The differences between analytic and continental philosophy (with Alexander Prescott-Couch)
Yosha Bach discusses intelligence as model-making for control, distinguishing it from rationality and wisdom. He explores how understanding our self-narratives and values can reduce suffering, and advocates for a computational view of philosophy and reality. The conversation also touches on AI's role in understanding the human mind and future directions.
Deep Dive Analysis
17 Topic Outline
Defining Intelligence: Models, Rationality, and Control
Dimensions of Intelligence and the Concept of Sentience
IQ Tests, Fluid vs. Crystalline Intelligence, and Pattern Recognition
Human Causal Inference vs. AI Statistical Models
The Role of Embodiment and Online Learning in Understanding the World
GPT-3 in a Robot Body: A Thought Experiment on AI Limitations
The Westworld Analogy: Separating Mind from 'Monkey' Desires
Eastern Philosophy: Suffering, Regulation, and Self-Understanding
Methods for Overcoming Emotional Pain and Disagreement
The Evolution of Understanding Love: Infatuation to Shared Sacredness
AI as a Philosophical Project: Mathematizing the Mind
Gödel's Incompleteness Proof and Computational Semantics
The Nature of Reality and Existence from a Computational Viewpoint
Disambiguating the Meaning of 'Existence' and 'Truth'
The Future of AI: Beyond Differentiable Programming
Limitations of Deep Learning and the Third Wave of AI Systems
Critique of the 'In-Group vs. Out-Group' Mentality
8 Key Concepts
Intelligence
The ability to make models, distinct from rationality, smartness, or wisdom. It operates in the service of control, allowing a system to predict future outcomes and choose actions that lead to preferable branches.
Sentience
The discovery of one's own nature as an agent in the world and understanding the relationships one has to that world. It is a specific capacity of an intelligent modeling system, distinct from consciousness.
Fluid Intelligence
The capacity that enables an individual to solve new problems and acquire new skills. It involves creatively seeing new patterns and connections in unfamiliar domains.
Crystalline Intelligence
Intelligence that is embodied in existing skills and knowledge. It represents the ability to use trained skills at a higher level, including reasoning and problem-solving based on accumulated experience.
Causal Structure
The underlying systems that produce observed patterns in the world. Humans primarily focus on discovering these generative systems rather than just recognizing correlations.
Computational Semantics
A philosophical perspective where truth is understood as a predicate assigned by executing an algorithm, making it a stateful notion tied to the procedure of its acquisition, rather than a stable, platonic concept.
Computable (Yosha's view)
A system or concept is computable if it can be described by states and transitions, meaning states can be mapped to other states in a regular fashion. This perspective implies that only results that an algorithm can actually compute have value.
Love (Yosha's view)
The ability to create next-level agency and shared purpose. It involves a discovery of shared sacredness, where individuals serve transcendent goals above their ego and are willing to sacrifice for common projects.
8 Questions Answered
Yosha Bach defines intelligence as the ability to make models, which is distinct from being rational, smart, or wise, and serves the purpose of control by allowing a system to predict and influence future outcomes.
Sentience is described as the discovery of one's own nature as an agent within the world and understanding the relationships one has to that world, distinct from consciousness.
Humans discover causal patterns by recognizing the flow of information and the rules by which it transforms over time, as the brain registers changes and what remains stable despite them.
While it's an open question, it's unclear if systems like GPT-3, by only learning correlations in text, can deduce the universe's structure and dynamics without embodied interaction, as human learning often starts with multimodal sensory experience.
The philosophical project of AI is to mathematize the mind by building machine minds that can perform philosophy with greater acuity than humans, using computational semantics to understand minds as executable mathematical models.
Computational semantics views truth as a stateful notion tied to the procedure or algorithm used to acquire it, rather than a stable, platonic concept, implying that truth is derived through a sequence of states.
Love is fundamentally the ability to create next-level agency and shared purpose, where individuals align on transcendent goals and are willing to make sacrifices for a common project, akin to how cells form an organism.
Spencer Greenberg believes the idea that one's in-group has all the right answers and the out-group is inherently bad is overrated, advocating for finding value and truth in diverse worldviews.
29 Actionable Insights
1. Detach from Fixed Identity
Recognize that your mind is separable from innate ‘monkey desires’ and a fixed identity; this allows you to choose who you want to be and explore different contexts, making you less powerful if you cling to a fixed identity.
2. Control Your Self-Narrative
Understand that your self is not determined by past stories; gain agency over the narratives you tell yourself to make them more truthful and relevant, thereby reducing suffering.
3. Minimize Suffering by Acceptance
Reduce suffering by understanding your true place in the world and focusing your efforts on regulating only what you can change, rather than fighting unchangeable aspects.
4. Separate Pain from Suffering
Practice separating physical pain from the suffering it causes by stripping away the narrative that labels the pain as ‘bad,’ viewing it instead as a neutral sensation.
5. Journal for Clarity
When overwhelmed by thoughts or emotions, write them all out to prevent circular thinking, quiet your mind, and enable access to higher levels of understanding and behavior.
