Hide table of contents

Podcast Link

There was recently a lengthy thread on the EA Forum about the value of forecasting as a potential cause area, between Eli Lifland and myself (Ozzie Gooen). We thought it would be interesting to expand on this in a podcast episode.

Some Summary Points

  • Open Phil's expanded forecasting grant-making program has sparked debate about the value and impact of this area.
  • The definition and boundaries of "forecasting" in EA are unclear, leading to differing opinions on its importance versus other priorities.
  • AI could significantly change forecasting, and integrating AI into forecasting pipelines is a key consideration.
  • Improving "epistemic infrastructure" is important, but the best approaches are uncertain, ranging from forecasting tournaments to broader efforts.
  • More work is needed on judgmental forecasting of AI risk and other key questions, but the tractability and impact are debated.
  • The prioritization of forecasting and the resources it deserves remain complex, unresolved questions requiring further research and experimentation.

Some Mentioned Organizations and Projects

  • Open Philanthropy - A foundation that aims to do as much good as possible with its giving, including grants for forecasting research and programs.
  • Metaculus - A community forecasting platform that hosts predictions on a wide range of topics, including AI and other EA-relevant areas.
  • CSET-Foretell - A project by the Center for Security and Emerging Technology (CSET) that uses forecasting to inform policy decisions related to emerging technologies.
  • Manifold Markets - A prediction market platform that allows users to create and trade on forecasting questions.
  • The Good Judgment Project - A research project that studies the principles of good judgment and forecasting, and trains "Superforecasters" who demonstrate exceptional prediction skills.
  • The Effective Altruism Forum - An online community and discussion platform for topics related to effective altruism, including forecasting and AI safety.
  • The Good Food Institute - A nonprofit organization that promotes the development of alternative proteins, including plant-based and cultivated meat, to improve food sustainability and animal welfare.
  • Fatebook - An easy way to make and track predictions.
  • Future Search - A startup that’s working on AI-assisted forecasting.

We don’t have an edited transcript. We do have an autogenerated transcript and widget from Descript, but note that it has a lot of errors in it, and it includes a lot of filler words and phrases. 




Sorted by Click to highlight new comments since: Today at 4:19 AM

Thanks Ozzie for chatting! A few notes reflecting on places I think my arguments in the conversation were weak:

  1. It's unclear what short timelines would mean for AI-specific forecasting. If AI timelines are short it means you shouldn’t forecast non-AI things much, but it’s unclear what it means about forecasting AI stuff. There’s less time for effects to compound but you have more info and proximity to the most important decisions. It does discount non-AI forecasting a lot though, and some flavors of AI forecasting.
  2. I also feel weird about the comparison I made between forecasting and waiting for things to happen in the world. There might be something to it, but I think it is valuable to force yourself to think deeply about what will happen, to help form better models of the world, in order to better interpret new events as they happen.

I'm interested to listen but I can't find the podcast on Spotify- am I missing something?