Gianfranco's Best of April 2024 Reading List

Gianfranco's Best of April 2024 Reading List

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

This month, I've curated my favorite essays I read this month across Society and Technology, Health and Wellness, Artificial Intelligence, Semiconductors, Finance and Economics, and Early-Stage Investing.

Society and Technology

• GenZ software engineers, according to older colleagues ($) (Pragmatic Engineer)

  • Excerpt: "GenZ seems more inspired by a strong sense of purpose and meaning, according to their older colleagues. Maybe this is why they’re seen as questioning everything: they want to know the why of tasks and decisions, and to do things for the right reasons!"

  • Read more

• The perverse incentives of euthanasia (Noahpinion)

  • Excerpt: "And having been a depressed person, I can tell you with confidence that depressed people are not fully capable of pushing back against a doctor’s suggestions. Depression — at least, the kind that I suffered — entails a lack of volition, willpower, and motivation. When I was depressed, I would go along with other people’s suggestions, simply because I didn’t have any motivation to do anything other than go along. Fortunately, nobody suggested self-harm to me. But if a doctor had suggested that I’d be better off dead, I can’t tell you with 100% certainty that I would have had the wherewithal to gainsay them."

  • Read more

• Meta 1Q'24 Update (Deep Dives)

  • Excerpt: "Meta AI is not a speculative/unproven idea on which Meta is aggressively investing; business messaging, and high intent data from chat bots, and hence getting more ad dollars from the bottom of the ad funnel, is likely to be a once-a-decade opportunity for Meta."

  • Read more

Health and Wellness

• The Wildly Asymmetric Impact of Diet and Nutrition (Graham Weaver)

  • Excerpt: "Simply put, I had been spending hours and hours exercising when I could have made a vastly larger impact on my body and my energy by spending just minutes being more thoughtful about what I put in my mouth."

  • Read more

Artificial Intelligence

• Gemini 1.5 and Google’s Nature ($) (Stratechery)

  • Excerpt: "The great thing about a monopoly is that a company can do anything, because there is no competition; the bad thing is that when the monopoly is finished, the company is still capable of doing anything at a mediocre level, but nothing at a high one, because it has become fat and lazy. To put it another way, for a former monopoly, 'big' is the only truly differentiated asset."

  • Read more

• Data Is All You Need (AI Changes Everything)

  • Excerpt: "Meta Llama 3’s edge in high-performance-for-size came from LLM training that was done on a massive amount of data: 15 trillion tokens. Phi-3’s edge in high-performance-for-size came from LLM training that was done on higher quality data: 3 trillion tokens. By being high-quality for their size, Llama 3 and Phi-3 cut the cost of inference. The recipe for a good GPT-4-level LLM is now known: Fifteen trillion tokens/words of data; 7M GPU hours of H100 compute; and the machine-learning team with expertise to train an LLM. Phi-3 shows an additional path: Do dataset optimization to improve the training LLM quality for a given amount of training compute."

  • Read more

• Where AI is Headed in 2025: A Builder’s Guide (B2BaCEO by Ashu Garg)

  • Excerpt: "Rather than relying on monolithic models to handle complex tasks end-to-end, the cutting edge of AI development is embracing a modular, system-level engineering approach. These compound AI systems employ multiple models that work together in iterative, looped processes, coordinating with each other and drawing on external tools to evaluate, refine, and improve their own results. By breaking down a large task into a series of subtasks, this approach allows builders to better optimize and troubleshoot each subtask, leading to higher overall performance."

  • Read more

Semiconductors

• Is Intel Back? Foundry and Product Resurgence Measured ($) (Semi Analysis)

  • Excerpt: "Intel says they need $25B to $30B of capex per 10,000 wafers per week for their new fabs. TSMC has stated they require about $42B of capex per 10,000 wafers for their 3nm in Arizona even after the 'cost overruns.' We are unsure what the delta is for Intel versus TSMC on these numbers, as the TSMC figures include site prep, shell, and tooling. Regardless, taking Intel’s numbers at face value, this is a huge wall given Intel always needs 150,000+ wafers per month of capacity on the leading edge to remain competitive in volumes with TSMC."

  • Read more

• The Data Center is the New Compute Unit: Nvidia's Vision for System-Level Scaling ($) (Fabricated Knowledge)

  • Excerpt: "The new Moore’s Law is about pushing the most compute into a rack. Also, looking at Nvidia’s networking moat as InfiniBand versus Ethernet is completely missing the entire point. I think the NVLink domain over passive copper is the new benchmark of success, and it will make a lot of sense to buy GB200 NV72 racks instead of just B200s."

  • Read more

• AI Accelerators: The Cambrian Explosion ($) (The Chip Letter)

  • Excerpt: "Drawing a historical analogy to Helen of Troy—'the face that launched a thousand ships'—we say tongue-in-cheek that TPU v1 'launched a thousand chips.' ‘A thousand chips’ is obviously an exaggeration, but the decade following the launch of Google’s TPU v1 in 2015 has certainly seen a lot of activity in the field of machine learning, and more specifically deep learning, accelerators."

  • Read more

• Alchemy is all you need (Airstreet)

  • Excerpt: "It could be that a world of smaller models with long context lengths end up becoming the go-to for the vast majority of common applications. Smaller models, in the 7B range, are compatible with a broad spectrum of GPUs, including the older V100 and A100, which are less powerful, but more cost-efficient than the newer H100. The Apple M-series chip, with its integrated CPU/GPU unified memory architecture, means it’s already possible to run Mistral-7B on some Macs."

  • Read more

Finance and Economics

• Anatomy of a credit card rewards program (Bits About Money)

  • Excerpt: "People love Amazon and Apple, and they love free Amazon and free Apple even more. This is true even among a portion of very sophisticated, wealthy, numerate CSR users, who love this idea so much they click a button designed for suckers. Why is that a sucker’s checkout button? Because CSR also includes a feature called Pay Yourself Back, which in the past prominently, and today a bit… less prominently, lets you cash out rewards at better than 1:1. You get a 25% bonus if you Pay Yourself Back by nominating past transactions at grocery stores: 10,000 points gets you $125 in statement credits if you are willing to do a trivial amount of clicking to show Chase $125+ in spend at grocery stores."

  • Read more

• The Indispensability of Risk (OakTree Capital)

  • Excerpt: "No one wants to be exposed and get hit. But most beginners play it too safe, and because they put so much emphasis on avoiding getting hit, they rarely win."

  • Read more

Early-Stage Investing

• YC's Secret SAFE (Law of VC)

  • Excerpt: "What if we simplified the SAFE? Scrap the Cap, the Discount, and the MFN, and replace those terms with one simple metric: the Conversion Percentage."

  • Read more

• The Arc Product-Market Fit Framework (Sequoia Capital)

  • Excerpt: "You take a pain point universally accepted as a hard fact of life, and see that it’s merely a hard problem that your product solves for the customer. Your customers have resigned themselves to just living with the problem. They’re not urgently engaged with trying to solve it. The status quo is just how it is, and change doesn’t seem like an option. You upend how things are done with an unexpected approach: Facts can’t be changed—but problems can be solved. The challenge to overcome is force of habit."

  • Read more

Writing as Craft

• Searching for True Words (Alchemy)

  • Excerpt: "I was writing at the intersection of money and the search for a meaningful life. How on earth did I manage to make that boring? Easy. The way you make any subject boring: by stripping it of its soul. Take something complex and treat it naively. Take something deeply human and pretend it is rational. Take something potent and pretend not to be affected by it."

  • Read more