Education and Charity (with Uri Bram)

Oct 13, 2020 Episode Page ↗
Overview

Spencer Greenberg speaks with Uri Bram about universities as cult institutions, global poverty interventions like deworming and GiveDirectly, and the effective altruism movement's approach to expected values. They also discuss a critique of GiveWell's deworming recommendation and GiveWell's reply.

At a Glance
16 Insights
51m 55s Duration
15 Topics
6 Concepts

Deep Dive Analysis

Universities as Cult Institutions

Critique of University Education's Practicality

University's Role in Cultural Acculturation

Opportunity Costs of Higher Education

The Nature of Knowledge and Memorization

GiveWell's Deworming Charity Recommendation

Critique of Deworming Evidence: Study Limitations

Critique of Deworming Evidence: Probabilistic Assessment

Critique of Deworming Evidence: Lack of Mechanistic Paths

Effectiveness of Direct Cash Transfers (GiveDirectly)

GiveWell's Outcome Metrics and Conservatism

Trustworthiness and Transparency of GiveWell

Evaluating High-Risk, High-Reward Interventions

Understanding Causal Structures in Problem Solving

GiveWell's Official Response to Critiques

University as a Cult

Uri Bram's perspective that universities function as institutions for indoctrinating elites into a specific culture, where esoteric knowledge is rarely used outside perpetuating the system itself.

Post-Hoc Rationalization (in education)

The idea that common defenses for university education, such as 'learning how to learn' or 'thinking skills,' are often justifications made after the fact to support an arbitrary set of academic disciplines, rather than being the primary, intentional reasons for their selection.

Cultural Acculturation (by universities)

The process by which universities immerse students, particularly those from different social classes, into the norms, values, and behaviors of the upper class, contributing to a significant social and cultural divide in society.

Expected Value Framework (in charity evaluation)

A method used by organizations like GiveWell to recommend charities, where the potential impact of an intervention (even if the probability of success is low) is multiplied by its cost-effectiveness to determine its overall value.

Mechanistic Paths (in interventions)

The specific, observable, and understandable ways an intervention is expected to produce its desired outcomes. A lack of identifiable mechanistic paths for an intervention, despite observed long-term effects, can increase uncertainty about its actual effectiveness.

GiveWell's Core Outcome Metrics

The ultimate goals GiveWell uses to evaluate charity effectiveness, which are primarily averting deaths and increasing incomes/consumption, rather than intermediate measures like improved literacy scores or other forms of general well-being.

?
Are universities like cults?

Uri Bram argues that if observed by aliens, universities would appear as cult-like institutions where elites are indoctrinated into a specific culture, gaining esoteric knowledge rarely used outside perpetuating the system.

?
Is the knowledge taught in universities practical for most careers?

Uri Bram suggests that the specific factual knowledge and skills learned in university are generally not used by most people in their careers, with only those entering academia directly applying it.

?
How does GiveWell evaluate charities like deworming programs?

GiveWell uses an expected value framework, considering the possibility that an intervention may have a large impact on long-term economic well-being, even if the probability of that impact is low, due to the low cost of the treatment.

?
What are the main criticisms of GiveWell's deworming recommendation?

Uri Bram's critiques include that the recommendation relies on very few, geographically overlapping studies, questions the validity of combining multiple uncertainties to arrive at a low probability of effect, and notes the lack of observed mechanistic paths (e.g., no improvement in grades or weight) despite projected long-term income increases.

?
What outcomes does GiveWell prioritize when recommending charities?

GiveWell primarily recommends charities based on their potential for averting deaths and increasing incomes/consumption, rather than intermediate measures like improved literacy scores or general well-being.

?
Does GiveWell only count long-term wealth/income increases for cash transfers?

No, GiveWell models both short-term and longer-term benefits to cash transfers, including immediate spending that improves quality of life, not just investments that increase future income.

?
Why is it hard to find highly effective, low-risk charitable interventions?

Spencer Greenberg suggests that highly effective interventions with a high chance of success are likely to be competitive and already implemented, making it more probable that remaining opportunities involve higher risks or complexities.

?
Why do many development interventions fail despite seemingly simple solutions?

Many interventions fail because they don't account for the complex causal structures of problems and human behavior, such as cultural resistance to new techniques or the difficulty of maintaining supply chains for simple solutions like water chlorination.

1. Understand Problem Causal Structure

When attempting to solve complex or persistent problems, thoroughly investigate their causal structure and understand why previous efforts have failed. This approach helps in developing effective strategies that address root causes and avoid repeating past mistakes.

2. Evaluate Education Opportunity Costs

Critically assess the opportunity cost of higher education, considering if the knowledge and skills gained are truly commensurate with the significant time and energy invested. This encourages a thoughtful approach to educational choices and their long-term impact.

