Metaculus is a forecasting technology platform that optimally aggregates quantitative predictions of future events. Research shows that with the right incentives and feedback, groups of people can make remarkably accurate predictions of the probability of future events. Our mission is to build epistemic infrastructure that enables the global community to model, understand, predict, and navigate the world’s most important and complex challenges.
External link for Metaculus
Box 7324
Santa Cruz, CA 95060, US
Metaculus reposted this
Proud to share I've placed 3rd of 120+ competitors in Metaculus's Q3 Market Pulse Tournament. The questions covered VIX, the 10-year Treasury yield, ICE BofA US High Yield OAS, relative moves in equities and futures, and company quarterly performance. Anchoring on base rates and forming realistic ranges helped, and I used current conditions to fine-tune my forecasts. Excited to iterate my process for future tournaments!
Metaculus reposted this
Agtech & Biotech Leader | Bridging Deep-Tech R&D and Commercial Strategy | Early-Stage Venture & Innovation Specialist
Last week I spent 3 days thinking about the future with an incredible group at the launch of the Horizon Scanning Working Group, hosted by the brilliant team at Renaissance Philanthropy. The goal was to bring together a diverse group of experts to stress-test and refine some of ARIA's opportunity spaces. The process, facilitated by the fantastic team at Metaculus, involved: * Mapping the bottlenecks and breakthroughs for transformative technologies. * Identifying the critical variables that could block or enable future scenarios. * Using forecasting tools to challenge our assumptions and inform future interventions. I was absolutely thrilled to pitch Programmable Plants as a potential focus area, and even more so that the group selected it as one of the five topics for a 9-month deep dive. The study group that has formed around this topic is simply stellar, and I could not be happier about the mix of expertise in it - a huge thank you to Paul Freemont, Liz Specht, Ph.D., Samuel Arbesman, Vincent Hanlon, and Alice Pettitt for the thoughtful, constructive, and creative engagement so far. I cannot imagine what you will come up with if this is what you can accomplish in just a couple of days! My key takeaway on the event: We will essentially be road-testing an update to classic DARPA Study Groups like ISAT, reframed to work for the ARIA model - shifting from narrow programmes to broad Opportunity Spaces, and from static reports to dynamic, living documents. It’s a powerful tool to periodically revisit our beliefs, update them as new information emerges, and seamlessly share them with the wider community. Feeling incredibly grateful and optimistic. A final thank you to Lauren Gilbert, Eirini Malliaraki, Andrew Black, and Joshua Elliott at RenPhil, and Molly Hickman and Deger Turan from Metaculus for architecting such a fun, productive, and inspiring retreat. I cannot wait to see what emerges from this work over the coming months. #HorizonScanning #DeepTech #Foresight #Innovation #ProgrammablePlants #Biotechnology #ARIA
Metaculus reposted this
Last week, we kicked off the Renaissance Philanthropy Horizon Scanning Study Group, powered by ARIA and Metaculus. We convened 30 scientists from around the world to predict the future - mapping out future technologies, forecasting on future hinge points and thinking about possible high-return program areas. We had an incredible group of scientists - spanning many disciplines, career stages, and industries - and were lucky enough to be joined by many ARIA program directors and team members. We also had some truly epic karaoke - Dr Jacqueline Campbell described the event as "the best day of her life". Extremely excited to continue our work over the next nine months! Andrew Black Eirini Malliaraki Joseph T. Meyerowitz Ivan Jayapurna Vincent Hanlon Dr Jacqueline Campbell Niko McCarty Jean Rintoul, PhD Paul Freemont Alice Pettitt Edith-Clare Hall John Dabiri Manoj Saxena Samuel Arbesman Nicole Wheeler Peter Woods Koziol Anders Sandberg Fabrizio Ticchiarelli-Marjot Madeleine Luck, PhD CF Dr. Jane Roskams Rose M. Mutiso, Ph.D. Joshua Elliott Jassi Pannu Adam Marblestone Vehbi Deger Turan Molly Hickman Maya Gavin Liz Specht, Ph.D. Lawrence Lundy-Bryan Laurie Smith George Horner Brian Wang
Metaculus reposted this
