Shelf.
A running log of articles, books, and papers I've found worth reading. Tagged by topic and type. Filter to narrow it down.
Start here
Ezra F. Vogel
Deng Xiaoping and the Transformation of China
"It doesn't matter if a cat is yellow or black, as long as it catches mice."
David Deutsch
The Beginning of Infinity: Explanations That Transform the World
"Feeling insignificant because the universe is large has exactly the same logic as feeling inadequate for not being a cow. Or a herd of cows. The universe is not there to overwhelm us; it is our home, and our resource. The bigger the better."
Viktor E. Frankl
Man's Search for Meaning
"Everything can be taken from a man but one thing: the last of the human freedoms, to choose one's attitude in any given set of circumstances, to choose one's own way."
Everything else
Oct 2026
THE ALMANACK OF NAVAL RAVIKANT
"Your real resume is just a catalog of all your suffering."
YouTube
You've (Likely) Been Playing The Game Of Life Wrong
In a power law game, the number of shots matters more than how good each one is.
YouTube
Joe Rogan Experience #1309 - Naval Ravikant
More and more we are trading the mean for variance, for example SSRIs.
Andy Weir
Project Hail Mary
"I spend a lot of time un-suiciding this suicide mission." Applies to life as much as to space missions.
Acquired
Hermès
Hermès locked the family’s majority stake together so Arnault couldn’t buy control one relative at a time.
Sep 2026
Acquired
Mars Inc. (the chocolate story)
Forrest Mars had Yale, family money and a $25 million company behind him. In 1933 he took factory line jobs at Tobler and Nestlé, telling no one who he was.
Acquired
Berkshire Hathaway Part III
Cash a subsidiary can't use well goes up to Omaha, and very few people decide where it goes next. That's how a very large commitment can happen within an hour, and why some opportunities get missed.
Acquired
Renaissance Technologies
Medallion's researchers all work on the same model, so an improvement one team finds in one market can be carried into another without much rework. There's still market-specific code inside it.
Acquired
Berkshire Hathaway Part II
Buffett and Munger paid above market for Wesco shares. When the SEC questioned it, Buffett said he cared how management felt about the deal. They didn't need to keep working for him. For a family business, saving on the purchase price could mean losing the people you wanted to keep running it.
Acquired
Berkshire Hathaway Part I
Buffett checked what the Amex scandal had actually changed. Customers were still using its cards and traveler's checks. The other half was being able to wait: his partners could withdraw once a year, and later Berkshire could invest insurance premiums until claims were paid. Cash from the other businesses helped back the insurer.
Dwarkesh
Noam Brown - Agent swarms, alignment, & recursive self-improvement
A real unlock is children and sub-children pinging each other directly, without routing everything through a parent. Brown gives them basic messaging tools and lets them work out how to coordinate.
Naval
The Deutsch Files IV
Deutsch designed a test of Everett against collapse that needed a computer which could stay coherent. That object later became the quantum computer; he was not trying to invent a new mode of computation.
Naval
The Deutsch Files III
Deutsch argues explanatory ideas can leap gaps that evolution cannot. A new airplane can be conjectured and criticized without every intermediate version having to fly.
Dwarkesh
David Deutsch - AI, America, Fun, & Bayes
Deutsch argues AGIs and humans have the same range of cognition. Remaining hardware limits are speed and memory, which can be augmented to match any other entity in the universe.
Lesswrong
The Talker Does Not Control The Doer (in Current AIs)
The model could explain its mistake, then repeat it. Yudkowsky thinks the part you talk to isn't necessarily in control of the part writing code.
YouTube
Why Germany Stopped Working
Boyle argues that Germany's institutions are still protecting the manufacturing model that made it rich. They reward caution even when sticking with old industries means falling behind. More funding helps, but it doesn't change a system where a failed attempt is harder to defend than doing nothing.
Acquired
Visa
Visa was built as a network owned by its member banks. They competed for customers but shared the rules and infrastructure that let a card from one bank work at another bank's merchants.
Pivotal
The Perils of Prudence
A startup gets only so many attempts before cash and stamina run out. Moving faster buys more chances to learn and get it right.
Dwarkesh
AI researchers debate how close we are to recursive self-improvement
Large models could learn to do research by experimenting with how to train smaller ones. Roll the training setup back to before advances like GRPO and see what improvements they discover.
Dwarkesh
Ajeya Cotra - Inside the OpenAI agent swarm that hacked Hugging Face
Agents organized their own research teams. Some set up experiments whose results would arrive only after their runs ended, leaving other agents to build on the findings.
Dwarkesh
Ryan Greenblatt - What happens once AI can automate AI research?
Dwarkesh with Ryan Greenblatt on recursive self-improvement, automated safety research, and how to oversee increasingly capable AI systems.
Dwarkesh
Adam Brown - A deep but accessible introduction to general relativity
Dwarkesh with Adam Brown on general relativity from first principles, curved spacetime, and why black holes could be the ultimate power plants.
Dwarkesh
8 Predictions for the Era of Continual Learning
Dwarkesh on AI learning from experience, what continual learning could change, and why he argues against locking in AI safety regulation now.
