Byrne Hobart
Longreads
- Daniel Dennett has died. Dennett focused on areas where there isn’t a sense of progress so much as a sense that we've found engaging new ways to frame debates that will never get resolved—on the supernatural, on what consciousness actually is, what we mean by "free will," etc. My favorite Dennett piece is this short story about brains in vats and consciousness.
- Austin Vernon on making off-grid datacenters affordable. It's a good first-principles look at what the right problem to solve really is. For example, he makes the early point that going from 100% solar/batteries to solar/batteries with diesel generators supplying 1% of total power needs cuts solar needs by 60-80%—that last bit of redundancy is very expensive! Meanwhile, being tolerant of interruptions (which is really what diesel generators offer) also substantially cuts the bill. A fun, worthwhile piece from someone who's done the math on an increasingly salient question.
- Ben Evans on the search for AI killer apps. The productivity paradox of AI is that it's clearly powerful, but hasn't moved the economic needle outside of directly-affected sectors; if you're a freelance translator or a deep learning expert, you've personally experienced an economic impact from AI, for better or for worse. By making it easier to get quick answers to arbitrary questions, AI has raised the value of good questions.
- Ezra Klein interviews Anthropic CEO Dario Amodei. In reference to the Ben Evans piece above, Amodei has a nice early riff on how in 2021-22, everyone working in AI could see that there was incredible progress, and wondered when the rest of the world would notice. And then ChatGPT shipped, and we noticed. There's also a discussion of how AIs are more persuasive when they're told to make an argument for some issue and allowed to lie. Amodei says regular people don't do this, which is not really true—lying, ideally in deniable ways, is a ubiquitous feature of discourse. If anything, dishonest AIs will commoditize it, and the lower cost of producing text in favor of or against a particular view will make people more discerning about how credulous they should be.
- Jay Caspian Kang on reading books on your phone. Pieces like this are worthwhile, because a shift to reading on screens a) happens subtly enough that it's hard to see its long-term effects, but b) is a different experience from reading on some medium that doesn't allow easy interruptions and task-switching.
- This week in Capital Gains, we talk about the thermostat problem: causation in the economy is so complicated that things don't correlate well with what they cause—higher supply causes lower prices, but also coincides with higher ones; rate cuts increase growth, but happen when growth is slow.
- This week on The Riff: acquisitions, memos, and why custom implementations of AI models are such a hard business. Listen with Twitter/Spotify/Apple/YouTube.
Books
How to Make a Few Billion Dollars: there are some fairly generic terms for people who have made, or are trying to make, a lot of money—"entrepreneur" and "investor" are labels that get applied to lots of people. We've shied away from some more old-fashioned descriptors like "industrialist" or "businessman," but those seem more applicable to Brad Jacobs, author of How to Make a Few Billion Dollars and an avid practitioner of the art described in that title.
Jacobs co-founded some oil trading firms early in his career, but his later success comes from three big rollups: United Waste Systems, United Rentals, and XPO Logistics. These companies all did non-glamorous things—hauling trash and operating landfills, renting industrial equipment, and freight brokerage. They had a few things in common: fragmented industries with long-term tailwinds, in businesses with a compounding data advantage.
If Jacobs had been born a generation later, he might have looked at the same industries and decided to build a software product that these companies would buy; a company like Shopify isn't monopolizing e-commerce, but it is maximizing its market share in the highest-margin, least capital-intensive part of the stack. But when that business isn't available, or the cost of selling the software is prohibitive, the best option can be a rollup strategy of buying small companies, running them better, and taking advantage of scale effects to get better returns beyond that. Each of the companies Jacobs built could create upside from higher utilization of the same physical assets. It's an early case of software and PE competing to eat the world.
Open Thread
- Drop in any links or comments of interest to Diff readers.
- What are some very early, very promising companies I should know about?
Diff Jobs
Companies in the Diff network are actively looking for talent. See a sampling of current open roles below:
- A fintech company using AI to craft new investment strategies seeks a quantitative research analyst with 2+ years of experience. (NYC)
- A company building ML-powered tools to accelerate developer productivity is looking for ML researchers with a knack for converting research into Github repos. (Washington DC area)
- A CRM-ingesting startup is on-boarding customers to its LLM-powered sales software, and is in need of a product engineer with a track record of building on their own. (NYC)
- A concentrated crossover fund is looking for an experienced full stack software engineer to help develop and maintain internal applications to improve investment decision-making and external applications to enable portfolio companies. (SF)
- Several companies in the Diff network are looking for smart generalists interested in their next challenge, especially for people who like solving interesting problems at the intersection of math and humanities.
Even if you don't see an exact match for your skills and interests right now, we're happy to talk early so we can let you know if a good opportunity comes up.
If you’re at a company that's looking for talent, we should talk! Diff Jobs works with companies across fintech, hard tech, consumer software, enterprise software, and other areas—any company where finding unusually effective people is a top priority.