Gianfranco's Best of September 2024 Reading List

Welcome to the September 2024 edition of my monthly reading list.

This month, I've curated my favorite essays across various categories: Artificial Intelligence, Semiconductors, Finance and Economics, Business and Technology, Startups and Founder Guidance, and Life and Meaning.

If you only have a few minutes, these three posts were my favorite, and are included in the list below:

The Trouble with Optionality (Mihir Desai)

• Desai challenges the conventional wisdom of maximizing career choices. He argues that this relentless pursuit of "optionality," while seemingly empowering, can become a gilded cage, trapping us in a cycle of risk aversion and delayed gratification. We become so fixated on keeping our options open that we lose sight of our true passions and postpone taking the leaps that lead to fulfillment. The question is: are we so busy building a safety net that we forget to live?

The Playing Field (Graham Duncan)

• Duncan reveals a hierarchy of mastery in the investing world, a journey that extends far beyond simply generating returns. He outlines five distinct levels, from the novice Apprentice to the seasoned Steward, emphasizing the importance of mentorship, self-awareness, and adapting your strategy to your unique strengths. But it's the Level 4 investors, the David Teppers of the world, who truly redefine the game. They operate with a rare blend of open-mindedness and conviction, laser-focused on extracting value, not on proving their brilliance. The question is: where do you stand on this playing field, and what steps can you take to level up?

DeepSeek's Open-Source Gambit: Can Collaboration Unlock AGI? (INF News)

• DeepSeek, a Chinese AI company, is charting a different course. While others race towards commercialization, DeepSeek has doubled down on fundamental research and open-source models. Their focus on architectural innovation, embodied in their MLA model, has sparked a price war in China and earned grudging respect in Silicon Valley. The question is: can this commitment to open collaboration and a focus on fundamental breakthroughs become a viable path to Artificial General Intelligence, even amidst fierce competition? Is DeepSeek's gamble a bold vision of the future or a risky deviation from the proven path?

Technological Idealism

• Unveiling DeepSeek: A more extreme story of Chinese technological idealism - inf.news
  • Excerpt: “In the face of disruptive technology, the moat formed by closed source is short-lived. Even if OpenAI closes its source code, it cannot prevent being overtaken by others. So we put our value on the team. Our colleagues grow in this process, accumulate a lot of know-how, and form an innovative organization and culture, which is our moat ... Open source and publishing papers actually do not lose anything. For technical personnel, being followed is a very fulfilling thing. In fact, open source is more like a cultural behavior rather than a commercial behavior. Giving is actually an extra honor. A company doing this will also have cultural appeal.”
  • Read More

Semiconductors

• Intel Honesty - Stratechery
  • Excerpt: “Here is the fundamental problem facing Intel, and by extension, U.S. dreams of controlling leading edge capacity: there is no reason for Intel Foundry to exist. Apple, Nvidia, AMD, and other leading edge fabless chip companies rely on TSMC, and why wouldn’t they? TSMC invested in EUV, surpassed Intel, and are spending tens of billions of dollars a year to continue pushing forward to 2nm and beyond. Yes, TSMC priced 3nm too low, but even if the company raises prices for future nodes, as I expect them to, the relative cost matters much less than TSMC’s superior customer services and demonstrated reliability.”
  • Read More

Artificial Intelligence

• OpenAI’s Strawberry, LM self-talk, inference scaling laws, and spending more on inference - Interconnects.ai
  • Excerpt: “What if you could lower your usage limit on ChatGPT or Claude, in exchange for every answer being a higher quality? This seems like a great paid feature to me.”
  • Read More
• OLMoE and the hidden simplicity in training better foundation models - Interconnects.ai
  • Excerpt: “In order to scale your training up to a GPT-4 level you need very stable training at the GPT-3.5 level. In order to train really cheap, really small models like Gemini Flash, Claude Haiku, and GPT-4o-mini, you need a larger model to distill from. In order to train a solid vision model, you need a base text model to fuse it with... The operating strategies of these labs are unique and different from both how technology has proceeded in the past and how people may expect by observing hiring patterns.”
  • Read More
• Scaling: The State of Play in AI - OneUsefulThing
  • Excerpt: “What OpenAI did was get the o1 models to go through just this sort of “thinking” process, producing hidden thinking tokens before giving a final answer. In doing so they revealed another scaling law - the longer a model “thinks,” the better its answer is. Just like the scaling law for training, this seems to have no limit, but also like the scaling law for training, it is exponential, so to continue to improve outputs, you need to let the AI “think” for ever longer periods of time. It makes the fictional computer in The Hitchhikers Guide to the Galaxy, which needed 7.5 million years to figure out the ultimate answer to the ultimate question, feel more prophetic than a science fiction joke.”
  • Read More