6. Resolve Conflict by Values
To resolve disagreements, especially political ones, delve two levels deeper to understand how both parties arrived at their values and opinions, enabling negotiation from a shared understanding of value construction.
7. Cultivate Mature Love
Cultivate mature love by understanding it as built on shared purposes and a sense of sacredness, moving beyond infatuation and projections to negotiate everyday life together.
8. Seek Good in All Worldviews
Actively look for positive aspects and valid points in all worldviews, even those you disagree with, to gain a more accurate and nuanced understanding of the world.
9. Gain Situational Understanding
To overcome emotional pain like lovesickness, seek a higher-level understanding of the situation by considering others’ perspectives and your role, moving beyond projections to see true relationships.
10. Identify Shared Sacredness
Discover shared ‘sacredness’ – purposes above ego for which you and others are willing to sacrifice – to foster non-transactional support and collaboration.
11. Love as Shared Purpose
Understand love as the ability to create next-level agency through shared purpose, fostering collaboration that transcends individual self-interest.
12. Seek Causal Structure
Beyond recognizing patterns, prioritize discovering the causal structures and underlying systems that produce those patterns, as humans primarily do this after early perceptual learning.
13. Perceive Information as Change
Understand that information is fundamentally about difference and change; focus on registering changes to build models of what remains stable in the world.
14. Recognize Information Flow
Direct your intelligence towards recognizing the flow and transformation of information, as this is how the universe becomes intelligible.
15. Define Intelligence Clearly
Understand intelligence as the ability to make models, distinct from rationality, smartness, or wisdom, to better conceptualize cognitive abilities.
16. Intelligence for Control
View intelligence as the ability to create models that serve control, enabling an agent to predict future outcomes and choose preferable actions.
17. Discover Your Place
To become sentient, model your own nature as an agent in the world and understand your relationships to your environment.
18. Understand Intelligence Types
Recognize the distinction between fluid intelligence (solving new problems) and crystallized intelligence (skills acquired), as IQ tests don’t fully measure all capabilities.
19. Assess Algorithmic Intelligence
Consider using the task of writing programs from scratch, especially for children who haven’t learned programming, as a predictor for later cognitive performance.
20. Embrace Computational Semantics
Shift from a stateless, platonic view of truth to a computational, stateful notion, understanding truth as tied to the procedure by which it is acquired, which has implications for understanding the universe and ourselves.
21. Value Computable Results Only
Adopt the principle that only computable results truly have value, fundamentally changing how one approaches mathematics and understanding of reality.
22. View Pi as a Function
From a computational perspective, understand irrational numbers like Pi not as fixed values, but as functions or procedures that can only be computed to a certain degree.
23. Distinguish Function and Value
Recognize the fundamental difference between a function, which is a procedure, and a value, which only exists to the degree that its generating procedure can be executed.
24. Avoid Infinite Language Contradictions
Understand that languages treating infinities as existing entities can become self-contradictory; rethink how you use language to express concepts involving infinity.
25. Use Computational Languages for Consistency
To create internally consistent languages and models, adopt a computational framework where everything is defined from first principles.
26. Model Continuous Space with Finite Lattices
When describing continuous space, work with finite-resolved lattices rather than assuming infinite resolution, as true continuous space is not computable in practice.
27. Use AI for Mind Understanding
Engage with artificial intelligence as a productive field to understand how minds work, by creating and testing computational models of cognitive processes.
28. Learn Minds from AI Failures
Build AI models, even those that fall short of human capabilities, to gain deeper insights into what human minds are and are not.
29. View Particle World as Quantum Implementation
Adopt the mental model that our observable particle universe is an inefficient implementation running on a more fundamental quantum substrate universe.
7 Key Quotes
Personally, I think of intelligence as the ability to make models, and it's distinct from being rational, right? A lot of irrational people are highly intelligent and vice versa.
Yosha Bach
Sentience is not the same thing as consciousness. It's basically just a discovery of your own nature as an agent in the world and the relationships that you have to the world.
Yosha Bach
The true message of Westworld is not a story about robots. It's a story about ourselves. Because we are like these robots.
Yosha Bach
Suffering is not the result of the universe doing something to you as an agent. The universe that you experience is not the physical universe.
Yosha Bach
I think that philosophy has been stagnating over the last maybe even couple hundred years, in part because we reach the limits of what the human mind can do without augmentation in philosophy.
Yosha Bach
The computer is teaching us how we are computers, of course. But the deeper insight is that by building models that fall short of how our minds work, we understand better what our minds are and what they are not.
Yosha Bach
I think one really overrated idea is that your in-group has the right answers and the out-group is bad and is undermining everything.
Spencer Greenberg
1 Protocols
Dealing with Emotional Pain and Suffering
Yosha Bach- Sit down and write out all thoughts and feelings to stop running in circles.
- Get present thoughts out of the way to allow for higher-level understanding.
- Go to a higher level of abstraction to perceive the situation, actors, and their true relationships more objectively.
- Understand how your own values and the values of others were constructed.
- Identify not with individual 'childish' behaviors or desires, but become the 'conductor' of your internal 'metal orchestra', nurturing parts as needed but not letting them dominate.