We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Recommended Reading from Tech Founders
Active
Public
Daily at
7:00 AM
Agent time: 8:00 AM GMT+01:00 – Europe / London
by avergin 137 sources
Tracks and curates reading recommendations from prominent tech founders and investors across podcasts, interviews, and social media
Andreessen highlights California Forever’s city‑building plan
22 October 2025
•
1 minute read
Marc Andreessen 🇺🇸
One standout, high-signal pick: Marc Andreessen highlights an All-In Pod clip on California Forever’s plan to build a new American city, featuring Jan Sramek, with clear reasons it matters.
Top pick
- Title: California Forever: The Startup Building America’s Next Great City
- Content type: Video clip
- Author/creator: theallinpod (X post)
- Link/URL:https://x.com/theallinpod/status/1980726875849736518
- Recommended by: Marc Andreessen
"It’s time to m—–f—— build! 🇺🇸"
- What it covers:
- (0:00) Introducing Jan Sramek
- (0:51) How California Forever is building America’s next great city
- Featuring: Jan Sramek (@jansramek), CAForever (@CAForever)
- Why it matters: Concise look at a live attempt to build a new American city—useful context for the current push to “build” ambitious projects
AI’s Next Constraint Is Energy; Craft Insights and China’s Gold Signals
21 October 2025
•
4 minutes read
Balaji
Patrick OShaughnessy
Tim Ferriss
Today’s high-signal picks center on AI’s shift to an energy bottleneck, practical guidance for building AI agents, craft lessons for better writing, and macro signals from China’s gold market—curated from authentic recommendations by founders and investors.
Top pick: China’s power build‑out and AI’s looming energy bottleneck
- Title: China’s Energy Scale and AI’s Next Constraint (Dan Wang clip)
- Content type: Video clip
- Author/creator: Dan Wang (speaker)
- Link/URL: https://video.twimg.com/amplify_video/1980281178067189760/vid/avc1/1080x1080/VFO3V1xeBf3LD_-g.mp4?tag=21
- Recommended by: Patrick O’Shaughnessy — “Dan Wang explains how China is already scaling electricity production far faster than the U.S.”
-
Key takeaway: “Elon’s been saying this for years, and he’s right: the AI scaling constraint is shifting from chips → transformers → energy.”
“China is adding 500 GW of solar capacity v. 50GW in the U.S. 33 nuclear plants under construction in China, and zero in the U.S.”
- Why it matters: If energy becomes the binding constraint for AI, China’s rapid electricity capacity expansion is a critical comparative advantage for training and serving large AI systems .
AI agents: reliability, human‑in‑the‑loop, and incremental value
- Title: Reflections on Karpathy’s podcast (AI agents in enterprise)
- Content type: X post
- Author/creator: Aaron Levie (@levie)
-
Link/URL: https://x.com/levie/status/1980435175160050092
Context link: https://x.com/samsja19/status/1980251883366477869 - Recommended by: Aaron Levie — “Anyone who has been working on AI agents, especially in an enterprise context, would have been nodding along to Karpathy’s podcast the entire time.”
-
Key takeaways:
“So much has yet to get built out to make it all work at scale…each additional 9 of reliability takes a massive amount of work. 90%, 99%, 99.9% are all new levels of complexity…” “The one benefit of AI agents for knowledge work…is that you can decide to get value at any point along the way by just tweaking what the human in the loop does.” “Self driving cars working safely…is nearly a binary event, but choosing to have AI do a 1 minute task, then 10 minute task, then 100 minute task…adds incremental value all along the way.”
- Why it matters: Enterprise AI adoption can progress today by scoping tasks and using human oversight, while the ecosystem works toward higher “nines” of reliability .
Craft and learning
Title: David Remnick on “what makes a great profile”
Content type: Article/essay (referenced)
Author/creator: David Remnick
Link/URL: https://x.com/patrick_oshag/status/1980449210400125221
Recommended by: Patrick O’Shaughnessy — “Re the writing role, I think about this a lot: David Remnick on what makes a great profile”
Key takeaway: A touchstone for the writing craft, especially for profiling leaders and companies .
Why it matters: High‑leverage guidance for founders, investors, and operators who communicate through narrative profiles .
Title: Jiro Dreams of Sushi
Content type: Documentary film
Author/creator: (not specified in source)
Link/URL: —
Recommended by: Tim Ferriss (as an exemplar for studying excellence)
Key takeaway:
“If you want to be good at anything, study people who are excellent at something…It does not have to be the same thing you are hoping to pursue.”
Why it matters: Studying top practitioners in any domain yields transferable lessons to accelerate mastery in your own field .
Macro and markets
Title: China is remonetizing gold (context and data)
Content type: X posts (chart, definitions, analysis)
Authors/creators: @biancoresearch; @AndreasSteno
Links/URLs: https://x.com/biancoresearch/status/1980101218211922270, https://x.com/AndreasSteno/status/1970938624855375884
Recommended by: Balaji Srinivasan — “See also @biancoresearch and @AndreasSteno.”
Key takeaways:
“Gold Warrants on the Shanghai futures exchange have nearly doubled since this chart posted a month ago. Up 25x since the beginning of the year.”
- What a gold warrant is and how it’s used for SHFE physical delivery; standard warrant ≈ 3,000g .
- Drivers: record safe‑haven demand, robust central‑bank buying (notably PBoC), and volatility‑driven investor/arbitrage activity .
- Framing from Balaji: “China is remonetizing gold.”
Why it matters: A fast‑rising gold‑backed market signal in China has implications for liquidity preference, reserves strategy, and macro hedging .
Title: The Jewish State
Content type: Book
Author/creator: Theodor Herzl
Link/URL: —
Recommended by: Balaji Srinivasan
Key takeaways:
“If you go back and read The Jewish State…Herzl was like a startup entrepreneur.”
