EA Forum Podcast (All audio)

Crossposted from my Substack.

Rational choice theory is commonly thought of as being about what to do in light of our beliefs and preferences. But our beliefs and preferences come from somewhere. I would say that we believe and prefer things for reasons. My evidence gives me reason to believe I am presently in an airport. Any reason I have to think a coin will land heads applies to tails, too, so I should have the same degree of belief in each. The suffering I expect to be averted by donating to charity gives me a reason to donate.

I think reflecting more closely on how our reasons enter into our decision-making can open up new directions for the theory of decision-making. It is particularly relevant, perhaps, for bounded agents who cannot generally act on the basis of expected utility calculations, but can still reflect on and [...]

---

Outline:

(01:58) Preliminaries

(03:14) Reasons for belief

(08:24) Explicitly reasons-based choice

(10:21) Cluelessness

(18:05) Acknowledgements

The original text contained 2 footnotes which were omitted from this narration.

---

First published:

Source:

Narrated by TYPE III AUDIO.

Over the years, I have learned many things that are rarely taught about doing cost-benefit or welfare analysis. Here are a few things that I often end up repeating when I mentor individuals or teams working on these kinds of projects:

A Point Estimate is Always Wrong

For any purpose other than an example calculation, never use a point estimate. Always do all math in terms of confidence intervals. All inputs should be ranges or probability distributions, and all outputs should be presented as confidence intervals.

Do not start with a point estimate and add the uncertainty later. From day one, do everything in ranges. Think in terms of foggy clouds of uncertainty. Imagine yourself shrinking the range of uncertainty as you gather more data.

This Google Sheets Template allows you to easily set up Monte Carlo estimations that turn probabilistic inputs into confidence-interval outputs.

Use Google Sheets

I [...]

---

Outline:

(00:20) A Point Estimate is Always Wrong

(01:04) Use Google Sheets

(04:19) One Individual, One Event

(05:53) Minimum Viable Product and Iterate

(07:27) Deliver What the Customer Needs

(10:05) Ruthlessly Simplify Literature

(11:11) Do Longform NPV

(14:17) Breadth Beats Rigor

(14:51) Use Multiple Specifications

(16:48) Become or Hire a Journalist

(20:55) Time Costs Usually Dominate

(21:56) Analyses are Worthless, but Analyzing is Everything.

(23:47) Ask for Help

---

First published:

Source:

Narrated by TYPE III AUDIO.

Audio note: this article contains 95 uses of latex notation, so the narration may be difficult to follow. There's a link to the original text in the episode description.

This is one of two posts I’m putting up today on how little economic theory alone has to say about the effects full automation would have on familiar economic variables.
The other is “The ambiguous effect of full automation on GDP growth”.

Introduction

Let full automation mean a state of affairs in which AI and robotics can single-handedly produce everything we like to consume, and which can produce any unit of output using inputs (e.g. electricity) that currently cost less than what humans currently earn just for the labor contribution to producing the same unit. Many people think that, though partial automation has greatly increased wages historically, full automation would inevitably lower wages. Some people have the [...]

---

Outline:

(00:31) Introduction

(02:25) Homogeneous factors

(04:33) Constant returns to scale

(07:28) Increasing or decreasing returns to scale

(08:11) Summary

(10:23) Heterogeneous factors

(10:27) Heterogeneous labor

(12:13) Heterogeneous capital

The original text contained 8 footnotes which were omitted from this narration.

---

First published:

Source:

Narrated by TYPE III AUDIO.

Summary

Open Philanthropy is launching a request for proposals to improve AI capability evaluations. We're looking to fund work on more demanding GCR-relevant benchmarks, better evaluation science, and improving third-party model access and infrastructure.

More details:

• Grant range: $200K to $5M per grant
• Timeline: Open until April 1, 2025
• Eligibility: Anyone can apply, including those in academia, nonprofits, industry, or working independently.
• Full RFP and application form: https://www.openphilanthropy.org/request-for-proposals-improving-capability-evaluations 

Below, we explain what we’re looking for, and why we think this work matters.

Why capability evaluations matter

The ability to accurately evaluate AI capabilities is becoming increasingly important, for three main reasons:

1. Evaluations are key inputs to AI governance

Many current governance proposals rely heavily on knowing what AI systems can and cannot do. “If-then commitments” are one prominent example — companies agree to take specific actions (like pausing training) if their systems display [...]

---

Outline:

(00:08) Summary

(01:02) Why capability evaluations matter

(02:36) Three current problems with capability evaluations

(03:43) What were looking to fund

(03:52) GCR-relevant capability evaluations for AI agents

(04:26) Improving the science of capabilities development and evaluations

(05:04) Improving third-party model access and evals infrastructure

(05:53) How to engage

---

First published:

Source:

Narrated by TYPE III AUDIO.

This is one of two posts I’m putting up today on how little economic theory alone has to say about the effects full automation would have on familiar economic variables.
The other is “The ambiguous effect of full automation on wages”.

Introduction

A lot of us are wondering what impact AI will have on global GDP growth (or would have if fully aligned and minimally regulated, in a world not destroyed by conflict). People have occasionally asked for my opinion on the question ever since I first wrote a literature review on the question five years ago—in fact before that, which is one of the reasons I wrote it! My answer has changed over time in many ways, but the outline remains similar.

