What listeners think Spencer's wrong about: steel-manning critiques
Host Spencer Greenberg addresses critiques of his work and views, discussing topics from cult definitions and Clearer Thinking's tools to AI watermarking, decision-making, and his life philosophy of valuism.
Deep Dive Analysis
20 Topic Outline
Introduction to Critiques and Episode Format
Defining Cults and the 'Good Cult' Argument
Cult-like Structures in Companies and Countries
Critiques on Clearer Thinking Tools' Efficacy
Clearer Thinking's Strategy: Free Tools vs. Direct Influence
Legitimacy of Clearer Thinking's Research and Peer Review
AI Text Watermarking: Purpose and Technical Feasibility
Prevalence of AI Cheating and Watermarking Limitations
Self-Questioning vs. Speaking Out: The Smart Person's Dilemma
Converting Probabilities and Uncertainties into Decisions
Critiques on Elon Musk's Hand Gesture Study Methodology
Spencer's Personal View on Elon Musk's Hand Gesture
Critiques on Coining New Terms: Importance Hacking
Critiques on Over-reliance on Self-Report in Research
Accounting for Individual Differences in Clearer Thinking Tools
Critiques on Valuism and Potentially Harmful Values
Critiques on the 'Is Magic Real?' Essay and Redefinition of Magic
Critiques on AI Image Generation Bias and Stereotypes
A Formative Personal Critique: Communicating Warmth
Strategies for Tackling the Climate Crisis
6 Key Concepts
Cult (Spencer's Definition)
Cults exist on a spectrum, characterized by a small group of leaders with special knowledge, separation from society (sometimes physically), discouragement of outside information, and demands for self-sacrifice from members that often benefits the leader. They typically require high conformity and involve unethical behavior by leaders or members.
Gray Thinking
This mental model involves seeing the world probabilistically rather than in true/false binaries. It encourages thinking about things as having both good and bad traits, rather than being all good or all bad, and comparing benefits and weaknesses.
Importance Hacking
A strategy used in academia to make a study appear valuable, novel, or interesting enough for publication when it might not be. Unlike p-hacking, which manipulates statistical results, importance hacking misleads reviewers about the nature of a result to get it published.
AI Text Watermarking
A technology where an LLM AI hides information within the text it generates, making it much more reliable to detect if the text was AI-generated. This is achieved by subtly raising or lowering the probability of certain words being generated at each step, in a pattern that can be reverse-engineered.
Valuism
A life philosophy with two core components: first, identifying one's intrinsic values, and second, taking effective actions to bring more of those intrinsic values into the world. It is presented as a framework for guiding life choices, rather than a normative moral system.
Magic (Spencer's Essay Definition)
In Spencer's essay, magic refers to situations where people perceive things as having non-psychological properties, but these properties are actually purely psychological. These 'magical' phenomena exist within our brain's internal simulation of reality, similar to how we experience colors.
10 Questions Answered
Spencer believes that most of the perceived benefits of 'good cults' are actually attributes of tight-knit communities with shared meaning and values, which he distinguishes from cults due to the latter's specific negative characteristics like leader benefit and unethical behavior.
Spencer believes their tools are important because critical thinking improves decision-making and societal impact, and modules on habits and goal setting have shown direct benefits, with ongoing efforts to improve teaching effectiveness and real-world application.
Clearer Thinking's strategy is to reach a broad adult audience interested in self-improvement, as this best fits their team's skill set and expertise, rather than focusing on children's pedagogy or political influence where they lack deep ties.
Spencer acknowledges peer review's benefits but finds it often marginally useful and ineffective at weeding out bad work, citing the replication crisis; Clearer Thinking instead relies on extensive public and beta tester feedback, and sometimes collaborates with academics for peer-reviewed publications.
Yes, Spencer was informed by an involved party that at least one company has effective technology, which works by subtly altering the probability of specific words during text generation, allowing for reliable detection of AI origin.
Spencer advocates for both: self-questioning and acknowledging mistakes are important for improvement, but they should not prevent smart people from contributing their knowledge and ideas to the world, as making mistakes is a natural part of human endeavor.
For most low-stakes decisions, one should aim to maximize expected value, which is mathematically proven to be the optimal long-term strategy; for high-stakes decisions, risk of ruin must also be considered, and seemingly binary choices often have nuanced implementation options.
Spencer notes that Clearer Thinking's research on intrinsic values found little evidence of inherently harmful values like creating suffering; while individuals might still cause harm in pursuing their values, most people also intrinsically value not harming others, and Spencer would personally oppose harmful actions based on his own values.
Spencer feels uncertain about whether Elon Musk intended to make a Nazi salute, acknowledging evidence both for (support for right-wing party, trolling behavior) and against (Nazism's extreme unpopularity), and believes the study he ran has little bearing on Musk's intent.
Spencer suggests three realistic strategies: electing leaders who support government collaboration (especially among major global powers), investing in better technology to make reducing emissions self-interested, and increasing pressure on megacorporations responsible for a large share of greenhouse gases.
18 Actionable Insights
1. Cultivate Critical Thinking
Develop critical thinking skills, such as nuanced and probabilistic thinking, to better understand the world and make informed decisions in all areas of life. This helps navigate contradictory information and avoid false beliefs.
