If you're:
• new to forecasting
• looking to skill up with fast feedback
• or, a parent of young children and feel like you just 'get' #Bluey on a deeper level than other parents

—then start forecasting in the new Q2 Quarterly Cup today:
metaculus.com/tournament/qua…

On June 25th, 2522, NASA predicts a leisurely long solar eclipse.

BUT — 500 years is a long time: What if we engineer the solar system in ways that invalidate the assumptions of the prediction?

So, total solar eclipse on June 25, 2522? 👍 or 👎?
metaculus.com/questions/111/…

Register now for the 𝗕𝗿𝗶𝗱𝗴𝗲𝘄𝗮𝘁𝗲𝗿 𝘅 𝗠𝗲𝘁𝗮𝗰𝘂𝗹𝘂𝘀 𝗙𝗼𝗿𝗲𝗰𝗮𝘀𝘁𝗶𝗻𝗴 𝗖𝗼𝗻𝘁𝗲𝘀𝘁:
metaculus.com/bridgewater/

+$25,000 prize pool
+Explore career opportunities with Bridgewater, a premier asset management firm
+Undergrad & Open leaderboards

Begins April 16th.

What will be the outcome of Donald Trump's New York 'hush money' criminal trial?

metaculus.com/questions/2206…

Are you a Bryan booster? A Caplan critic?
Come predict whether he wins his current slate of bets: metaculus.com/project/caplan/

Timeline of AI forecasts

What to expect in AI capabilities, potential harms, and societal impact, according to the median prediction of forecasters on @Metaculus and Manifold

Explore all >250 predictions on theaidigest.org/timeline

This US presidential election season:
Scroll articles, read polls, tune into podcasts—and bookmark Metaculus's interactive map aggregating live, state-by-state election forecasts:
metaculus.com/experiments/el…

How capable will top AI models be in 2025?

Forecast:
• LLM agents' autonomous replication & adaptation (ARA) abilities
• model performance on benchmarks like GPQA & GAIA

— in AI Benchmarks, a collaboration with the AI Safety Student Team at Harvard: metaculus.com/project/3054/

How many people will work in these AI-automatable jobs (e.g., interpreters, transcriptionists, file clerks) in the US in May 2025?

May 2019: 477,230
May 2022: 441,140

May 2025 (forecast from Metaculus): ~333,000

metaculus.com/questions/2142…

Pre-register your forecasts for the next generation of AI capabilities at metaculus.com/project/3054/ (by Romeo Dean, drawing from my list of benchmarks at foxy-scout.com/a-template-for…)

Predicting the progress of scientific funding proposals (especially if they are high-risk/high-payoff) is a crucial issue that I was very happy to be able to contribute to in forecasting on this @ScientistsOrg project with other forecasters from @Metaculus!

'...lawmakers and executives have begun to converge on an ostensibly straightforward plan: test the AI models to see if they are indeed dangerous.'
—from TIME's Nobody Knows How to Safety-Test AI
time.com/6958868/artifi…

Which labs will get evals by 2026?metaculus.com/questions/2122…

Economist Bryan Caplan has won all his public bets—a nearly two-decade winning streak. Now you can predict the outcomes of his bets on AI-supercharged growth, climate change, civil wars, and more in our new Bryan Caplan Collection.

Get started:
metaculus.com/project/caplan/

Science funding agencies are biased against risk, making transformative research difficult to fund. With Metaculus, we piloted an alternate, expected utility-based approach to grant review that could give high-risk, high-reward research a fairer shot.

📚fas.org/publication/pr…

Russia Elections Predictions 🪩

You can now predict candidate percentage shares in 2024 Russian Presidential elections at Metaculus.

metaculus.com/questions/2188…

Resolved!
metaculus.com/questions/2091…

Next up:
𝗪𝗶𝗹𝗹 𝗦𝘁𝗮𝗿𝘀𝗵𝗶𝗽 𝗿𝗲𝗮𝗰𝗵 𝗼𝗿𝗯𝗶𝘁 𝘁𝗵𝗶𝘀 𝘆𝗲𝗮𝗿?

Forecasters updated probabilities this week, raising the aggregate of 786 predictions by 7% to reach 82%:
metaculus.com/questions/2077…