NewsMatch makes it easy to find relevant context for questions — so when the story changes, so can your forecast. It uses a semantic similarity model to match Metaculus questions to news stories from across the globe, presenting you with a list of articles most likely to be relevant.
We previously released NewsMatch as an experimental feature. Now, after further testing and refinement, we've reintroduced it to help you update predictions, incorporate information from a wide variety of sources, and identify new lines of reasoning.
NewsMatch was created in collaboration with our friends at Improve The News, a non-profit news aggregator developed by MIT researchers that gives readers more control over their news intake.
More matches will roll out over time. For now, you can check out questions like:
The machine-learning model driving the article matching is always improving — you can accelerate its improvement by sharing your feedback and clicking the rate buttons on the right of NewsMatch stories.