2. Pursue Intrinsic Values
Adopt valuism by first identifying your intrinsic values, then actively taking effective steps to manifest those values more fully in the world. This provides a framework for guiding your life, especially if you don’t believe in objective moral truth.
3. Form Healthy Habits
Utilize the free “Daily Ritual: A Habit Creation System” program from Clearer Thinking to learn simple techniques for forming new, beneficial daily habits. Research shows this program helps create stickier habits over time.
4. Maximize Expected Value
For most low-stakes decisions, aim to maximize expected value given probabilities, as this strategy is mathematically proven to be optimal in the long term. This approach helps you make choices that yield the best outcomes on average.
5. Act Despite Imperfection
Don’t let self-questioning or the fear of making mistakes prevent you from contributing to the world or expressing your viewpoints. Making mistakes is a natural human experience and not a reason to hold back.
6. Set Goals Effectively
Improve your ability to achieve desired outcomes by first understanding your intrinsic values, then applying effective goal-setting strategies. This direct approach helps you align your actions with what truly matters to you.
7. Seek Diverse Feedback
Actively solicit and seriously consider feedback and critiques from a wide variety of people, including experts and users, to significantly improve the quality and effectiveness of your work. This process provides valuable insights that can make your work much better.
8. Design Adaptive Policies
When making decisions or policies with uncertainty (e.g., 75% sure), build in adaptive mechanisms like monitoring or alternative pathways to account for potential errors. This approach allows for flexibility and correction if the initial plan proves less effective.
9. Distinguish Community from Cult
When seeking benefits like strong group identity and meaning, differentiate between healthy tight-knit communities and cults. Cults, unlike communities, often involve harmful elements like isolation, extreme conformity, and leader-benefiting self-sacrifice.
10. Identify Cult-Like Groups
Evaluate groups on a spectrum based on properties such as charismatic leadership, separation from society, discouragement of outside information, self-sacrifice benefiting leaders, high conformity, and unethical behavior. This helps in understanding the potential cult-like nature of a group.
11. Avoid Single Ideology Traps
Be wary of falling into false beliefs by thinking that one simple ideology has all the answers to life’s questions. Over-reliance on a single ideology often leads to detrimental outcomes because no single framework is universally comprehensive.
12. Account for Self-Report Bias
When interpreting self-report data, be mindful of its inherent unreliability due to factors like incentives to lie, memory limitations, and ego-driven exaggerations. Employ strategies such as anonymous reporting to improve data accuracy.
13. Optimize for Average Benefit
When developing tools or interventions, prioritize optimizing for average user benefit while actively working to avoid harm. Recognize that achieving perfect individual adaptation is often unfeasible and that average positive impact is a worthy goal.
14. Express Positive Feelings Clearly
Make a conscious effort to communicate your positive feelings towards others, especially if your facial expressions are not naturally very expressive. This helps ensure others understand your warmth and avoids creating unintended insecurity.
15. Advocate AI Text Watermarking
Support and advocate for major AI companies to implement AI text watermarking technology. This creates a crucial barrier against the abusive use of AI for cheating and misinformation, making it harder to pass off AI-generated content as human work.
16. Recognize AI System Biases
Be aware that AI systems, particularly those trained on vast internet data, tend to reflect existing societal biases unless specifically corrected. Understand that even with correction efforts, biases can still manifest in unexpected ways.
17. Engage in Realistic Climate Action
To make a plausible impact on climate change, focus your individual efforts on strategies such as electing leaders who promote government collaboration, investing in or developing better technology, or increasing pressure on megacorporations. These approaches offer more realistic avenues for change.
18. Steel Man Critiques First
When addressing criticism, first attempt to make the strongest possible argument for the opposing position (steel man it) before responding. This protocol ensures a thorough, fair, and well-considered response to the critique.
5 Key Quotes
The best lack all conviction while the worst are full of passionate intensity.
Spencer Greenberg (quoting Yeats)
If you wait until you're perfect to do things well, then you'll be waiting forever.
Spencer Greenberg
If you put a lot of ideas out in the world, you'll be wrong sometimes. And that's just something you have to accept. And that's okay.
Spencer Greenberg
I think it's bad when we can't tell what's generated by AI. I think, for example, a recent thing came out where it turned out people were using AI to generate lots of responses on Reddit. And I think this is a bad thing when AI can start appearing everywhere without us realizing it.
Spencer Greenberg
I think self-report really does have all these issues. But first of all, they're smart strategies to get around a lot of the issues. And second of all, the issues don't always apply.
Spencer Greenberg
2 Protocols
Strategies for Improving the World with Educational Content
Spencer Greenberg- Go mass market, focusing on adults (e.g., Clearer Thinking's approach: targeting top 25% of reflectivity and interest).
- Focus on educating children early to change their life trajectories.
- Focus on people with higher levels of influence (e.g., politicians, business leaders) for outsized impact.
Strategies to Tackle Climate Change
Spencer Greenberg- Government collaboration: Major governments (China, US, India, Europe) make a pact to seriously address climate change, potentially by electing interested leaders.
- Better technology: Invest in technologies that make it in people's self-interest to reduce greenhouse gas emissions (e.g., starting or joining relevant companies, carbon capture).
- Increasing pressure on megacorporations: Influence the few hundred companies responsible for a large proportion of greenhouse gases through agreements or investor pressure.