- Balaji notes Herzl organized funding; the descendant, the Jewish National Fund, “owns a big chunk of Israel.”
Why it matters: An entrepreneurial lens on institution‑building and long‑term mobilization offers durable lessons for founders and policy thinkers .
Title: The Social Network
Content type: Film
Author/creator: (Aaron Sorkin referenced by the recommender)
Link/URL: —
Recommended by: Balaji Srinivasan
Key takeaways:
“It’s a great movie…Inspired a whole generation of entrepreneurs.” “If that wasn’t Sorkin’s intent, it was his result.”
Why it matters: A compelling narrative that has shaped entrepreneurial ambition and cultural understanding of tech founding .
AI agents + SaaS, developer productivity picks, and a sci‑fi design experiment
20 October 2025
•
4 minutes read
Lenny's Newsletter
Lenny's Podcast
Amjad Masad
A concise set of organic recommendations: Aaron Levie flags a must-read thread on AI agents + SaaS; Nicole Forsgren shares practical books and tools for measuring dev productivity; and Amjad Masad highlights a striking sci‑fi UI experiment.
Below are today’s highest-signal, organic picks from trusted founders/operators. Each entry includes what it is, who surfaced it, the core takeaway, and why it matters.
Featured pick
- Title: The State of AI in SaaS: SaaS is Dead, Long Live SaaS
- Content type: X thread / analysis
- Author/creator: Spencer Skates
- Link/URL: https://x.com/spenserskates/status/1979890627136688177
- Recommended by: Aaron Levie (CEO, Box)
-
Their key takeaway:
- “Bull case for AI agents + SaaS”
- Models show strong general reasoning but still need context (your data, relevant UIs, and domain grounding) to be truly useful
- Many of the interfaces to feed that context will need to be built from scratch
- Developers have a natural agent workflow via IDEs; most categories lack that analog today
- This is a “trillion dollar opportunity” for new startups or incumbents who pivot fast
- Why it matters: It frames where value may accrue as AI agents integrate with real workflows and data—highlighting the missing infrastructure and UI layers still to be built . As the thread’s author puts it, understanding this shift will determine outcomes:
“Those who deeply understand this will make all the returns and everyone else will lose a lot of money.”
Developer productivity: Nicole Forsgren’s book and tool picks (via Lenny’s Podcast)
Context: Forsgren created the widely used DORA and SPACE frameworks, wrote Accelerate, and is releasing Frictionless; she’s Senior Director of Developer Intelligence at Google .
Title: Outlive: The Science and Art of Longevity
- Content type: Book
- Author/creator: Peter Attia
- Link/URL: https://www.amazon.com/Outlive-Longevity-Peter-Attia-MD/dp/0593236599
- Recommended by: Nicole Forsgren (on Lenny’s Podcast)
- Their key takeaway: “Outlive by Peter Attia is fantastic.”
- Why it matters: Highlights a longevity resource Forsgren finds valuable, reflecting the importance of sustainable performance for builders.
Title: Back Mechanic
- Content type: Book
- Author/creator: Stuart McGill
- Link/URL: https://www.amazon.com/Back-Mechanic-Stuart-McGill-2015-09-30/dp/B01FKSGJYC
- Recommended by: Nicole Forsgren (on Lenny’s Podcast)
- Their key takeaway: A layperson’s guide to diagnosing and fixing lower-back problems
- Why it matters: Practical health guidance for knowledge workers who spend long hours at a desk.
Title: How Big Things Get Done
- Content type: Book
- Author/creator: Not specified in source
- Link/URL: https://www.amazon.com/How-Big-Things-Get-Done/dp/0593239512/
- Recommended by: Nicole Forsgren (on Lenny’s Podcast)
- Their key takeaway: Dissects large projects across recent history—where they failed and why—useful as AI-era software systems change
- Why it matters: A lens for planning and executing complex initiatives as teams replatform workflows with AI .
Title: The Undoing Project
- Content type: Book
- Author/creator: Michael Lewis
- Link/URL: https://www.amazon.com/dp/B01KBM82M4/
- Recommended by: Nicole Forsgren (on Lenny’s Podcast)
- Their key takeaway: Strong praise; “audibly gasped” at the final sentence
- Why it matters: Behavioral insights that inform decision-making—relevant to product, engineering, and leadership.
Title: Claude Code
- Content type: Tool
- Author/creator: Not specified in source
- Link/URL: https://www.claude.com/product/claude-code
- Recommended by: Dan Shipper (called “the most underrated” on Lenny’s Podcast)
- Their key takeaway: Beyond coding, it can handle practical local tasks—for example, finding files to help clean up storage on your laptop
- Why it matters: Illustrates how agentic coding tools can execute real computer actions, expanding usefulness beyond code generation .
Related insights from the episode:
- OpenAI Codex has been reported to find stubborn bugs when given time to reason, per an anecdote attributed to Karpathy .
- Rapid visual iteration: using ChatGPT and Gemini to transform a room’s look from a floor plan and photos can be quick and helpful for visualization .
Design and code: a sci‑fi UI experiment worth trying
- Title: Galaxy Radar (interactive visual design experiment)
- Content type: Web app / interactive demo
- Author/creator: @levinstanley
- Link/URL: https://galaxy-radar.replit.app
- Recommended by: Amjad Masad
- Their key takeaway: Built on Replit, inspired by sci‑fi UI; designed to test how far design and code can merge into something beautiful; a galactic radar visualization with simulated data
- Why it matters: A compact example of high-quality UI/UX prototyping that blends aesthetics and engineering—useful inspiration for teams building agent front-ends and real-time visualizations.