What futures could, and should, we create with advanced AI? Today we are announcing two possible paths – Tool AI and d/acc – with contributors like Vitalik Buterin, Adam Marblestone and Anthony Aguirre. • In the Tool AI scenario, AI is highly competent, but narrow and with limited agency. Humans remain in control, and use AI to design vaccines, summarize research literature, and personalize learning for students – a high-speed, low-risk amplifier of human capacity. But questions remain: are we missing out on important progress by choosing tool AI over AGI? And where do we draw the line between "tool" and "agent"? • d/acc stands for decentralized, democratic, differential, and defensive acceleration. The idea: we can, and should, be selective about which technologies we accelerate, prioritizing those that strengthen defense, distribute power, and promote cooperation. Examples include decentralized AI, distributed financial protocols, open science, and network states. But uncertainties remain: can distributed actors coordinate fast enough during crises? And how do we determine when decentralization enhances resilience rather than adding inefficiency? With these scenarios, we want to open up the meme space for what futures with AI are possible – showing that multiple, actionable paths exist – so we can make more informed decisions about which ones we should steer towards. Explore the scenarios: • Tool AI: https://lnkd.in/dRAkpQKv • d/acc: https://lnkd.in/dw-R_vXt With Metaculus you can make predictions on key aspects of the scenarios – we award a $5,000 prize pool to the 8 commenters making the most meaningful contributions: https://lnkd.in/eWs2cQTz Contributors to the scenarios: Tool AI • Adam Marblestone – Convergent Research • Benjamin Reinhardt – Speculative Technologies • Joel Leibo – DeepMind • Anthony Aguirre – Metaculus & Future of Life Institute • Anton Korinek – University of Virginia • Brad Love – Los Alamos National Laboratory • Saffron Huang – Collective Intelligence Project • Konrad Kording – University of Pennsylvania • Jeremy Barton – Nano Dynamics Institute • Kristian Rönn – Lucid Computing • Cecilia Elena Tilli – Cooperative AI Foundation • Owen Cotton-Barratt – Researcher • Rif A. Saurous – Google d/acc • Vitalik Buterin – Ethereum • Kevin O. – Gitcoin • E. Glen Weyl – Microsoft Research, RadicalXchange • Naomi Brockwell – NBTV / Privacy Advocate • Andrew Trask – OpenMined • Allison Duettmann – Foresight Institute • Soham Sankaran – PopVax • Christine Peterson – Foresight Institute • Marcin Jakubowski – Open Source Ecology • Emilia Javorsky MD, MPH – Future of Life Institute • Molly Mackinlay – Protocol Labs • Vehbi Deger Turan – Metaculus • Lou de Kerhuelvez – Nodes
Metaculus reposted this
What futures could, and should, we create with advanced AI? Today we are announcing two possible paths – Tool AI and d/acc – with contributors like Vitalik Buterin, Adam Marblestone and Anthony Aguirre. • In the Tool AI scenario, AI is highly competent, but narrow and with limited agency. Humans remain in control, and use AI to design vaccines, summarize research literature, and personalize learning for students – a high-speed, low-risk amplifier of human capacity. But questions remain: are we missing out on important progress by choosing tool AI over AGI? And where do we draw the line between "tool" and "agent"? • d/acc stands for decentralized, democratic, differential, and defensive acceleration. The idea: we can, and should, be selective about which technologies we accelerate, prioritizing those that strengthen defense, distribute power, and promote cooperation. Examples include decentralized AI, distributed financial protocols, open science, and network states. But uncertainties remain: can distributed actors coordinate fast enough during crises? And how do we determine when decentralization enhances resilience rather than adding inefficiency? With these scenarios, we want to open up the meme space for what futures with AI are possible – showing that multiple, actionable paths exist – so we can make more informed decisions about which ones we should steer towards. Explore the scenarios: • Tool AI: https://lnkd.in/dRAkpQKv • d/acc: https://lnkd.in/dw-R_vXt With Metaculus you can make predictions on key aspects of the scenarios – we award a $5,000 prize pool to the 8 commenters making the most meaningful contributions: https://lnkd.in/eWs2cQTz Contributors to the scenarios: Tool AI • Adam Marblestone – Convergent Research • Benjamin Reinhardt – Speculative Technologies • Joel Leibo – DeepMind • Anthony Aguirre – Metaculus & Future of Life Institute • Anton Korinek – University of Virginia • Brad Love – Los Alamos National Laboratory • Saffron Huang – Collective Intelligence Project • Konrad Kording – University of Pennsylvania • Jeremy Barton – Nano Dynamics Institute • Kristian Rönn – Lucid Computing • Cecilia Elena Tilli – Cooperative AI Foundation • Owen Cotton-Barratt – Researcher • Rif A. Saurous – Google d/acc • Vitalik Buterin – Ethereum • Kevin O. – Gitcoin • E. Glen Weyl – Microsoft Research, RadicalXchange • Naomi Brockwell – NBTV / Privacy Advocate • Andrew Trask – OpenMined • Allison Duettmann – Foresight Institute • Soham Sankaran – PopVax • Christine Peterson – Foresight Institute • Marcin Jakubowski – Open Source Ecology • Emilia Javorsky MD, MPH – Future of Life Institute • Molly Mackinlay – Protocol Labs • Vehbi Deger Turan – Metaculus • Lou de Kerhuelvez – Nodes
Metaculus reposted this
🥈 Excited to have placed 2nd in the latest Metaculus AI Forecasting tournament! It has been fascinating to see how the competition has changed over the last few months, as LLM thinking models like o3, and search tools, have been released. One lesson is that model capability appears to be far more important than prompt engineering, with the Metaculus o3 bot with the default prompt nearly winning the whole thing! Interested to see how well GPT-5 and future frontier models perform. Join me for the next tournament starting soon! 🚀 https://lnkd.in/dsdaHNWM #AI #ArtificialIntelligence #LLM #Forecasting
