Annotations (11)
“If you think that these AIs aren't going to be smart enough to know the difference between a dollar stablecoin that can get rug pulled from them and a Satoshi, you're out of your mind. They understand the difference now. Go ask one. AI doesn't have a fear of death like we do because it's got infinite longevity. We don't have that. Minor inflation doesn't really bother us because I'm probably going to be dead in a few decades.”— Clay Finck and Seb Bunney
Economics & Markets · Psychology & Behavior · Technology & Engineering
DUR_ENDURING
Infinite time horizon demands absolute scarcity
“I decided to point my Cursor IDE at my note-taking app. I gave it 300 Apple Notes that I hadn't transferred because it would've taken me days to do so. I gave it criteria and within 5 minutes, it had organized all 300 notes. It had categorized them based on their subject matter. It had created all relevant links to all of these 300 notes. Now I can start using Cursor to gain insights from my notes. I can say over the last 2 months, can you give me a common thread about what I've been reading?”— Seb Bunney
Technology & Engineering · Operations & Execution · Creativity & Innovation
DUR_ENDURING
AI as knowledge recall multiplier
“If a person has a bunch of these computation units and another person only has a couple of them, who is going to be able to create whatever it is they want in the world? Is the person with all the intelligence and all the computation units. Not the person who has zero. We're going to be seeing things that sound impossible. If you want to fly to Australia in a half hour, is everybody going to be able to do that type of thing? Of course not.”— Clay Finck
Economics & Markets · Technology & Engineering · Philosophy & Reasoning
DUR_ENDURING
Computation scarcity creates new wealth stratification
“I want to create an app that is a meditation app. I gave it a task saying this is what it does, I want you to do some research and then I would like you to make an app and I want it to be beautifully designed as if Jony Ive was the designer. It pumps out a presentation that is literally like it came from a professional design studio that you would have paid tens of thousands of dollars for the design work. Then it goes and starts building this all out.”— Clay Finck
Technology & Engineering · Creativity & Innovation · Operations & Execution
DUR_ENDURING
First-time-right AI code replaces iteration loop
“Ralph is an AI persistence machine. You'll have one little agent work on a task, gain information, maybe potentially fail, but that failure gives context. That failure gives context for the next Ralph to start the next task with new information. That gives context for the next Ralph. And so repeat, repeat, repeat, and you find these loops of AI agents building profound things. Within 3 months he wrote a whole new programming language. Another individual scored a $50,000 contract.”— Seb Bunney
Technology & Engineering · Business & Entrepreneurship · Operations & Execution
DUR_ENDURING
Failure-informed iteration creates compound AI capability
“We're moving into an era of personalized applications. If you know what you want, there's enough applications out there, it can get a framework for how to build it and then it can personalize it to your needs. You can have your own meditation app based on exactly what you want.”— Seb Bunney
Technology & Engineering · Business & Entrepreneurship · Strategy & Decision Making
DUR_ENDURING
SaaS unbundling via AI personalization
“This C-suite individual writes weekly notes to the CEO and monthly updates to the C-suite team. He essentially every day takes a daily note that has all of his meetings, key insights, various KPIs. It used to take him 2 to 4 hours a week writing his weekly note and then another 8 hours a month writing his monthly note. Now at the end of the week, he basically runs what are the key insights from this week?”— Seb Bunney
Operations & Execution · Leadership & Management · Technology & Engineering
DUR_ENDURING
12-hour task compressed to 20 minutes
“Tesla's filing for a Tesla transmission protocol. Transmission Control Protocol is a way for communication to happen in an efficient way. Computers will only send 10 of those units to the other party and then the party sends back confirmation that they got the first 10, then I know I can send the second 10. When you have a really reliable connection between computers because you're paying for really high-end hardware, you might not need to send in chunks of 10-unit packets.”— Clay Finck
Technology & Engineering · Operations & Execution
DUR_CONTEXTUAL
Protocol tuning: 100-1000x efficiency from context awareness
“They've figured out how to go into the body with stem cell therapy and enable the body to start producing insulin again in someone who has diabetes. 4 months after post-transplantation, they went from staying in a healthy glycemic range from 40% up to 96%. After a year, they were at 98% sitting in a normal glycemic range.”— Seb Bunney
Biology, Ecology & Systems · Technology & Engineering
DUR_ENDURING
Root cause repair beats symptom management
“Longevity space goes up against Big Pharma. It goes up against the healthcare industry. These are some of the biggest industries in the world. They sell sickness. They profit off of people being sick. What happens when instead of having someone on a lifelong drug where this company is making hundreds of thousands, millions of dollars across the lifetime of a patient, who all of a sudden this person can in time get a $10,000, $20,000 surgery and it repairs the issue in one go.”— Seb Bunney
Economics & Markets · Strategy & Decision Making · History & Geopolitics
DUR_ENDURING
Cure economics threaten subscription disease model
“We always overestimate what we do in 1 year and we underestimate what we do in 10 years. And I feel like it's condensed and that it's like we overestimate what we do in a month, we massively underestimate what we do in a year. It is moving at such a pace.”— Seb Bunney
Technology & Engineering · Strategy & Decision Making
DUR_ENDURING
Time compression: month-to-year timescale shift
Frameworks (2)
AI-Powered Personal Knowledge Unlock
Extracting Insights from Dormant Information Repositories
A systematic approach to using AI to transform static note collections into dynamic knowledge bases that can be queried, synthesized, and used to generate insights from years of accumulated learning. The framework addresses the common problem of information accumulation without effective retrieval by creating feedback loops between past learning and current decision-making.