3. Prioritize Concepts Over Facts

In learning, focus on grasping fundamental concepts and ‘building block’ knowledge rather than just memorizing isolated facts. This enables flexible application of knowledge and problem-solving, as specific details can often be referenced when needed.

4. Combine Memorization with Rules

To effectively learn and solve problems, integrate the memorization of basic ‘building blocks’ (e.g., simple arithmetic) with flexible rules and principles. This allows for efficient decomposition and resolution of more complex challenges.

5. Cultivate Optimization Mindset

Adopt an ‘optimization mindset’ by conceptualizing problems as finding the highest peak on a metaphorical mountain, and consider various algorithmic approaches to move from your current state to an optimal solution. This framework can be applied to diverse challenges in everyday life.

6. Write for Mass Audiences

When writing, aim to create content that is inherently interesting and engaging for a broad audience, rather than just for a single required reader. This involves focusing on crafting compelling narratives and selecting topics that resonate with people who have a choice to read.

7. Skepticism for Unclear Mechanisms

Increase your uncertainty about the efficacy of interventions, especially those with long-term effects, if clear and plausible mechanistic paths for the observed outcomes cannot be identified. A lack of understanding of ‘how’ an intervention works can indicate a spurious or unreliable result.

8. Question Combined Uncertainties

Be critical of analyses that combine multiple uncertainties by multiplying probabilities, particularly when each potential issue could independently lead to a zero outcome. This method may inflate the expected value of highly speculative interventions.

9. Seek High-Risk, High-Reward

Consider pursuing opportunities that have a high chance of failure (e.g., 90%) but promise a massive positive impact if successful, similar to a startup venture. These less competitive areas may offer significant value and are often overlooked by others.

10. Beware ‘Easy’ Problem Solutions

Approach seemingly ’easy’ solutions to persistent problems with caution, as they often mask hidden complexities or cultural barriers. Deep investigation into why previous attempts have failed is crucial to developing truly effective strategies.

11. Understand Cultural Adoption Barriers

When introducing new solutions or technologies, thoroughly understand the local cultural context and potential resistance to change. Even highly beneficial innovations may not be adopted if they conflict with existing practices or beliefs.

12. Prioritize Real-World Impact Metrics

When evaluating charitable interventions, focus on ultimate, objective impact metrics like long-term income or child nutrition, rather than easily quantifiable intermediate variables (e.g., number of people served) that may not directly correlate with improved well-being.

13. Read Full Expert Analyses

To make truly informed decisions and engage in meaningful debate, commit to reading the complete, detailed analyses and blog posts from expert organizations, rather than assuming others have already done so.

14. Leverage Transparency for Critique

Actively engage with and critique transparent organizations, as their openness allows for detailed scrutiny that can help them identify flaws in reasoning and continuously improve their methods and recommendations.

15. Address Capital Constraints

Recognize that poverty often prevents individuals from making long-term beneficial investments, even if they are more cost-effective over time. Providing direct cash transfers can empower capital-constrained individuals to make choices that improve their long-term financial well-being.

16. Ensure Intervention Supply Reliability

When implementing interventions that require ongoing access to resources, establish a robust and reliable supply chain. Inconsistent availability can lead to people abandoning the intervention, even if it is beneficial.

If some aliens came down from space... they would basically just see this as a cult.

Uri Bram

The entire teaching apparatus had been geared around facts, and not concepts so much.

Uri Bram

It's not the actual formula or something like that, that we have to remember. It's that we have to kind of have a hook in our mind that there is something to know about on a topic.

Spencer Greenberg

My critique of universities is actually a lot more cultural than it is about the specific things they teach.

Uri Bram

I often ask people like, do you know anyone dating anyone who didn't go to college? And the answer is almost overwhelmingly no.

Uri Bram

The problem with being poor is that you can't afford like good boots, and therefore your boots keep wearing out. And therefore you end up spending way more money on boots than the rich guy who buys a nice pair of leather boots that last him the rest of his life.

Uri Bram

The only reason I think it's worth, you know, reading GiveWell's blog posts and critiquing them is because yeah, like I really, really respect them.

Uri Bram
2%
Probability that deworming has the projected long-term income effect (GiveWell's model) GiveWell's own calculated probability, which they argue is still worthwhile due to the intervention's low cost and high potential impact if successful.
Less than $1
Cost per deworming treatment According to GiveWell, contributing to its high cost-effectiveness in their expected value framework.
10%
Time spent in college as a percentage of a typical working career Based on 4 years of college for a roughly 40-year career, highlighting the opportunity cost of higher education.
~34%
Percentage of US adults with a college degree Mentioned by Spencer Greenberg, illustrating the significant social divide between college-educated and non-college-educated populations.
15-20 years
Time until deworming impact on income is observed The duration after childhood deworming treatment that studies observed increased adult incomes.