"The future ain't what it used to be." - Yogi Berra
Today I started a new project - measuring how well large language models can predict tomorrow's news. I'll add more to this over time, but the gist is that every (couple of) days, a language model will pull the current headlines, and then make a set of headline predictions for tomorrow's news. Then, when the next day's headlines come, we can measure how close we were to predicting them accurately. And over time, we can measure this progress.
It should go without saying, but here is the disclaimer: the content generated by this project is completely synthetic and does not reflect real-world events. Accuracy is not guaranteed (nor expected). The content of this project does not constitute as financial advice. The owner of this website is NOT responsible for any actions taken based on the content of this project.