1. It seems much more likely than not to me that advanced enough AI would eventually result in GWP growth above the fastest rates of sustained [...]

---

Outline:

(00:20) Introduction

(02:10) Example

(05:23) Discussion

(10:59) More on the motivation

The original text contained 8 footnotes which were omitted from this narration.

---

First published:

Source:

Narrated by TYPE III AUDIO.

The Mental Health Funding Circle just wrapped up its Autumn 2024 round. We’re delighted to support organisations doing highly impactful work in the space of mental health and wellbeing. This round we supported six organisations for a total of $560,000. We received a large number of excellent applicants, and were not able to fund all of the organisations which applied – many of which we consider to be doing exceptional work.

Our next round's applications are open! Applications will close March 15th - you can find the application on our website. Whilst Will Troy will continue to chair the MHFC, from the next round onwards Peter Brietbart and Lily Yu of the Bloom Wellbeing Fund will co-chair, bringing more capacity and experience into the project.

About the MHFC: Global mental health continues to be highly neglected despite contributing more to cumulative suffering than many more highly-prioritized [...]

---

First published:

Source:

Narrated by TYPE III AUDIO.

Crosspost form an article on my tooooootally amaaaaaaaazing blog that you should definitely check out!

I’ve spent a few recent articles arguing that PEPFAR is really valuable and attempts to cut it would be disastrous. PEPFAR is an international program that combats HIV—it's saved many millions of lives. Probably the most common response has been roughly the following: “sure PEPFAR saves a few kids from diseases, but it doesn’t address the root cause of the problem which is [gestures vaguely at whatever bit of one's ideology explains why bad things happen].”

Now, when it comes to PEPFAR, this is simply factually wrong. It is addressing the root of the problem. You can see this if you look at a graph of HIV deaths over time:

People also make the same objection to effective altruism. They’ll say: “effective altruism only is a palliative solution, that doesn’t address the [...]

The original text contained 3 images which were described by AI.

---

First published:

Source:

Narrated by TYPE III AUDIO.

---

Images from the article:

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

Crosspost form an article on my tooooootally amaaaaaaaazing blog that you should definitely check out!

I’ve spent a few recent articles arguing that PEPFAR is really valuable and attempts to cut it would be disastrous. PEPFAR is an international program that combats HIV—it's saved many millions of lives. Probably the most common response has been roughly the following: “sure PEPFAR saves a few kids from diseases, but it doesn’t address the root cause of the problem which is [gestures vaguely at whatever bit of one's ideology explains why bad things happen].”

Now, when it comes to PEPFAR, this is simply factually wrong. It is addressing the root of the problem. You can see this if you look at a graph of HIV deaths over time:

People also make the same objection to effective altruism. They’ll say: “effective altruism only is a palliative solution, that doesn’t address the [...]

---

First published:

Source:

Narrated by TYPE III AUDIO.

---

Images from the article:

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

Open Philanthropy is launching a big new Request for Proposals for technical AI safety research, with plans to fund roughly $40M in grants over the next 5 months, and available funding for substantially more depending on application quality.

Applications (here) start with a simple 300 word expression of interest and are open until April 15, 2025.

Overview

We're seeking proposals across 21 different research areas, organized into five broad categories:

1. Adversarial Machine Learning
   • *Jailbreaks and unintentional misalignment
   • *Control evaluations
   • *Backdoors and other alignment stress tests
   • *Alternatives to adversarial training
   • Robust unlearning
2. Exploring sophisticated misbehavior of LLMs
   • *Experiments on alignment faking
   • *Encoded reasoning in CoT and inter-model communication
   • Black-box LLM psychology
   • Evaluating whether models can hide dangerous behaviors
   • Reward hacking of human oversight
3. Model transparency
   • Applications of white-box techniques
   • Activation monitoring
   • Finding feature representations
   • Toy models for interpretability
   • Externalizing reasoning
   • Interpretability [...]

---

First published:

Source:

Narrated by TYPE III AUDIO.

Summary

Task: Make an interesting and informative Fermi estimate
Prize: $300 for the top entry
Deadline: February 16th, 2025
Results Announcement: By March 1st, 2025
Judges: Claude 3.5 Sonnet, the QURI team

Motivation

LLMs have recently made it significantly easier to make Fermi estimates. You can chat with most LLMs directly, or you can use custom tools like Squiggle AI. And yet, overall, few people have taken much advantage of this. Today we’re launching a competition to encourage exploration.

What We’re Looking For

Our goal is to discover creative ways to use AI for Fermi estimation. We're more excited about novel approaches than exhaustively researched calculations. Rather than spending hours gathering statistics or building complex spreadsheets, we encourage you to:

• Let AI do most of the heavy lifting
• Try unconventional estimation techniques
• Experiment with multiple approaches to find surprising insights

The ideal submission might be as simple as [...]

---

Outline:

(00:04) Summary

(00:31) Motivation

(00:54) What We're Looking For

(01:37) Task

(02:21) Submission Format

(02:57) Examples

(03:10) Important Notes

(03:55) Support and Feedback

(04:11) Appendix: Evaluation Rubric and Prompts

---

First published:

Source:

Narrated by TYPE III AUDIO.