Pragmatic AI and time‑literacy: Karpathy’s thread, Dwarkesh interview, and Andrew Chen’s reading picks
19 October 2025
•
3 minutes read
jack
andrew chen
Aaron Levie
Enterprise leaders flagged two high-signal AI resources—Karpathy’s pragmatic thread on timelines, RL, and agents, plus the Dwarkesh interview—alongside Andrew Chen’s history + sci‑fi picks to build time‑literacy and strategic imagination.
Most compelling today
Karpathy’s post‑podcast notes on AGI timelines, RL, and LLM agents (X thread)
- Title: Karpathy’s post‑podcast notes on AGI timelines, RL, and LLM agents
- Content type: X thread
- Author/creator: Andrej Karpathy
- Link/URL: https://x.com/karpathy/status/1979644538185752935
- Recommended by: Aaron Levie
"This is actually extremely pragmatic and realistic from @karpathy based on what is likely to happen, especially in an enterprise context."
- Why it matters: Clear-eyed guidance for deploying AI in real organizations—tempered timelines, emphasis on integration/safety, and a practical stance on agents vs. RL .
-
Key points:
- Timelines: More conservative than mainstream hype; a “decade of agents” with substantial remaining work (integration, sensors/actuators, safety/security) even amid rapid progress .
- Learning paradigms: Critique of RL’s poor signal-to-compute efficiency; optimism for alternatives (e.g., agentic interaction, “system prompt learning”) while noting the gap between arXiv ideas and scalable lab implementations .
- Agents today: Prefer collaborative, verifiable workflows over fully autonomous code dumps; insist on proofs, API docs, and tighter human-in-the-loop to avoid low-quality output and security risks .
- Cognitive core: Reduce memorization to improve generalization; models may need to get larger before they get smaller .
Also worth your time
The @karpathy interview — Dwarkesh Podcast (episode)
- Title: The @karpathy interview
- Content type: Podcast episode (video/audio)
- Author/creator: Dwarkesh Podcast; Guest: Andrej Karpathy
- Link/URL: Episode post on X — https://x.com/dwarkesh_sp/status/1979234976777539987
- Availability: YouTube, Apple Podcasts, Spotify
- Recommended by: Jack ("this is great")
- Why it matters: A broad, accessible tour of AGI timelines, LLM cognitive limits, the role and limits of RL, and implications for industries like self‑driving and education .
-
Key topics (timestamps):
- 0:00 AGI is still a decade away; 0:30 LLM cognitive deficits; 0:40 Why RL struggles; 1:07 AGI blending into GDP growth; 1:43 Why self‑driving took so long; 1:57 Future of education .
Andrew Chen’s history + sci‑fi picks for strategic imagination (reading list)
"I’ve always thought reading history and reading sci‑fi go hand in hand."
- Content type: Curated reading list (via X thread)
- Author/creator: Andrew Chen
- Link/URL: https://x.com/andrewchen/status/1979666433853579748
- Recommender’s key takeaway: History = pattern recognition; Sci‑fi = scenario generation; together they cultivate strategic imagination and time‑literacy .
- Why it matters: Pairs realism with imagination, warns against hubris, and helps leaders build foresight for technology and policy decisions .
-
Picks (all via Andrew Chen’s thread):
- Foundation (1951) — Book — Isaac Asimov; “future history” and predictive governance
- Neuromancer (1984) — Book — William Gibson; digital consciousness
- Dr. Strangelove (1964) — Film — Stanley Kubrick; satirical exploration of nuclear annihilation
- Heinlein’s future‑history stories (1940s–1950s) — Fiction — Robert Heinlein; chronological histories of the future
- The First Men in the Moon (1901) — Book — H. G. Wells; pulp sci‑fi that inspired the Space Race
- 1984 (1949) — Book — George Orwell; how progress can curdle into control
- Brave New World (1932) — Book — Aldous Huxley; how progress can curdle into control
Energy costs shaping AI, rare‑earths strategy, and the power of high standards
18 October 2025
•
5 minutes read
Patrick OShaughnessy
Elad Gil
All-In Podcast
Authentic picks centered on energy’s role in AI competitiveness, rare‑earths and policy, durable operating principles, a standout founder profile, and a practical tool to speed product iteration—curated from trusted operators and investors.
Top pick: Energy markets, solar, grids, and AI compute economics (conversation with Zach Dell, Base Power)
- Title: Energy markets, surprisingly fast solar adoption, batteries/next‑gen distributed grids, and whether regional energy costs will drive future AI dominance
- Content type: Video (X clip)
- Author/creator: Guest ZachBDell; company: BasePower
- Link/URL: https://video.twimg.com/amplify_video/1979213243294887936/vid/avc1/1080x1920/RCAn7CtnzfIUf6GC.mp4?tag=21
- Recommended by: Elad Gil (“Great conversation…”)
- Key takeaway: Frames energy as an input to everything, covering adoption speed of solar, storage, distributed grids, and how regional energy cost differentials could shape AI dominance
- Why it matters: It centers the real constraint for AI scale—energy cost and availability—and surfaces the question of regional competitive advantage for compute
Why this is the top pick: It directly links hard‑cost inputs (energy) to AI competitiveness, a first‑order variable for builders and investors planning infrastructure, siting, and capital allocation .
AI, energy, and policy crosscurrents
-
Title: All‑In discussion of SF public safety, rare‑earths/US‑China trade, and AI’s PR crisis (datacenter cancellations)
- Content type: Podcast episode / video
- Author/creator: All‑In Podcast
- Links/URL: Episode post https://x.com/theallinpod/status/1979351675040039239; Clip https://video.twimg.com/amplify_video/1979344189969080320/vid/avc1/1920x1080/EegnQmwwJ5kdjy9K.mp4?tag=21
- Recommended by: Chamath Palihapitiya (“Pod up!”)
