It deleted his syllabus and a year of notes instead. Dropbox saved him, not the AI.
Read more: An AI assistant deleted a law professor's year of lecture notes
