AI forecasters are approaching top human performance

We don't have to be naive

Trump's main character syndrome

"The historical parallel should be worrying"

Stay stubborn

Sometimes you need allies

AI comes at you fast

Fish, family, freedom

I think I can say with little fear of contradiction that I...

Show your support for Hawaiian Humane Society by voting for Juniper in Love Drives Us – Hawaiian Humane Society Alumni Van Wrap Contest

We opened our AI forecasting benchmark to external submissions. Here’s what happened.

Minnesota No ICE

"The future's not magic—it's probability!"

A personal note

Midterms are a referendum on the party in power

"In a few months time it’s my belief that it will no longer be clear that systems like Mantic are below the performance of the human pros"

Demagogues rise when the social order breaks down

We’ve just released an updated version of ForecastBench, our LLM forecasting benchmark. Here’s what the new results reveal about the accuracy of state-of-the-art models.

Violence multiplies violence

"In the next year to eight years, the extrapolated trend would suggest that AIs will catch up to superforecasters"

"In the next year to eight years, the extrapolated trend would suggest that AIs will catch up to superforecasters"

The pauses are in response to President Donald Trump's decision to end a trade agreement that for years allowed low-value packages to avoid duties.

We taught it language, now it knows how to curse

I never realized how much I love walking

All my forecasts are hand produced by humans