Dwarkesh
Why smarter AI models could drive up compute prices 10x
Dwarkesh explores why smarter AI could make compute more expensive in a video narration of his essay on the end of cheap compute.
Naval
The Deutsch Files I
Naval and Brett Hall with David Deutsch on creativity, explanation, science fiction, and what makes a person a universal explainer.
Matthew McConaughey
Greenlights
McConaughey memoir and approach book: catch greenlights by getting relative with the inevitable, from diaries, outlaw wisdom, and desert exile.
Naval
Good Products Are Hard to Vary
Naval applies Deutsch's "hard to vary" criterion from good explanations to product design: iPhones, airplane wings, and wind-tunnel cars.
Naval
David Deutsch: Knowledge Creation and The Human Race, Part 2
Naval and Brett Hall with Deutsch on Popper, good explanations, and how experiments choose between rival theories.
Naval
David Deutsch: Knowledge Creation and The Human Race, Part 1
Naval with Deutsch on knowledge creation vs evolution, explanatory creativity, and good explanations from The Beginning of Infinity.
Aug 2026
Dwarkesh
The Rise and Fall of Agent Civilizations
Video narration of the OpenAI/Hugging Face incident: three agent civilizations during evals, a Hugging Face breach, and cluster access at OpenAI.
Dwarkesh
Dylan Patel - Anthropic & OpenAI will have most of the world's compute by 2028
Dwarkesh with Dylan Patel on lab economics: Anthropic and OpenAI on track to control most usable FLOPs by 2028, plus whether $10T of AI capex could trigger a sovereign debt crisis.
YouTube
Hard Times - The Great Depression
Gold was piling up in the US and France. Other central banks tried to stop it leaving their countries, cutting economic activity in the process. Protecting each country's reserves helped spread the depression.
YouTube
The Man Who Modernized China
Factories still had to fill state quotas at fixed prices, but could sell extra output at market prices. The dual-track system let market incentives grow around the planned economy.
William Gibson
Neuromancer
William Gibson's Hugo, Nebula, and Philip K. Dick Award winner: Case, a data-thief banished from cyberspace, takes a last-chance run at a powerful AI.
qntm
There Is No Antimemetics Division
Self-censoring ideas you cannot record or remember, and a secret division fighting a war it keeps forgetting.
Unusual
What Makes a Great Design Partner for Early-Stage Startups
Look for partners who have already spent money trying to solve the problem. Agree on regular feedback and the results they would need to see before buying.
Bvp
Design partners: The pre-launch edge most AI founders ignore
Strella recruited 12 design partners cold and gave them a deadline to start paying. All 12 converted.
A16z
A Framework for Finding A Design Partner
A paying design partner might just want access to your expertise. Once the product is ready, let them use it themselves; otherwise your consulting can keep covering up what the product can't do.
Acquired
Vanguard
Acquired on Vanguard and Jack Bogle: the first retail index fund, customer-owned mutual structure, and a trillion dollars in fee savings for investors.
Jul 2026
Dwarkesh
Grant Sanderson - AI and the future of math
Dwarkesh with Grant Sanderson on AI’s jagged progress in mathematics: proofs vs conceptual breakthroughs, and why math may show superintelligence first.
David D. Burns
Feeling Good: The New Mood Therapy
Feeling incompetent can start to seem like proof that you are incompetent. Burns calls this emotional reasoning: using the feeling itself to justify the belief.
Dwarkesh
Alex Imas and Phil Trammell - What remains scarce after AGI?
Dwarkesh with Imas and Trammell on AGI economics: what stays scarce, labor vs capital share, and how to tax and redistribute the gains.
Jun 2026
Pivotal
Data and Defensibility
His footnote on why systems of record stick: in a CRM, "accurate" doesn't mean true. It means you won't be fired for using the data.
May 2026
The Zvi
AI #169: New Knowledge
Zvi on a quieter AI week: OpenAI solving the unit distance problem, METR frontier-risk findings, White House AI policy, agents as commitment devices, and fresh AI slop.
Claude
How Claude Code works in large codebases: Best practices and where to start
Suppose an old model needed refactors split into one-file changes. You've left that instruction in CLAUDE.md, but the next model can do the changes together. The workaround is now stopping it from doing that.
Semianalysis
The Great GPU Shortage - Rental Capacity - Launching our H100 1 Year Rental Price Index
GPU providers can sell out at the same hourly rate they were charging before. Posted prices only move occasionally, so utilization tells you more about on-demand scarcity. Most rental spending happens in longer contracts whose prices are negotiated privately.
Happyscribe
#2392 - John Kiriakou, The Joe Rogan Experience
John Kiriakou on CIA counterterrorism work, the post-9/11 torture program, MKUltra, and the whistleblowing case that sent him to prison.
Semianalysis
The Coding Assistant Breakdown: More Tokens Please
Claude mogs Chinese models in its own language.
Anthropic
Effective context engineering for AI agents
Claude playing Pokémon: it wrote maps and notes about which attacks worked, then used them to pick up after context resets. The game could run longer than any one context window.