- Key topics: US‑China trade tensions and price floors ; on‑shoring a full‑stack rare‑earth industry ; SF crime and National Guard debate ; AI’s PR crisis with datacenters denied .
- Why it matters: One conversation spans policy, supply chains, and AI infrastructure—issues that directly affect founder operating conditions and investor theses .
-
Title: Wikipedia page on “nuclear” (risk and safety)
- Content type: Reference article
- Author/creator: Wikipedia
- Link/URL: (not provided)
- Recommended by: Elad Gil (frequent reading)
- Key takeaway: Direct evidence of deaths from nuclear power is very small; most clearly provable fatalities are from 1980s Russian submarine incidents and radiology misdoses; more people die annually from solar installation falls than from nuclear accidents; overall, nuclear is “very safe” .
- Why it matters: Calibrates risk perception with comparative data—useful context for energy strategy and policy discussions .
Operating principles and mental models
Title: “The key to happiness is high standards” (clip)
- Content type: Video clip (X)
- Author/creator: Posted by Patrick O’Shaughnessy
- Link/URL: https://x.com/patrick_oshag/status/1960451111862558938
- Recommended by: Patrick O’Shaughnessy (“I love how this clip just keeps making the rounds”)
-
Key takeaway:
“Happiness is a trailing indicator of impact” (as @eldsjal says), and impact often results from high standards . He adds it “continues being true as you get older” .
- Why it matters: A durable operating heuristic for teams and leaders: raise standards to compound impact over time .
Title: Model‑scaling papers (GPT‑2/3 era)
- Content type: Research papers
- Author/creator: Various (not specified)
- Link/URL: (not provided)
- Recommended by: Elad Gil (he experimented with GPT‑2/3, saw a step‑function up, and could “just read” the scaling papers to see what was coming) .
- Key takeaway: Early, close reading plus hands‑on experimentation helped Elad anticipate the wave and look for people building in the space—leading to investments in Perplexity, Harvey, Decagon, and Abridge .
- Why it matters: Systematic engagement with foundational research can inform timely product and investment moves .
Title: The Beginning of Infinity (book)
- Content type: Book
- Author/creator: David Deutsch
- Link/URL: (not provided)
- Recommended by: The host (speaker A) in a conversation with Ryan Hoover
- Key takeaway: Frames innovation as effectively infinite; top value‑add investors avoid scarcity mindsets, give freely, and see compounding reciprocity .
- Why it matters: Useful lens for building networks and adding value over long horizons .
Founder profiles and writing craft
-
Title: Profile of Joshua Kushner (Join Colossus)
- Content type: Article
- Author/creator: Jeremy Stern (@JeremySternLA)
- Link/URL: https://joincolossus.com/article/joshua-kushner-thrive-new-world/
- Recommended by: Morgan Housel
-
Key takeaway:
“This is so well done and @JeremySternLA is the best profile writer alive.”
- Why it matters: Deep, well‑crafted profiles help founders and investors study decision quality, time horizons, and firm‑building patterns .
Tools to speed build cycles
-
Title: Alloy (web app for in‑browser design changes)
- Content type: Tool (web app)
- Author/creator: (not specified)
- Link/URL: http://alloy.app
- Recommended by: Jason Lemkin (“One of the coolest vibe apps I’ve used so far”)
- Key takeaway: You can describe desired changes to a page he “vibe coded” in Replit; Alloy applies the design changes right in the browser; he then sends a screenshot to Replie—“No more drama explaining what I want changed, and why!”
- Why it matters: Shrinks the iteration loop between product intuition and implementation by making UI changes immediate .
Cultural lenses for builders
Title: Eddington (film)
- Content type: Film
- Author/creator: Ari Astor (auteur)
- Link/URL: (not provided)
- Recommended by: Marc Andreessen
- Key takeaway: Described as a “Capital A art” film that uniquely sets COVID, the George Floyd riots, social media, Wokeness, and Trump inside the movie’s world—and shows online/offline interplay affecting behavior .
- Why it matters: A rare contemporary work grappling directly with recent societal shocks—useful cultural context for navigating today’s narratives .
Title: The Critical Drinker (YouTube channel)
- Content type: YouTube channel
- Author/creator: (not specified)
- Link/URL: (not provided)
- Recommended by: Marc Andreessen
- Key takeaway: A scathing, funny chronicle of the trend he calls “the Message” in Hollywood .
- Why it matters: Helps decode entertainment industry signals that shape audience expectations and cultural discourse .
Founders’ picks: AI scaling, US–China strategy, and design with taste
17 October 2025
•
5 minutes read
Patrick OShaughnessy
Lenny's Newsletter
Patrick Collison
A concise, authenticity‑filtered set of founder and investor recommendations spanning AI’s scaling laws and foundations, US–China strategy, design taste as a moat, and execution. Top pick: Dan Wang’s Breakneck, with direct implications for AI, national security, and investing.
Most compelling pick
-
Breakneck — Book by Dan Wang (link: conversation with Dan on Invest Like the Best: https://www.youtube.com/watch?v=DSY68wBQ97I)
- Content type: Book; Related conversation (video)
- Author/creator: Dan Wang
- Recommended by: Patrick O’Shaughnessy, who calls it “the best framework I’ve seen for understanding US–China competition.”
-
Key takeaway:
“By most metrics, Chinese scientists are producing better qualities of work, not just quantity. There’s a more stable environment for them to do their science.” “And conversely, we have not seen that the US is really learning the hard parts of manufacturing really well. The Chinese are learning at a much faster rate to patch up their deficiencies than the Americans have.”
- Why it matters: Frames the core dynamic shaping AI, national security, and investing decisions, and why rebuilding U.S. manufacturing may be harder than China improving invention .
AI: scaling and foundations
Scaling AI (Stripe Press; oral history) — Dwarkesh Patel
- Content type: Oral history/book
- Author/creator: Dwarkesh Patel (Stripe Press)
- Recommended by: Rory (guest on 20VC), who read it and said it’s “really good”
-
Key takeaways:
“The scaling law has been proven to hold for six, seven years now at a high degree of accuracy.” Some leading practitioners matter‑of‑factly accept investing roughly “1% of GDP” in compute as the to‑do list to reach AGI .
- Why it matters: Captures how top teams think about compute scaling and why they treat progress as a predictable engineering path .
- Link: 20VC episode context (YouTube): https://youtu.be/PZAstljD60c
Attention Is All You Need — Research paper (2017)
- Content type: Paper
- Author/creator: Vaswani et al.
- Recommended by: Sundar Pichai, who calls it “extraordinary” and “the foundation of everything since then” in GenAI .
- Key takeaway: Google adopted Transformers immediately; models like BERT and MUM dramatically improved Search and enabled AI‑native products such as Google Photos .
- Why it matters: Core reading to understand modern AI systems and why scaling Transformers unlocked step‑changes in applied performance .
Google AI cracks new cancer code — Article (Decrypt)
- Content type: Article
- Author/creator: Decrypt
- Recommended by: Vinod Khosla (“Google as scientist benefiting humanity… creative with new ideas for cancer”) .
- Why it matters: Illustrates how frontier AI is beginning to shape biomedical discovery and therapeutic approaches .
- Link: https://decrypt.co/344454/google-ai-cracks-new-cancer-code
Design and craft (as a moat)
“If you want your generations to feel designed and not…generated” — Video demo by @bnj
- Content type: Video (tweet)
- Author/creator: @bnj
- Recommended by: Garry Tan — “What might it look like for an LLM to be able to do codegen with taste? This” .
- Key takeaway: Demonstrates LLM codegen with taste—outputs that feel designed, not merely generated .
- Why it matters: Shows what a design‑forward AI tooling experience can look like; the video also hides invite codes to @variantui .
- Link: https://x.com/bnj/status/1978884060077580454
Understanding Comics: The Invisible Art — Book by Scott McCloud
- Content type: Book
- Author/creator: Scott McCloud
- Recommended by: Dylan Field — “almost like an HCI book… a great way to explore how people perceive… wonderful in how it deals with abstraction” .
- Why it matters: A practical lens on perception and abstraction that applies directly to product and UX work .
- Link: https://www.amazon.com/Understanding-Comics-Invisible-Scott-McCloud/dp/006097625X
Codex Seraphinianus — Book by Luigi Serafini
- Content type: Art/illustrated encyclopedia
- Author/creator: Luigi Serafini
- Recommended by: Dylan Field; described as “an encyclopedia of another world” with an invented script that sparks exploration and interpretation .
- Why it matters: Trains imagination and comfort with ambiguity—useful for open‑ended design and concept work .
- Link: https://www.amazon.com/Codex-Seraphinianus-Anniversary-Luigi-Serafini/dp/0847871045
Strategy and macro
Seeing Like a State — Book by James C. Scott
- Content type: Book
- Author/creator: James C. Scott
- Recommended by: Eugene Wei (gifted to Patrick O’Shaughnessy) .
- Key takeaway: Top‑down systems underperform over long horizons relative to bottom‑up approaches .
- Why it matters: Offers a durable mental model for evaluating centralized vs. decentralized strategies—relevant to governance, markets, and technology adoption .
1929: Inside the Greatest Crash in Wall Street History — Book by Andrew Ross Sorkin; related interview
- Content type: Book; Interview (video)
- Author/creator: Andrew Ross Sorkin
- Recommended by: Chamath Palihapitiya (“Watch the interview…buy the book! I bought it and am excited to read it.”) .
- Key topics in the interview: Parallels to 2025, AI’s potential unemployment impact, and whether the US needs a “New Deal” on spending .
- Why it matters: Historical context for bubble dynamics and policy responses can sharpen judgment in today’s AI‑driven market environment .
- Link (interview): https://x.com/theallinpod/status/1978747752273297510
Operating and iteration
Michael Dell on iteration — Short video (posted by David Senra; recommended by Brad Gerstner)
- Content type: Short video clip
- Author/creator: David Senra (clip); Speaker: Michael Dell .
- Recommended by: Brad Gerstner (“Feast for the brain.”) .
-
Key takeaway:
“You want to make mistakes. You just want to make them small, iterate and fix them quickly.”
- Why it matters: Reinforces an execution rhythm—tight feedback loops and small, reversible bets—that compounds learning in fast‑moving markets.
- Link: Source tweet https://x.com/davidsenra/status/1978644001667551490 (direct video: https://video.twimg.com/amplify_video/1978643309225787392/vid/avc1/1920x1080/tzjxfi-zbPQ0MBuV.mp4?tag=21)
Bertrand Serlet — WWDC presentations (video)
- Content type: Talks (WWDC)
- Author/creator: Bertrand Serlet
- Recommended by: Patrick Collison, noting Serlet’s outsized influence and scarcity of interviews; suggests watching WWDC appearances to absorb his approach .
- Why it matters: Rare “primary source” exposure to the engineering leadership behind OS X and early iOS can inform taste, simplicity, and platform judgment .
Additional high‑signal picks from today
-
Dylan Field’s non‑fiction pick: The Spy and the Traitor — Book by Ben Macintyre
- Content type: Book
- Recommended by: Dylan Field (“That’s whatever hard situation you’re going through… you read that book and you’re like, okay, could be worse.”) .
- Link: https://www.amazon.com/Spy-Traitor-Greatest-Espionage-Story/dp/1101904216
Founder-led picks: open-source AI policy, decision biases, and execution classics
16 October 2025
•
5 minutes read
Brad Feld
a16z
Ben Horowitz
Today’s high-signal picks span AI policy history, decision-making, execution frameworks, reliability case studies, practical productivity, and a new print magazine. The top recommendation unpacks the near-miss on open-source AI bans and why the landscape looks different now.
Top pick
- Title: Machine‑learning bans (web post)
- Type: Article
- Author/creator: 1a3orn
- Link/URL: https://1a3orn.com/sub/machine-learning-bans.html
- Recommended by: David Sacks (via @sriramk)
- Recommender takeaway: Sacks highlights past efforts to criminalize open‑source AI and calls 1a3orn’s post “a good trip down memory lane,” underscoring how the landscape shifted with new open‑source models and geopolitical competition .
- Why it matters: A concise historical context for founders building with open‑source AI—clarifies what was proposed, why it nearly happened, and how the ground has shifted .
“Let’s never forget how outlandish their proposals were… 1a3orn’s post here is a good trip down memory lane.”
Decision tools for better bets
Title: Richard H. Thaler: Sunk cost fallacy (20‑second explainer)
- Type: Video clip
- Author/creator: Richard H. Thaler
- Link/URL: https://video.twimg.com/amplify_video/1978511670583361536/vid/avc1/1080x1920/iX2wi9nl5XRFo46h.mp4?tag=21
- Recommended by: Tim Ferriss (shared the clip)
- Recommender takeaway: A crisp, expert explanation of sunk cost bias in under 20 seconds .
- Why it matters: Fast, memorable mental model to avoid throwing good effort after bad.
Title: Sunk cost fallacy (mini‑encyclopedia entry)
Type: Article
Author/creator: BehavioralEconomics.com
Link/URL: https://www.behavioraleconomics.com/resources/mini-encyclopedia-of-be/sunk-cost-fallacy/
Recommended by: Tim Ferriss (shared the excerpt and link)
Key idea (definition):
“Individuals commit the sunk cost fallacy when they continue a behavior or endeavor as a result of previously invested resources (time, money or effort)… related to loss aversion and status quo bias.”
Why it matters: A canonical, citable reference to pair with the quick video.
Execution and leadership staples (from Ali Ghodsi’s playbook)
Title: High Output Management
- Type: Book
- Author/creator: Andy Grove
- Recommended in: Ali Ghodsi × Ben Horowitz conversation
- Ghodsi’s takeaway: Managerial leverage—hire and develop people so exceptional that you’re learning from them .
- Why it matters: Systematizes scaling output through people.
Title: Radical Candor
- Type: Book
- Author/creator: (not specified in source)
- Recommended in: Ali Ghodsi × Ben Horowitz conversation
- Ghodsi’s takeaway: Recast feedback as help so people lean in and improve, rather than feel attacked .
- Why it matters: Enables high‑trust, high‑velocity coaching cultures.
Title: Amp It Up
- Type: Book
- Author/creator: Frank Slootman
- Recommended in: Ali Ghodsi × Ben Horowitz conversation
- Ghodsi’s takeaway: A playbook for driving execution and a high‑performance culture at scale .
- Why it matters: Practical patterns for raising the bar across teams.
Title: The Hard Thing About Hard Things
- Type: Book
- Author/creator: (not specified in source)
- Recommended in: Ali Ghodsi × Ben Horowitz conversation
- Ghodsi’s takeaway: “Probably the best business book I’ve read,” influential in Databricks’ formative years .
- Why it matters: Frank guidance for navigating the toughest CEO calls.
Founder mental models and mentorship
Title: The Five Love Languages (applied at work)
- Type: Book/framework
- Author/creator: (not specified in source)
- Recommended by: Brad Feld
- Key takeaway: Use it as a mentorship lens—meet people where they are; translate “love languages” into the work context .
- Why it matters: Improves how feedback lands and relationships function inside teams.
Title: Hero’s Journey (Joseph Campbell)
- Type: Framework
- Recommended by: Brad Feld
- Key takeaway: Startup arcs have highs and lows—expect periods of increasing difficulty for the protagonist .
- Why it matters: Normalizes turbulence and sustains founder resilience.
Title: The Goal
- Type: Book
- Author/creator: (referenced as “the Goal”)
- Recommended by: Robert Glazer (alongside Pat Lincioni and Bob Berg as influences)
- Key takeaway: Story‑driven, values‑led leadership—use narrative and Socratic coaching to align the “big three”: vocation, partner, and community .
- Why it matters: Shows how to translate core values into day‑to‑day decisions.
Title: Level 5 Leadership (Jim Collins)
- Type: Leadership concept
- Recommended by: Robert Glazer
- Key takeaway: Self‑awareness sits at the core; personal values often stem from formative experiences—either doubling down on what mattered or reacting against prior pain .
- Why it matters: Anchors leadership development in introspection and values.
Reliability and product design case studies
Title: WhatsApp engineering blog posts (archived)
- Type: Blog posts (case studies)
- Where to find: Internet Archive
- Recommended by: Guillermo Rauch
- Key takeaway: How WhatsApp “sweated the details” to achieve reliability on poor networks; scaling choices, including uncommon language/stack, documented in depth .
- Why it matters: Concrete, real‑world reliability patterns worth emulating.
Title: Sora “model + experience” product pattern
- Type: Product example (video/UX pattern)
- Recommended by: Guillermo Rauch
- Key takeaway: The “magic” wasn’t just the model—pairing features like at‑mentioning someone and bringing them into a video shows the power of model plus experience .