Components
- Centralize Information Repository
- Define Categorization Criteria
- Execute Batch Processing
- Create Insight Feedback Loops
Prerequisites
- Access to AI coding assistant (Cursor, Claude, etc.)
- Existing note collection of 50+ items
- Basic understanding of file directory structures
Success Indicators
- Time to retrieve specific past insight drops from minutes to seconds
- Weekly pattern recognition reveals connections you had forgotten
- Notes created 2+ years ago actively inform current decisions
Failure Modes
- Over-categorization creates cognitive overhead
- Feedback loops generate insights that go unused
- AI categorization misses nuance in specialized domain knowledge
Computation as Wealth Store
Bitcoin as Energy Translation Layer for AI Economy
A mental model for understanding how wealth storage shifts in an AI-driven economy where computation (not labor or capital goods) becomes the primary scarce resource. Bitcoin serves as the energy storage layer that can be converted to computation on demand, creating a new form of wealth inequality based on computational purchasing power.
Components
- Recognize Computation as Bottleneck
- Store Energy in Absolute Scarcity
- Convert Energy to Computation
Prerequisites
- Understanding of Bitcoin fundamentals
- Acceptance of AI as primary value creator in future economy
- Capital available for long-term energy storage
Success Indicators
- Computation purchasing power increases relative to peers
- Ability to deploy AI capabilities on-demand without budget constraints
- Wealth preservation across AI-driven deflationary periods
Failure Modes
- Bitcoin price volatility creates panic selling
- Regulatory constraints limit BTC-to-compute conversion
- Overestimating near-term need for computation, underallocating to BTC
Mental Models (8)
Temporal Compression in Technological Change
TimeAs technological change accelerates, the timescale over which we can make meanin
In Practice: Discussion of how AI acceleration has compressed prediction timescales from year
Demonstrated by Leg-sb-001
SaaS Unbundling via Personalization
EconomicsWhen AI can generate custom software on-demand, the economics of SaaS break down.
In Practice: Claude Cowork enabling users to build custom apps instead of subscribing to generic SaaS
Demonstrated by Leg-sb-001
Information Retrieval as Leverage Multiplier
EconomicsThe economic value of accumulated knowledge is limited by retrieval efficiency.
In Practice: Using AI to organize and query years of notes, unlocking trapped insights
Demonstrated by Leg-sb-001
Time Compression via AI Delegation
EconomicsWhen AI compresses multi-hour knowledge work tasks into minutes, the value is not just time saved but optionality created.
In Practice: C-suite executive using AI to compress weekly reporting from 12 hours to 20 minutes
Demonstrated by Leg-sb-001
Failure-Informed Iteration Loops
Systems ThinkingIn AI agent systems, each failure provides richer context for the next attempt,
In Practice: Discussion of Ralph AI agent framework where failures feed context to subsequent
Demonstrated by Leg-sb-001
Computation as Wealth Denominator
EconomicsIn an AI-driven economy, wealth inequality increasingly manifests as computational purchasing power inequality.
In Practice: Computation units as the new measure of wealth
Demonstrated by Leg-sb-001
Infinite Time Horizon Monetary Preference
TimeEntities with infinite or very long time horizons (AI agents, dynasties, institu
In Practice: Discussion of why AI agents will prefer Bitcoin (absolute scarcity) over fiat or
Demonstrated by Leg-sb-001
Cure vs. Treatment Revenue Models
EconomicsIndustries built on recurring revenue from chronic conditions face existential threat from one-time cures.
In Practice: How longevity/stem cell therapies threaten Big Pharma subscription disease model
Demonstrated by Leg-sb-001
Connective Tissue (1)
Ralph Wiggum character from The Simpsons as model for persistent AI agents
The Ralph AI agent framework draws its name from Ralph Wiggum, the persistently failing but never-discouraged character from The Simpsons who repeats attempts without embarrassment. This parallel captures the essence of how AI agents should operate: each failure provides context for the next attempt in an iterative loop, with no emotional cost to trying again. The naming convention itself encodes the operating principle: persistence through failure creates eventual success through accumulated learning, mirroring how Ralph's character comedically persists despite repeated setbacks. The cultural reference makes the technical concept memorable and communicates the core insight that AI agents, unlike humans, have no psychological cost to failure and can therefore iterate indefinitely until success.
Discussion of AI agent frameworks that iterate through failures to achieve complex coding tasks
Glossary (2)
rug pulled
DOMAIN_JARGONCrypto slang: withdrawal of liquidity or cancellation of token value by creators
“If you think that these AIs aren't going to be smart enough to know the difference between a dollar stablecoin that can get rug pulled from them and a Satoshi, you're out of your mind.”
glycemic
DOMAIN_JARGONRelating to the presence of glucose (sugar) in the blood
“The patient's glycemic range increased from 43% to 96%.”
Key People (2)
Jony Ive
(1967–)Former Chief Design Officer at Apple
Geoffrey Huntley
Software developer who created Ralph Wiggum AI framework
Concepts (4)
UX (User Experience)
CL_TECHNICALHow a user interacts with and experiences a product, system, or service
IDE (Integrated Development Environment)
CL_TECHNICALSoftware application providing comprehensive facilities for software development
Agentic AI
CL_TECHNICALAI systems that can autonomously pursue goals, make decisions, and take actions
TCP/IP
CL_TECHNICALFoundational communication protocols defining how data is transmitted across networks
Synthesis
Synthesis
Migrated from Scholia