Finance and Economics

• Shall We Repeal the Laws of Economics? - Oaktree Capital
  • Excerpt: “This is a good time for me to cite the economist’s adage that “the best solution for high prices is high prices.” This isn’t a joke; far from it. In general, high prices mean demand is strong relative to supply. Eventually, those high prices will encourage producers to produce more and consumers to consume less, and the depressant impact on prices from both directions is obvious. We see this all the time in the oil market, for just one example.”
  • Read More
• Finance as culture - John Luttig
  • Excerpt: “Importantly, low discount rates most heavily impact the valuations for companies with cash flows far in the future. Nikola, for example, has practically no income but promises cash flows on a multi-decade timescale, earning a $7b market cap. A 1% change in interest rates makes a one-year bond change in value by ~1%, but could change Nikola’s value by 20% or more. Stated differently, if you don’t discount cash flows, you believe the future is all but certain.”
  • Read More
• When is a loss not a loss? - Acadian Asset
  • Excerpt: “Denial is not a river in Egypt. When anyone, even Warren Buffett, uses a six-syllable euphemism (“quotational shrinkage”) for a one-syllable concept (“loss”), that’s a red flag.”
  • Read More
• Invest like the worst: Wealth-destroying portfolio concentration - Acadian Asset
  • Excerpt: “Sharpe ratios increase with IQ … High-IQ investors’ superior Sharpe ratios thus arise mostly from lower portfolio volatility … cognitive skill increases the number of stocks held.”
  • Read More
• What I Learned When I Stopped Watching the Stock Market - WSJ
  • Excerpt: “When you don’t watch the market every day, you can finally see with unquestionable clarity that what you would have expected to happen didn’t. The unexpected did.”
  • Read More
• One Thing You Wish People Better Understood About Venture Capital - Hunter Walk
  • Excerpt: “But not every VC firm and individual VC is primarily incentivized by the Power Law. Firms with very large funds make a lot of money for themselves in management fees before realizing returns. Individuals get promoted for successfully out-competing other investors and winning over founders long before knowing whether those were great investments to make. Those incentives are often misaligned with the interests of founders and limited partners.”
  • Read More

Business and Technology

• Boomer Apple - Stratechery
  • Excerpt: “Software, specifically AI, is what will drive differentiation going forward, and even in the best case scenario, where Apple’s AI efforts are good enough to keep people from switching to Google, the economics of software development push towards broad availability on every iPhone, not special features for people willing to pay a bit more.”
  • Read More
• How pour-over coffee got good - Works in Progress
  • Excerpt: “As much as some coffee connoisseurs hate to admit it, a well-made but relatively inexpensive machine like the 1971 Moccamaster can often brew coffee better than you can with a pour-over.”
  • Read More

Startups and Founder Guidance

• The Idea Maze is a Useless Idea - Commoncog
  • Excerpt: “Good entrepreneurs understand that the whole game of entrepreneurship is a game of improvisation. There is no knowledge to be had; no advice one can read. There’s simply no guaranteed answer to anything: you take lots of action to generate information and stay alive, and then roll with whatever comes your way. You are prepared to do this for years.”
  • Read More

Life and Meaning

• Are you serious? - Visakanv Substack
  • Excerpt: “One of the stupidest things I’ve repeatedly seen smart people do – I call this Advanced Stupid – is that they get so fixated on trying to optimize something complicated, that they forget to optimize for survival, and then they get knocked out of the game. Like having a startup trying to do something complicatedly clever, and then running out of money. The thing here is to study the failure conditions – how and why have past players failed to do what you’re trying to do? – and be conscientious about avoiding that.”
  • Read More

Interesting Graphs:

Memes: