3 Comments

As always, you included news that I had not seen before. Thanks.

Expand full comment
Feb 3, 2022·edited Feb 3, 2022Liked by Nuño Sempere

You (and until recently, I) missed this big news, an Air Force SBIR for a combined forecasting/news aggregation system in similar to Good Judgment. Inc.'s FutureFirst, but leveraging NLP/AI to greatly increase throughput and coverage. https://www.sbir.gov/node/2101893. Phase III is focused upon commercialization. This would be a great leap forward in the forecasting world, but also could be seen as a threat to GJI's FutureFirst and in general to the use of superforecasters.

We have shown in " What do forecasting rationales reveal about thinking patterns of top geopolitical forecasters?" https://www.sciencedirect.com/science/article/abs/pii/S0169207021001473, that NLP/random forest applied to forecaster rationales can turn a relatively ordinary set of human forecasters such as Amazon Prime workers into a prediction powerhouse similar to superforecasters. The AF SBIR should build upon our research.

As a superforecaster myself, I'm not entirely thrilled by this progress, but on the other hand, it promises to empower "regular" folks, and I like that more.

Expand full comment
author

Nice, thanks for the reference!

Expand full comment