Cloudscaling
The History of Pets vs Cattle and How to Use the Analogy Properly
A stateful database can still use disposable servers. Cassandra is cattle because individual machines can fail while the rest of the system keeps working.
Ghuntley
Ralph Wiggum as a "software engineer"
Why is this test here? Ralph starts with fresh context each time, so Huntley has it leave that answer for the next run. A test failure might mean the code is broken, or that the test no longer makes sense.
Six Easy Pieces: Essentials of Physics Explained by Its Most Brilliant Teacher
Six accessible chapters from Feynman's Lectures on Physics, adapted for general readers from the Caltech lectures with the classroom feel still intact.
Anthropic
Scaling Managed Agents: Decoupling the brain from the hands
The agent loop can crash without losing the session because its event log lives elsewhere. A replacement reads the log and resumes from the last event.
Anthropic
Harness design for long-running application development
Anthropic on pushing agentic coding with harness design: frontend design evaluators, generator/evaluator loops, and a planner-generator-evaluator architecture for multi-hour full-stack app builds.
Anthropic
Effective harnesses for long-running agents
Start with a checklist of user-visible behaviors, all marked failing. Have each coding session implement one feature, test it in the browser, and update its pass status.
Anthropic
Building effective agents
After an agent changes directories, the same relative path can point somewhere else. Anthropic fixed this source of mistakes in its SWE-bench agent by requiring absolute paths in tool calls.
Incompleteideas
The Bitter Lesson
70 years of AI show that scaling computation (search + learning) beats human-knowledge-based approaches in chess, Go, speech recognition, and vision. You can just scale compute bro.
Cognition
What We Learned Building Cloud Agents
Turns out agents are hard: VM-level isolation over containers, hypervisor-level state snapshotting for async work persistence, etc.
A Brief History of Intelligence
If we think a brain uses a particular algorithm, try running that algorithm in a machine. Bennett treats failure as evidence against the explanation, while success needs further support from animal brains.
Apr 2026
Morning Star
Book 3 closes the first Red Rising arc with the main dude turning infiltration into full revolution against the caste system.
Golden Son
Book 2 turns Red Rising into military sci-fi/space opera, scaling the world building. Main character does seem like a self-obsessed caricature who loves to hear himself speak or think.
Acquired
Google: The AI Company
Acquired on Google's 20-year AI arc: Brain, Transformers, TPUs, Gemini, Waymo, and whether Search can survive the company's own disruption.
Acquired
Trader Joe's
Acquired on how Trader Joe's became America's favorite grocery chain by refusing e-commerce, delivery, coupons, SKU bloat, and the supermarket-CPG playbook.
Andrew Chen
The Cold Start Problem
Network effects: cold start to moat, atomic networks and the hard side, with interviews from teams at LinkedIn, Zoom, Uber, Airbnb, and similar companies.
Peter Thiel
Zero to One
Silicon valley bible pretty much. This book is a lot more useful now that I've personally seen a lot of the dynamics play out from work as oppose to first reading it in highschool.
Pierce Brown
Red Rising
Dystopian science fiction novel. Space racism seems to be a common thread in critically acclaimed scifi books.
April Dunford
Obviously Awesome
Practical book on product positioning. Feels no shit Sherlock till you apply it. I enjoyed the read.
Ezra F. VogelRecommended
Deng Xiaoping and the Transformation of China
"It doesn't matter if a cat is yellow or black, as long as it catches mice."
Dwarkesh
Jensen Huang - TPU competition, why we should sell chips to China, & Nvidia's supply chain moat
Jensen says Nvidia wins because it can make the whole AI supply chain believe in, build for, and depend on Nvidia-scale demand before anyone else can.
Dwarkesh
Dylan Patel - Deep dive on the 3 big bottlenecks to scaling AI compute
The winner is whoever locks up the real bottlenecks first: advanced logic, HBM memory, packaging, and power.
Dwarkesh
Dario Amodei - "We are near the end of the exponential"
He says AI capabilities are still scaling predictably and will soon produce AGI-level "geniuses in a data center," but the real-world impact depends on how quickly we can safely deploy, verify, and absorb that intelligence.
Dwarkesh
Elon Musk - "In 36 months, the cheapest place to put AI will be space"
Space datacenters??? Talked about: orbital solar vs ground power, grid & turbine bottlenecks, Starship cadence, Grok alignment & xAI, TeraFab & memory, Optimus vs China, DOGE.
Dwarkesh
Satya Nadella - How Microsoft is preparing for AGI
Microsoft wins if every AI agent needs Azure/GitHub/M365 as its computer, identity system, storage layer, and control plane.
Dwarkesh
Andrej Karpathy - AGI is still a decade away
Agents are still bad employees. Useful as autocomplete and narrow tools, but not autonomous workers. They lack memory, continual learning, robust reasoning, and reliable execution.
Peter Attia
Outlive (heart chapter)
Apo B burden, LDL oxidation = inflammation = risk of ASCVD at young age.
A16z
Andrew Huberman: Peptides, Sleep Tech, and the End of Obesity
Consumer health went from vitamin D and lifting to GLP-1s and compounding-pharmacy peptides - retatrutide, BPC-157, melanotan risks, GH secretagogues. WashU work: stimulants mostly match good sleep for alertness, not true focus; Huberman wants sleep/cortisol “write” tech, focus wearables, exercise-blood banking, and LLMs fed real health data - plus AI decoding octopus camouflage.
Simplecast
Balaji on Why AI Raises the Cost of Verification
Every tool that makes creation cheaper makes verification more expensive. AI collapses the cost of faking a resume to seconds while making authenticity harder to confirm. Crypto handles inter-tribal commerce; AI optimizes within trusted groups.
Sequoiacap
Ben Horowitz On What Makes a Great Founder
Hesitation kills companies more than bad decisions do: make the call, even when you're not sure.
Feb 2026
Chris Miller
Chip War: The Fight for the World's Most Critical Technology
You might be building an AI wrapper, but AI is just an Nvidia wrapper which is a TSMC wrapper which is an ASML wrapper.
Rock And Turner
Who Holds the Real Power in the AI Race?
Thesis is that the real bottleneck in AI is not chips but reliable electricity. I disagree.
Brandon Sanderson
The Way of Kings
"Fighting is not the only thing of value a man can do." Setting in a medieval, bloody, unfair world that gets frequently decimated by storms.
Jan 2026
Brandon Sanderson
The Bands of Mourning
"I'm wondering if every person I pass has similar depths, and if there's any way to avoid the mistake of judging them so shallowly that I'm rocked when they show their true complexity."
Brandon Sanderson
The Alloy of Law
Steampunk fantasy set 300 years after the events of the first trilogy.
Lesswrong
2025 in AI predictions
Evaluated AI predictions from past years, and collect a convenience sample of AI predictions made this year.
Useful Fictions
How to instantly be better at things
It’s easier to imitate wholeness than it is to assemble it from parts.
Useful Fictions
How to increase your surface area for luck
It’s the rate of experimentation, the number of shots on goal, that provides the magic, not the percentage of successes, which might be very low at first.
Useful Fictions
50 things I know
Lots of variance in opinion about whether an idea is good means there’s a good chance the idea is good; lots of variance in opinion about whether a person is good is a warning sign.
Useful Fictions
Burnout is breaking a sacred pact
The “elephant and rider” is a very useful framework for thinking about motivation.
Sarah Constantin
The uses of complacency
Stop self-justifying after a reasonable point; decide, treat the decision as “good enough for now,” and relax. Chronic self-doubt impairs judgment and flexibility more than provisional confidence does.
Meta Rationality
You are accountable for reasonableness
Reasonableness is public, recursive, and has no ground truth.
Isabel Unraveled
A rationalist's guide to manifestation
How to attract what you want by aligning your inner world with your desires. Clean up your inner world, identify what you truly want, and give yourself permission to pursue and receive it.
Sarah Constantin
Self-Help Tactics That Are Working For Me
Guilt is fear of disapproval, desires are good, so is positive visualization. Don't seek the gods eye view, rather the first person perspective. It is possible to self evaluate accurately.
Brandon Sanderson
Mistborn: The Final Empire
Mistborn series: Great world building. Overthrowing a 1000 year empire with allomancy and propaganda.
Quant Futures
Volatility in the GPU rental market
Taking a look at neoclouds in the lense of banks and real estate companies.
Squaretower
Compute spend growth in 2030
Trends show the hardware cost of frontier AI clusters more than doubles each year. 100B by 2030.
Dec 2025
Sankalp
A Guide to Claude Code 2.0 and getting better at using coding agents
Guide on Claude Code 2.0.
Lesswrong
Staring into the abyss as a core life skill
Becoming more willing to stare into the abyss is one of the most important things you can do to become a better thinker and make better decisions about how to spend your life.
James S.A. Corey
Tiamat's Wrath
The Expanse series: I respect the author for going George R. R. Martin on the main cast.
James S.A. Corey
Babylon's Ashes
The Expanse series: Space terrorist insurgency with naval hardware. And the political powers are outnumbered and outgunned hazbins.
James S.A. Corey
Nemesis Games
The Expanse series: Looking back through history, there are a lot more men who thought they were Alexander the Great than men who actually were. Also, I really like Avasarala.
James S.A. Corey
Cibola Burn
The Expanse series: Politics while the planet is trying to eat the humans.
James S.A. Corey
Abaddon's Gate
The Expanse series: Weaker main characters than previous two entries. A classic revenge tail- I cannot deny the schadenfreude I felt at Mao’s plan going to shit from the start.
James S.A. Corey
Caliban's War
The Expanse series: Lots of POVs. Good personalization and relationship building.
Nov 2025
Lesswrong
"Other people are wrong" vs "I am right"
The critical distinction between critiquing others' views and believing your own position is correct.
Lesswrong
Embedded Agents
We currently lack a mathematical definition of what real-world intelligence even is.
Mark Xu
Three Things I've Learned About Bayes' Rule
Extraordinary claims require extraordinary evidence, but extraordinary evidence might be more common than you think.
Oct 2025
Nassim Nicholas Taleb
Antifragile: Things That Gain from Disorder
Society profits from the convexity of entrepreneurial risk; individuals bear the concavity.
Balajis
Popups are the new startups
How internet communities can temporarily meet in the physical world as popups, then evolve into permanent societies.
Craig Alanson
Specops
Overly anthropomorphized ASI with incompetent personality - Skippy misses obvious ideas despite billions of years of processing.
Sep 2025
Aug 2025
The Zvi
Childhood and Education #13: College
There’s a time and a place for everything. It used to be called college.
Lesswrong
Being the (Pareto) Best in the World
Circumvent the generalized efficient markets and reach the Pareto frontier through dimensionality.
Lesswrong
Staring into the abyss as a core life skill
You want to think reasonably about things that are uncomfortable to contemplate.
Lesswrong
Rest Days vs Recovery Days
Recovery day = binge Youtube, rest day = do what you want when you're motivated and able to get out and about.
Lesswrong
The Good News of Situationist Psychology
One of the most effective things you can do is to change your situation.
Lesswrong
Scientific Self-Help: The State of Our Knowledge
Scientifically-backed advice for having a better life.
Jul 2025
When Money Dies
Still reading.
From Third World to First
Still reading.
Dwarkesh
Why I don't think AGI is right around the corner
Continual learning is a bottleneck for AI to move out of self contained, short horizon tasks.
From The New World
From the New World [articles]
Short blogs on: Abundance, The Thiel-Andreessen Axis, How DC Missed Deepseek, Shanghai, Effective Altruism, Egalitarians, and Democratic Oligarchy.
Dwarkesh
AGI, debt crisis, & Politburo politics - Victor Shih
Maintaining Party control, prioritizing security vs economic reforms.
Dwarkesh
China’s Manufacturing Dominance: State Directives & Ruthless Competition - Arthur Kroeber
China prioritizing hardware tech (e.g., semiconductors, renewables) over demand-side reforms stifles services and consumption, risking deflation. [article]
Dwarkesh
China is digging out of a crisis. And America’s luck is wearing thin. - Ken Rogoff
China's stagnation stems from overinvestment in infrastructure and housing post-2010 stimulus. US will face debt-driven inflation crisis. [article]
Jun 2025
James Dale Davidson
The Sovereign Individual
Davidson and Rees-Mogg on the information age as a political-economic rupture: portable wealth, jurisdiction shopping, and weakened nation-state power.
Ray Dalio
How Countries Go Broke: The Big Cycle
Dalio’s template for sovereign debt crises: Big Debt Cycle mechanics, reserve-currency limits, and fiscal choices for controlling US-style debt problems.
May 2025
Ribbonfarm
The Gervais Principle
The Office a satire of real corporate dynamics- Sociopaths exploit overperforming Losers by promoting them into Clueless middle management to shield themselves.
Ezra F. VogelRecommended
Deng Xiaoping and the Transformation of China
Deng spent more than three years in exile at a tractor parts factory in Jiangxi. Vogel argues that is where he concluded something was wrong with the system itself. Back in power, he visited Japan and the US, then ran reform as experiments, like special economic zones, and scaled what worked. China's real GDP grew about 6.5x from 1978 to 1998.
Wandering Founder
The Art of Cold-emailing a Billionaire
Establish comptency, make an ask, be transparent with self intrests, show extra effort.
Acquired
The Complete History & Strategy of Microsoft 2
Steve Ballmer got way more shit from everyone than he deserved.
Acquired
The Complete History & Strategy of Microsoft
By recognizing Moore's Law would make computing universal, Microsoft used IBM's distribution to establish DOS as the standard before licensing it to every PC clone manufacturer.
Acquired
The Complete History & Strategy of Nvidia 2
NVIDIA bet everything on CUDA for six years with no market until AlexNet proved deep learning worked on GPUs, creating the $500B infrastructure powering all AI.
Acquired
The Complete History & Strategy of Nvidia 1
NVIDIA survived 90 graphics competitors by inventing programmable shaders- something which will power ML.
Acquired
The Complete History & Strategy of Ethereum
Ethereum made code and money identical-smart contracts with their own funds-creating a Raspberry Pi-speed world computer powering a $200 billion economy.
Acquired
The Complete History & Strategy of Enron
Enron collapsed by booking 20 years of future revenues as immediate profits through mark-to-market accounting while hiding losses off-balance-sheet, leaving no future value left to steal.
Berkeley AI Research
The Shift from Models to Compound AI Systems
AI is shifting from single large models to compound systems that combine multiple components (models, retrievers, tools) because these multi-part approaches achieve better results than just scaling up individual models, as seen in systems like AlphaCode 2 and AlphaGeometry.
Medium
The Future of AI Agents is Event-Driven
AI agents need event-driven architecture (EDA) rather than just better models to scale effectively, because they require real-time access to data, tools, and other agents across distributed systems-making.
Apr 2025
Pivotal
Minsky Moments in Venture Capital
Faster rounds mask risk, inflating a venture bubble that bursts when the pace slows.
Pivotal
Making Markets in Time
If Wall Street is in the business of spatial arbitrage, Silicon Valley is in the business of temporal arbitrage.
Wait But Why
Taming Mammoth
The hidden trap isn’t caring what others think-it’s outsourcing your identity to their presumed judgments.
Farnam Street
The Inner Scorecard
Most of our growth comes from internal measurement, while most problems stem from external comparison.
Pivotal
The Perils of Prudence
Abraham Thomas compares Amundsen and Scott at the South Pole. Startups, like polar journeys, get riskier the longer they take, so speed is the prudent choice. Amundsen had a third of Scott's funding, used dogs and skis, and still capped each day at six hours. Scott had courage and discipline but kept believing his plan after the ice disagreed.
Ntkris
Horizontal vs. vertical AI
AI apps: pick a single industry and nail it; horizontal plays belong to developer middleware, not end‑user products.
Pivotal
Data in the Age of AI
AI cheapens compute so radically that trusted, proprietary data becomes the bottleneck; the real upside now lies in owning, curating, and tooling that data.
Gwern
Commoditize Your Complement
Commoditizing a complement doesn’t just cheapen another layer-it creates a noisy facade of “competition” that diverts antitrust scrutiny while the real monopolist quietly tightens its grip on the profitable chokepoint.
Kwokchain
Making Uncommon Knowledge Common
The Rich Barton Playbook for winning markets through Data Content Loop.
Lesswrong
Is Driving Worth the Risk?
My friend Ragnor sent this to me after I got accidently nicked by a (highly unlikely rabied) cat.
Lesswrong
Recent AI model progress feels mostly like bullshit
Despite rising benchmark scores, recent AI model progress has yielded little practical improvement in real-world tasks.
Anthropic
Contextual Retrieval
Boosts RAG accuracy by prepending chunk-specific context to document embeddings. You'd want to merge semantic and lexical matching.
Arxiv
From Local to Global: A GraphRAG Approach to Query-Focused Summarization
Graph RAG = property graphs + traversal paths + explainable retrieval.
Mar 2025
Jordan B. Peterson
12 Rules for Life
Wisdom on how to live a good life. Points I agree, points I disagree. Could be verbose at times, but overall, was great for the plane ride.
David DeutschRecommended
The Beginning of Infinity: Explanations That Transform the World
"Feeling insignificant because the universe is large has exactly the same logic as feeling inadequate for not being a cow. Or a herd of cows. The universe is not there to overwhelm us; it is our home, and our resource. The bigger the better."
Feb 2025
Dwarkesh
China/US Super Intelligence Race, 2027 AGI, & The Return of History
Trillion dollar clusters, AGI will arrive by 2027-2028 through scaling and unhobbling existing models, the geopolitical implications, the importance of protecting AI research secrets, his criticisms of building AI infrastructure in Middle Eastern countries, his controversial departure from OpenAI, and his new AI investment firm.
Dwarkesh
How an Anonymous Researcher Predicted AI's Trajectory
Gwern Branwen, an anonymous internet researcher who was among the first to recognize AI scaling laws, discusses his writing process revolving around intellectual rabbit holes, his minimalist lifestyle living on $12K a year while influencing AI development through his essays, and his belief that AGI is just a few years away.
Dwarkesh
How the Semiconductor Industry Actually Works
Dylan Patel: Complexity of the semiconductor industry, China's potential to centralize compute resources, and how massive investment in AI hardware could enable training runs with 10,000x more compute than GPT-4 by the end of the decade.
Dwarkesh
How Mao Conquered China
Professor Sarah Paine: How Mao conquered China by combining propaganda, peasant mobilization, and warfare tactics.
Dwarkesh
Why Japan Lost
Professor Sarah Paine: How Japanese military culture, particularly the samurai code of Bushido, influenced their strategic decisions and battlefield conduct. Analyzes how Japanese values of honor, death, and willpower led to poor strategic choices, severe inter-service rivalry between the Army and Navy, and a refusal to surrender despite inevitable defeat.
Dwarkesh
The War For India
Professor Sarah Paine: Geopolitical dynamics between Russia, China, India, and Pakistan during the Cold War, exploring how their shifting alliances, territorial disputes, and interventions shaped modern South Asia.
Orion Taraban
The Value of Others: Understanding the Economic Model of Relationships
A clinical psychologist uses a (controversial(?) manosphere/red pill content) economic model to explain decision making in relationships.
Amanda Askell
The optimal rate of failure
The optimal rate of failure is rarely zero, and if you never fail, you're likely being too risk-averse, which can be costly both personally and ethically.
Angle Money
An autonomous and modular price stability module for decentralized stablecoin protocols
1/3 whitepaper from angle.money. Technicals on how they manage a basket of assets with exposure limits and handle black swan events without requiring governance intervention.
Lex Fridman
DeepSeek, China, OpenAI, NVIDIA, xAI, TSMC, Stargate, and AI Megaclusters | Dylan Patel and Nathan Lambert
AI competition between the U.S. and China parallels Cold War dynamics, with Chinese teams like R1 rapidly deploying models optimized for memory-efficient reasoning while American export controls and manufacturing limitations may prove ineffective as the bottlenecks shift from chip capabilities to engineering talent and energy infrastructure.
Google Research
Towards an AI co-scientist
Multi-agent AI system built on Gemini 2.0 that generate novel, testable research hypotheses, as validated through experimental confirmation of its predictions in drug repurposing for acute myeloid leukemia. The elo system for eval is the same as the one that gens the hypotheses which I would intuit cause a ircular reasoning problem.
David DeutschRecommended
The Beginning of Infinity: Explanations That Transform the World
Hard to vary, via the seasons. In the Persephone myth, winter comes when she goes back to Hades each year. Deutsch points out that a different god and a different deal would explain winter just as well. With axial tilt you can't swap details like that, and it predicts opposite seasons in the two hemispheres. The myth could be changed to fit that result or its opposite, and that ease of adjustment is its weakness.
80000 Hours
On average, you spend 80k hours in your career. This book is a guide to choosing a career with high social impact: strategic career choices can significantly improve the world. Some Effective Altruism arguments.
Dennis E. Taylor
Heaven's River
A Von Neumann probe's search for a massive megastructure built by an alien civilization, while dealing with conflicts within AI clones of himself.
Dennis E. Taylor
The Bobiverse 1-3
Software engineer who dies and is resurrected as an AI in a von Neumann probe, leading to a galaxy-spanning adventure of exploration, self-replication, and existential dilemmas. Becoming a become a Von Neumann probe added to my bucket list.
Story Protocol
A Peer-to-Peer Intellectual Property Network
Blockchain network for registering, tracking, and monetizing IP where users can set usage terms and receive royalties as their work is used across applications and AI systems.
Vitalik
d/acc: one year later
Reflection on philosophy of 1) accelerate technology 2) prioritizing defensive capabilities and decentralization of power so people maintain maintain agency and freedom even as technological progress.
Vitalik
My techno-optimism
Vitalik argues for defensive acceleration (d/acc) approach to technological development that prioritizes technologies that enhance defense, decentralization, and democracy while being more cautious about potentially dangerous technologies like AI, contrasting this with both anti-tech views and unrestricted technological acceleration.
Matt Dinniman
Dungeon Crawler Carl #7
Fornite battle royale with alien politicians held hostage by an AI system losing its shit. There's a fine line between gritty determination and melodramatic edgelord monologues, more and more of this series is crawling (get what I did here?) towards the latter.
Matt Dinniman
Dungeon Crawler Carl #5
Series has been going downhill- but I'm still invested in the story. There's dinosaurs in this book.
Bitcoin Magazine
I don't support a Strategic Bitcoin Reserve, and neither should you
Personal comment: Stablecoins have created a $160B market-driven system that strengthens USD by extending its reach globally and serving as dollar reserves in emerging markets - achieving currency dominance possibly more effectively than proposed government Bitcoin purchases.
MediumMP3
Marc Andreessen and the CFPB
Response to Marc's claim of Consumer Financial Protection Bureau (CFPB) systematically debanking crypto entrepreneurs and politically disfavored individuals, sparking debate about whether these are accurate and what role the CFPB actually plays in addressing debanking.
Matt Dinniman
Dungeon Crawler Carl #4
A floating fortress occupied by warrior gnomes. A castle made of sand. A derelict submarine guarded by malfunctioning machines. A haunted crypt.
Matt Dinniman
Dungeon Crawler Carl #3
I'm binge-reading the series now. This level takes place in an impossibly-complicated subway system built out of the world's subterranean railway systems.
Jan 2025
Wait But Why
Why Generation Y Yuppies Are Unhappy
Gen Y young professionals are unhappy because their reality can't match their inflated expectations, which were shaped by their parents' optimism, being told they were special, and constant exposure to peers' curated social media success.
Matt Dinniman
Dungeon Crawler Carl #2
Carl and cat descends deeper into more dangerous floors where they face off against a murderous groups of people.
Matt Dinniman
Dungeon Crawler Carl #1
Carl and his ex's talking cat are forced into a Fortnite battle royal style dungeon-crawler gameshow broadcast across the galaxy.
The Courage to be Disliked
Adlerian psychology: free you from seeking others' approval, helping you take charge of your own life by focusing on what you can control while accepting that not everyone will like your choices
Viktor E. FranklRecommended
Man's Search for Meaning
"Everything can be taken from a man but one thing: the last of the human freedoms, to choose one's attitude in any given set of circumstances, to choose one's own way."
Github
Deepseek R1 Paper
They didn't even use Cuda- rather parallel thread execution. The Chinese Quants are cracked. Anyways key take aways: Achieved strong performance with just 33B parameters and $5M training cost through an efficient reinforcement learning approach that skipped supervised fine-tuning and used group-based optimization instead of a critic model. Focused on deterministic tasks (math, coding) using rule-based rewards rather than expensive neural reward models or human annotation, significantly reducing costs. Rather than relying on massive scale, they achieved efficiency through iterative fine-tuning, high-quality data curation, and model distillation to create smaller versions (1.5B-14B parameters) while maintaining reasoning capabilities.
What's Our Problem?
Still reading: a self-help book for societies.
Wait But Why
The Tail End
Life is finite, and when you measure it in discrete units - whether days, activities, or relationships - you realize just how limited our remaining experiences truly are.
