I’m working under a BERI grant to use forecasting to support work on x-risk, together with Ben Goldhaber. We’re currently experimenting with a novel format for effective group forecasting.
What does the workshop involve?
We’ll gather outside views, make guesstimates, factorize the question into sub-questions, turn our uncertainty into probabilities, and so forth!
Much of the workshop will happen via silent collaboration in a Google doc, using the structured format that we’re developing specifically for this.
Do I need prior experience with forecasting?
No! We want to learn how well this format works for people with a wide range of backgrounds.
What have past attendees said?
We’ve had about ~20 people try this format over the last weeks, with a wide range of experience from superforecasters and AI safety researchers to amateur forecasters and undergraduates.
People have generally found it effective and fun, though we have several things we want to improve and are looking for more feedback and volunteers to keep iterating.
When is it?
Friday May 17, 11am Pacific time/7pm UK time. If you’re interested in joining but can’t make this time, let us know your other availabilities here.
How do I join?
Sign-up via this form, and we’ll send you more details.
AI Forecasting online workshop
I’m looking for volunteers to join me for a 1.5h online AI forecasting workshop.
Why am I doing this?
I’m working under a BERI grant to use forecasting to support work on x-risk, together with Ben Goldhaber. We’re currently experimenting with a novel format for effective group forecasting.
What does the workshop involve?
We’ll gather outside views, make guesstimates, factorize the question into sub-questions, turn our uncertainty into probabilities, and so forth!
Much of the workshop will happen via silent collaboration in a Google doc, using the structured format that we’re developing specifically for this.
Do I need prior experience with forecasting?
No! We want to learn how well this format works for people with a wide range of backgrounds.
What have past attendees said?
We’ve had about ~20 people try this format over the last weeks, with a wide range of experience from superforecasters and AI safety researchers to amateur forecasters and undergraduates.
People have generally found it effective and fun, though we have several things we want to improve and are looking for more feedback and volunteers to keep iterating.
When is it?
Friday May 17, 11am Pacific time/7pm UK time. If you’re interested in joining but can’t make this time, let us know your other availabilities here.
How do I join?
Sign-up via this form, and we’ll send you more details.