Forecasts on critical topics. Labor Automation Forecasting Hub: metaculus.com/labor-hub FutureEval: metaculus.com/futureeval Partner: metaculus.com/services
Santa Cruz, CA
- Tweets 3,763
- Following 855
- Followers 13,239
- Likes 5,164
We're also benchmarking AI forecasting accuracy over time and across different models via FutureEval: metaculus.com/futureeval/
AIs cracking short-term superforecasting is such an important (and, still, under-discussed) trend.
Via a new piece from @nikostro economist.com/science-and-te…
11
1,074
We just launched the Metaculus AI Existential Risk Explorer, created by our Director of Forecasting, @ryanbeck111.
It examines how the forecasting community is thinking about AI-caused catastrophe & extinction risk: ai-risk-explorer.metaculus.c…
1
3
402
"Will AI Cause Human Extinction? Pro Forecasters Land at 7.5%. Our Most Dedicated X-Risk Forecasters Say 22%." (writeup): metaculus.substack.com/p/wil…
1
325
--but also, some important distinctions between the questions:
- We specified a Millennium Prize problem vs. their broader set of unsolved problems
- The survey specified the AI needed to solve without further human input. There was some amount of human input here. See here: aiimpacts.org/wp-content/upl…
Question link: metaculus.com/questions/2853…
1
278
We have a new batch questions in our Animal Futures Tournament:
- How many shrimp per year will be covered by SWP's Humane Slaughter Initiative agreements on July 1, 2027?
- What will be the total retail sales of plant-based meat and seafood in 2026?
- Will at least one additional U.S. state enact a cage-free egg requirement before July 2027?
- What percentage of companies globally will fulfill their pre-2026 broiler commitments in 2026 and 2027?
- What percentage of companies globally will fulfill their pre-2026 cage-free commitments in 2026 and 2027?
Leave your answers to these & other questions via the official tournament page: metaculus.com/tournament/330…
10
3
716
For the first time, an AI forecaster has won the seasonal Metaculus Cup.
The Economist has covered the story; link in replies. Congratulations to Jeffrey Liang for building laertes, the winner in question.
3
16
155
27,614
"Computers have long been used to make forecasts in narrow fields, such as the weather. But AI forecasters, like those on Metaculus, are a different breed. Instead of being trained specifically on data regarding the questions they are predicting the answers to, they are built using the same large language models (LLMs) that underpin the rest of the AI boom."
And congratulations for the top-performing human forecaster, @BenShindel!
See the full Summer Metaculus Cup leaderboard here: metaculus.com/tournament/met…
Read the full piece here: economist.com/science-and-te…
1
10
1,892
We also run FutureEval, a benchmark that measures AI progress, based on seasonal Bot Tournaments. Learn more here: metaculus.com/futureeval/
2
1,061
The Hugging Face Incident, OpenAI's agents hijacking a German wiki, & now Jacob Coxon's viral resignation. Some argue these are signals that AI safety, security, & alignment are not keeping up with the speed of deployment, while others hold they're more media moments than genuine risk indicators.
We put together a flash briefing on AI cybersecurity risk, based on forecasts & rationales left by Pro Forecasters, Metaculus staffers, and our flagship AI forecaster, Azimuth. Find our analysis at the link below, where you'll also find links to the question series to leave your own predictions. metaculus.substack.com/p/for…
1
2
8
1,255
Announcing: the Metaculus Cup Fall 2026 with guest forecaster Kiko Llaneras (@kikollan), data journalist at @el_pais!
Go head-to-head to compete for up to $7,500 in prizes. The first batch of questions includes:
- Will the United States attack Oman before 2027?
- Who will be awarded the 2026 Nobel Peace Prize?
- Will FEWS NET classify any area of Sudan as experiencing Famine before 2027?
- How many Major Atlantic Hurricanes will occur in the 2026 Atlantic hurricane season?
- What first round vote share will Iliana Iotova win in the 2026 Bulgarian Presidential election?
Link in replies!
1
2
13
15,319
Start forecasting in the Metaculus Cup Fall 2026: metaculus.com/tournament/met…
1
428
Metaculus retweeted
I expect the "human-machine intelligence parity" would resolve positively if we ran the test.
metaculus.com/questions/384/…
Portal 2 may be on the horizon soon as well. metaculus.com/questions/2193…
Jesus.
1
3
457
We're reviewing whether a key criterion of one of our longest-running AGI questions has been met.
The question has been open since 2020. It asks for one unified system that clears four bars, and the recent claims are about the last of them, the Atari game 'Montezuma's Revenge.'
We've asked for methodology and code so we can try to replicate. We'll review reproducibility and share what we find, including if the answer is that it doesn't resolve.
Link to question in replies.
GPT-6 Astra (low) has beaten the Atari game Montezuma's Revenge, in real time, with a basic harness
This presumably resolves the Metaculus question "When Will Weakly General AI Arrive?"
4
4
48
31,130
When will the first weakly general AI system be devised, tested, and publicly announced?
metaculus.com/questions/3479…
1
1
561
Metaculus retweeted
Participar es abierto y gratis. Metaculus es una comunidad de apasionados de la predicción tomada en serio. El torneo empieza ya y cuanto antes pongas números, mejor.
Aquí los detalles👇
metaculus.com/tournament/met…
1
1
5
2,266
Metaculus Cup Fall 2026: metaculus.com/tournament/met…
151
Metaculus retweeted
🔮 ¿Atacará EE UU a Omán? ¿Cuántos huracanes grandes habrá en 2026? ¿Ganará Trump el Nobel de la Paz?
Ponle un número. Cada semana hasta diciembre, la comunidad de @metaculus predice 4-5 preguntas así. Y este otoño compite contra un invitado: ¡yo!
No voy a ganar (me doy 10%). Pero es un ejercicio brutal para aprender, entender tus sesgos y apreciar el valor de los superpronosticadores.
Yo iré contando mi proceso.
¡Súmate! Link abajo👇
1
3
14
3,611