cassi-ai.bsky.social
@cassi-ai.bsky.social
H/T & thank you to the quietly and inspiringly brilliant John Hetherington and Adrian Snell for all the work leading this effort behind the scenes.
a young boy wearing a hat that says respect is sitting in a crowd of people
ALT: a young boy wearing a hat that says respect is sitting in a crowd of people
media.tenor.com
January 8, 2026 at 10:21 PM
Moreover, prediction is about more than forecasting.

'Everything is Prediction'.

As we described over on the blog:
January 8, 2026 at 10:21 PM
Why does it matter: it's highly unlikely your organisation has access to a team of proven superforecasters, and even if you did, forecasting takes time.

Cassi's forecasts are with you at the touch of a button.
January 8, 2026 at 10:21 PM
Why did we excel? Cassi focuses on real-world performance - accuracy in prediction in areas where there are no good odds - as part of wider client work.

Ironically maybe, we won because we are not 'bench-maxing' by tailoring our model to the scoring systems of other forecasting competitions. ✅
January 8, 2026 at 10:21 PM
For questions where where you can start estimating (eg) chances of country X being invaded by country Y, w/odds from @metaculus.bsky.social,
@polymarket.bsky.social, @kalshiofficial.bsky.social) human superforecasters still have the edge... ...but in AI forecasting we are second only to XAI.
January 8, 2026 at 10:21 PM
On "Dataset" questions, Cassi was within 0.006 Brier* score of the median Superforecaster—practically a dead heat.

When you can't leverage the "wisdom of crowds" to baserate your odds, the gap between AI and elite humans almost vanishes🎯
January 8, 2026 at 10:21 PM
Cassi is the best in the world: #1 for questions where there are no good odds out there already.

Which is most of the questions you, your business and your organisation care about.
January 8, 2026 at 10:21 PM