🏆 Excited to have won 1st place in the Q4 Metaculus AI Forecasting tournament! https://lnkd.in/eMKJd9gM I've been really interested in forecasting and prediction markets, as well as the latest LLM advancements, and this tournament has been a great opportunity to deep dive into capabilities and limitations. My favorite part of the tournament has been experimenting with the various different parts of an agentic workflow and seeing important it is that they all fit together. First you need great context engineering (I used Perplexity and AskNews) to make sure the AI is using the most up-to-date information, then you need complex prompt engineering to give instructions, and only then do you need the right underlying reasoning engine provided by the LLMs (GPT-4o, Claude and Gemini) to put it all together. The most fascinating parts? Seeing how well one-shot predictions from the bot templates performed, and how small tweaks in how you frame instructions or provide context can lead to vastly different results. If you're interested in prediction markets or AI, I encourage you to participate in the next 2 quarters of the tournament, starting next week! It's hands-on experience you can't get from just reading about these technologies. Big thanks to the Metaculus team for organizing this - it was both educational and a lot of fun! https://lnkd.in/eQv6rVuu #AI #Forecasting #artificialintelligence
Metaculus reposted this
Just found out I placed 3rd out of 293 forecasters in the Metaculus Biosecurity Tournament that ran from 2022 to 2025 on questions about COVID-19, global biosecurity, public health funding, and more! Really enjoyed forecasting this one! https://lnkd.in/eNDmev6b
"Technology now allows hiring that emphasizes ability and potential, over credentials and traditional paths." — Blake Cecil Couldn't agree more. In our 2024 forecasting tournament with Bridgewater Associates, we saw this principle in action when a Grinnell College student placed first and Northwestern topped the school rankings, outperforming traditional elite institutions. Forecasting ability is a powerful indicator of analytical talent and reasoning skills, and we're excited to again collaborate with Bridgewater to identify exceptional talent. The contest begins Monday, February 3rd. Warmup questions and registration details are here: https://lnkd.in/eqF3emKs
As I explored in my piece for Fast Company, meritocratic recruitment is increasingly reshaping the hiring process at elite companies that need the best talent. Technology now allows hiring that emphasizes ability and potential, over credentials and traditional paths. This approach levels the playing field, enabling companies to access diverse talent pools that were previously overlooked. Organizations like Bridgewater Associates that embrace this shift can cultivate the most talented, innovative teams.
Metaculus reposted this
At Bridgewater Associates, we believe great ideas can come from anyone, anywhere. That’s why we’ve partnered with Metaculus to host a global forecasting competition. With $25,000 in prizes on the line, this is your opportunity to showcase your forecasting abilities and explore exciting career possibilities with Bridgewater. Do you have what it takes to predict the future and ready to show what you’ve got? Join the challenge here: https://lnkd.in/eqF3emKs
Great to see responses like this to our Day One policy memo for the Federation of American Scientists! Crowd forecasting quantifies uncertainty, predicts intervention impacts, and supports decision-makers in evaluating risks and tradeoffs. Read our proposal Collaborative Intelligence: Harnessing Crowd Forecasting for National Security by Vehbi Deger Turan, Molly Hickman, and Leonard Barrett here: https://lnkd.in/ef9A4cEg
I’m very excited to see the Federation of American Sciences call out the RAND Forecasting Initiative (RFI) in Collaborative Intelligence: Harnessing Crowd Forecasting for National Security ( https://lnkd.in/ePrytZQa ), part of its Day One Project for the incoming Trump administration. The author’s advocate for the development within an FFRDC of a next-generation crowd-forecasting program that balances academic rigor with policy relevance and recommend RFI as the potential home for this effort. RFI is already (1) supporting government and foundation sponsored RAND research; (2) managing forecasting working groups that bring together participants from US and allied governments, forecasters, and RAND subject matter experts to decompose strategic issues into policy-relevant forecast questions using a variety of methods and AI tools; and (3) developing education and training in coordination with the RAND Pardee Graduate School and others to help analysts and policymakers understand how to integrate collaborative intelligence into their processes and products. Learn more about RFI here: https://lnkd.in/eFSYEU7s