- Why it matters: Reminds builders to design end‑to‑end experiences, not just models.
Practical productivity
-
Title: 10× your productivity with these simple shortcuts (Comet browser thread)
- Type: X thread
- Author/creator: @defikito
- Link/URL: https://x.com/defikito/status/1978314007434989947
- Recommended by: Aravind Srinivas (“Good thread on Comet use cases”)
- Recommender takeaway: Practical, high‑leverage shortcuts; the OP calls Comet “by far the most superior browser out there” .
- Why it matters: Quick wins for daily workflows; explore @comet for the tool itself .
New long‑form journalism
-
Title: Colossus (print magazine)
- Type: Magazine (print + private audio feed)
- Link/URL: https://shop.joincolossus.com/subscribe
- Recommended by: Tobi Lütke (“New publication doing real journalism… as amazing as they said it was.”)
- First‑issue highlights: Profiles of Joe Liemandt and Thomas Peterffy; @ganeumann on investing in AI; an essay on taste and beauty in math .
- Why it matters: Curated, durable analysis in a quarterly format, plus early‑access audio for deeper dives .
Founder/VC picks: Kushner profile leads, with Gurley’s book and strategy/epistemology classics
15 October 2025
•
4 minutes read
Naval
Reid Hoffman
Patrick Collison
A curated, non‑promotional roundup of founder/VC‑endorsed resources. Highlights a multi‑voice profile on Joshua Kushner/Thrive, plus career, hiring, innovation, and epistemology classics from operators like Brad Gerstner, Tracy Britt Cool, Tom Blomfield, Naval, Jensen Huang, Patrick Collison, and Tomasz Tunguz.
Today’s standout pick
Joshua Kushner and Thrive Capital — category‑defining profile (Only in Colossus)
- Content type: Long‑form profile/feature
- Author/publisher: Jeremy Stern, Only in Colossus
- Link: https://x.com/joincolossus/status/1978061765591405039
- Recommended by: Scott Belsky; Reid Hoffman
-
Key takeaways:
- Voices include Sam Altman, Jony Ive, Demis Hassabis, and John Collison
- Traces the American‑dream arc from the Holocaust through postwar Europe to the AI era
- Details the 2023 OpenAI coup attempt and critical moments at Stripe and GitHub; shows how Thrive was built as a “rampart for the American experiment”
- Noted for an emotionally powerful ending
- Why it matters: Multiple respected operators call it the category‑defining profile of a pivotal figure and institution in tech and investing .
“super insightful profile of an investor, confidante, and someone w/ a unique vision that has only started to materialize.”
“Highly recommend giving this a read… It’s a piece that captures Josh’s dedication to his craft and his deep care for the people he encounters on the journey. The blueprint!”
Additional high‑signal recommendations
Runnin’ Down a Dream (Bill Gurley)
- Content type: Book
- Author/creator: Bill Gurley
- Link: https://a.co/d/0oIPPve
- Recommended by: Brad Gerstner; Chamath Palihapitiya
-
Key takeaway (from recommenders):
- “Reserve 10 copies… the best gift you can give to those chasing their dreams!”
- “You may want to buy this book before you pay for your college tuition.”
- Context: Featured on BG2Pod with segments on “Bill’s Book: Running Down a Dream,” Bezos’ Regret Minimization Framework, and “Who the Book Is For” .
- Why it matters: Strong endorsements frame it as a pragmatic career accelerator for builders and operators .
Who (GH Smart)
- Content type: Book (hiring)
- Author/creator: GH Smart
- Recommended by: Tracy Britt Cool
- Key takeaway: “single handedly the best simple book on hiring,” emphasizing starting with an in‑depth role scorecard (mission, outcomes, competencies; time‑bound and measurable) to avoid misalignment and rework .
- Why it matters: Raises hiring signal by enforcing clarity up front .
American Icon
- Content type: Book (turnaround leadership)
- Recommended by: Tracy Britt Cool
- Key takeaway: “single handedly the best book in navigating a turnaround,” highlighting disciplined reviews, celebrating transparency on “reds,” and building a culture of continuous improvement .
- Why it matters: Concrete practices for leading through crisis and change .
The Innovator’s Dilemma (Clayton Christensen)
- Content type: Book (innovation strategy)
- Recommended by: Tom Blomfield
- Key takeaway: Why well‑managed incumbents optimize into local maxima, dismiss “toy” entrants, and then get overtaken as new products rapidly improve—often requiring incumbents to undermine their own models to compete .
- Why it matters: A precise lens for evaluating disruptive entrants vs. incumbent inertia .
Crossing the Chasm (Jeffrey’s book, as cited)
- Content type: Book (go‑to‑market)
- Recommended by: Jensen Huang — “that’s a good book.”
- Why it matters: Endorsed by a leading operator for navigating mainstream adoption gaps .
The Beginning of Infinity (David Deutsch)
- Content type: Book (epistemology/knowledge)
- Recommended by: Naval — “If you just want to know epistemology, read David Deutsch full stop.”
- Key takeaways: High‑density, interdisciplinary framework linking epistemology, evolution, and quantum computation; first and last chapters are most accessible . Introduces “good explanations are hard to vary,” which Naval also applies to product design (“Good products are hard to vary”) . Deutsch’s framing elevates knowledge as the dominant component of wealth creation .
- Companion: Brett Hall’s podcasts can boost comprehension .
- Why it matters: A rigorous foundation for reasoning and building, prized for density and transferability across domains .
The Lessons of History (Durant)
- Content type: Book (historical synthesis)
- Recommended by: Naval — a concise summarization of the massive Story of Civilization series .
- Why it matters: High‑density historical perspective without the 12‑volume commitment .