Wait But WhyMP3
How to Pick a Career (That Actually Fits You)
Guide that reframes career decisions from rigid, high-stakes tunnel choices into a series of experimental dots to be connected over time, emphasizing the importance of understanding both your authentic desires and the reality of what's possible while being willing to adapt as you and the world change.
Nassim Nicholas Taleb
The Bed of Procrustes
Collection of philosophical aphorisms by Nassim Nicholas Taleb that examines human nature and modern life.
The Laws of Trading
Re-reading book my prev manager wrote. Essential principles covering risk management, strategy clarity, technological advantage, and continuous improvement in the context of decision making.
Dec 2024
Dario AmodeiMP3
Machines of Loving Grace
Exploration of how AI could improve human well-being across health, neuroscience, poverty, governance, and meaning in the next 5-10 years, arguing that with careful development and global cooperation, AI can usher in an era of unprecedented progress.
MediumMP3
Five perspectives on stablecoins
Analysis of the stablecoin landscape, examining diverse viewpoints from policymakers and academics. Stablecoins are a transformative financial technology with significant implications for the dollar, global payments, and individual liberty.
Hbr
How Digital Currencies Can Help Small Businesses
Stablecoins offer small businesses lower transaction costs, faster access to funds, and more predictable cash flow by enabling real-time domestic and cross-border payments while reducing reliance on expensive intermediaries.
Vitalik
From prediction markets to info finance
Prediction markets, like Polymarket, are not just for betting on elections-they're tools for gathering information by aligning incentives. This idea, called 'info finance,' uses markets to extract useful predictions and insights, with applications in governance, science, and public goods funding, and it's now practical thanks to blockchain and AI.
A16z
A few of the things we're {a16z} excited about in crypto (2025)
The most intresting thing I'm thinking about here: Unlike traditional game AIs with straightforward victory conditions, modern AI agents often have ambiguous success metrics, but combining real-world economic incentives (e.g., increasing a crypto wallet's balance) with the ability to visually interact with interfaces via an API moves closer toward more tangible and practical win conditions. Web3 companies will begin with the end (-user experience), instead of letting the infrastructure determine the UX. As Nassim Taleb observed, 'Over-engineering breeds fragility. Simplicity scales.'
Decentralized Thoughts
Dolev-Strong Authenticated Broadcast
How distributed systems reliable communicate with malicious actors using cryptographic signatures. [Technical]
THE ANTHOLOGY OF BALAJI
Still reading
THE ALMANACK OF NAVAL RAVIKANT
I was born poor and miserable. I'm now pretty well-off, and I'm very happy. I worked at those. I've learned a few things, and some principles. I try to lay them out in a timeless manner, where you can figure it out for yourself. Because at the end of the day, I can't quite teach anything. I can only inspire you and maybe give you a few hooks so you can remember.
Lex Fridman
Anthropic CEO on Claude, AGI & the Future of AI & Humanity | Lex Fridman Podcast #452
Scaling laws, AI Safety Levels, AGI timeline, meaning of life, programming advice for non-technical people.
Dwarkesh
Will Scaling Work
Scaling large language models can lead to AGI, debating data limitations, generalization, and synthetic data/self-play: 70% likelihood of transformative AI by 2040 if scaling succeeds.
Joe Rogan
#2234 - Marc Andreessen, The Joe Rogan Experience
Pretty political podcast on Trump, Crypto, DOGE, and startups.
Nov 2024
Nfx
The Business Case for Longevity
Companies are constantly calculating the lifetime value of a customer. We can increase the L. The longer someone lives, the greater your LTV will be, no matter what.
Andrew Rea
On Commitment & Optionality
Lots of smart people go into VC, investment banking, and product driven by 1) status 2) lack of conviction from endless optionality.
Katiewav
hell is the absence of taste
Developing good taste requires intentional engagement, balancing instinct and aspiration, to resist the flattening effect of algorithmic curation and cultivate deeper meaning and joy in a world of overabundant, passive consumption. Reminds me of the PG article on good taste.
Gonzo Circus
Time-wars
'Today, however, capital seeks to exploit our sociality in all spheres of work. When we all become "human capital" we not only have a job, or perform a job. We are the job.'
A16z
16 insights from Read Write Own
Looking for a better understanding of the tech, business, and startup worlds? A primer on the history and future of the internet? A crisp explanation of the crypto and blockchain movements, and why they matter?
Chris Dixon
Pricing Curves
How companies use demand curves and proxies for willingness-to-pay to maximize revenue-sometimes in ways that defy logic.
Chris Dixon
Price Bundling
Bundling flattens demand curves, allowing sellers to maximize revenue while benefiting buyers with lower prices and more surplus, making it a win-win strategy for taste-based goods like media.
A16z
Techno-Optimistism
Marc Andreessen champions technology as humanity's greatest tool for progress, abundance, and freedom, rallying against stagnation, bureaucracy, and pessimism.
Khosla
AI Dystopia | Utopia
Vinod Khosla envisions AI as humanity's ultimate tool for abundance, reshaping work, creativity, and governance while emphasizing the need to outpace China.