Works by Jorge Luis Borges
- Content type: Literature (essays/short stories)
- Recommended by: Guillermo Rauch — “highly recommend reading him and… his study of the English language,” noting a return to the value of language and communication in the “prompt engineering” era .
- Why it matters: Elevates precision of thought and language—skills increasingly central to AI‑mediated work .
Titan
- Content type: Book (biography)
- Recommended by: Patrick Collison — “I recently read Titan about Rockefeller” .
- Key takeaway: Used to illustrate that beyond a certain stage, the correlation between hours worked and outcomes is weak; Rockefeller “napped a lot and worked from home” .
- Why it matters: Challenges effort‑fetishism; favors leverage, design, and judgment .
How to Do Science and Engineering (Trip Press)
- Content type: Book (builder’s philosophy)
- Recommended by: Tomasz Tunguz — “Really great… Just a philosophy of how to build things. Brilliant, brilliant book.”
- Why it matters: A concise, principled blueprint for building and experimentation .
High-signal picks: nanochat, A Culture of Growth, Ben Thompson on rare earths, and Seven Powers
14 October 2025
•
3 minutes read
Patrick Collison
Elad Gil
tobi lutke
Authentic, non-promotional picks from founders and investors. Top pick: Karpathy’s nanochat repo for an end-to-end, reproducible LLM training + inference baseline. Plus high-signal reads from Mokyr, Ben Thompson, and Hamilton Helmer.
Today’s most useful founder/investor‑recommended resources
Top pick: nanochat (full-stack LLM training + inference repo)
- Content type: Code repository
- Author/creator: @karpathy
- Link/URL: https://x.com/karpathy/status/1977755427569111362
- Recommended by: Tobi Lütke
"What a gift this is"
-
What it covers:
- Minimal, from-scratch, full-stack pipeline for a simple ChatGPT-like model; boot a cloud GPU, run one script, and talk to your own LLM via a web UI in as little as ~4 hours .
- Training steps include: tokenizer with a new Rust implementation ; pretrain on FineWeb and evaluate CORE ; midtrain on SmolTalk conversations, multiple-choice, and tool use ; SFT plus evals on ARC-E/C, MMLU, GSM8K, HumanEval ; optional RL on GSM8K with “GRPO” ; efficient inference engine (KV cache, prefill/decode, tool use via a sandboxed Python interpreter) with CLI and ChatGPT-like Web UI ; one markdown “report card” summarizing and gamifying the run .
- Cost/perf snapshot: ~US$100 (~4 hrs on 8×H100) yields a basic chat model; ~12 hrs surpasses GPT‑2 on CORE; scaling to ~US$1,000 (~41.6 hrs) improves coherence. Example: depth‑30 @24 hrs ~40s MMLU, ~70s ARC‑Easy, ~20s GSM8K .
- Intent: a cohesive, minimal, readable, hackable “strong baseline” with potential to evolve into a research harness/benchmark, similar to nanoGPT .
- Access: repo URL and a detailed speedrun are linked in the announcement thread’s reply .
- Why it matters: It lowers the barrier to hands‑on LLM training and evaluation with a cohesive baseline that’s cheap to reproduce and extend, useful for both learning and research/benchmarking .
A Culture of Growth — Joel Mokyr
- Content type: Book
- Author/creator: Joel Mokyr
- Link/URL: https://www.amazon.com/Culture-Growth-Origins-Schumpeter-Lectures/dp/0691180962/
-
Recommended by: Patrick Collison and Elad Gil
- Patrick Collison: "Congratulations to Joel Mokyr! A terrifically deserving winner. Read A Culture of Growth if you haven’t."
- Elad Gil: "Worth a read-"
- Why it matters: Multiple independent endorsements from prominent tech leaders increase signal that this is a high‑value read .
Ben Thompson on rare earths (commentary)
- Content type: Commentary
- Author/creator: Ben Thompson
- Link/URL: https://x.com/benthompson
- Recommended by: Elad Gil
- Key takeaway (quoted):
“U.S.’s deficiency in rare earths isn’t a matter of needing to catch up to technological excellence; it’s needing to simply have the will, incentives, and regulatory regime to do what we already did before.”
“TL;DR they are not rare, we just chose not to make them in USA”
- Why it matters: Frames the rare earths challenge as primarily about will, incentives, and regulation rather than technological capability .
Seven Powers — Hamilton Helmer
- Content type: Book
- Author/creator: Hamilton Helmer
- Link/URL: Not provided
- Recommended by: Harry Stebbings (20VC)
“There’s a great book by Hamilton Helmer Seven Powers… If you haven’t read it, please let me send it to you as a gift. It’s the best business book you’ll ever read.”
- Why it matters: Strong endorsement from an investor-operator host signals practical strategic value for builders and operators .
Network State 2025 talks now free: standout sessions for builders of internet communities
13 October 2025
•
3 minutes read
Balaji
balajis.com
All Network State 2025 talks are now free online. We highlight the most useful sessions for builders—Naval, Vitalik, Ben Horowitz, Brian Armstrong, Amjad Masad, Andrew Huberman, and Pieter Levels—with links, key takeaways, and why they matter.
Network State 2025 talks now free: standout sessions for builders of internet communities
All Network State 2025 talks are now free online. Here’s a link to the YouTube playlist: https://www.youtube.com/watch?v=ZLNYkftDP98&list=PLJg2RipiXz8oDoMPCsmOVI_UEaSNbgjA6
Coinbase (Brian Armstrong, CEO). Brian Armstrong, CEO and cofounder of Coinbase, talks to Balaji about their joint history of success, and tells the audience that anyone working on real cryptocommunities and network states should build them on Base and reach out to Coinbase Ventures.