- Matt Bornstein, Guido Appenzeller, and Martin Casado at A16Z ask who retains the best economics in generative AI? In some ways, the industry has echoes of search and social media, where the winners monetized just enough to demonstrate that they could earn money, but not so much that it impeded growth. But that model is harder when the addressable market is bigger and the fixed costs of the service are higher. They conclude that it's hard to find any unique competitive advantages (aside from relatively weak ones that mostly come down to scale that can be replicated with capital). For a Diff look at AI as a commodity industry, and ways it might not be, see this Diff piece from a few weeks ago.
- Sebastian Bensusan has a good post on "high-variance management." This is part of the generally useful principle that if everyone's focused on the same moment of the distribution, you should care about some other one: when everyone's fixated on averages, look at standard deviations; when everyone's focused on standard deviations, spend some time on averages—but also look at skewness. Depending on the distribution, you may not be able to collect meaningful data, but knowing which metric is most closely connected to the payoff you care about is useful.
- Jacob Bernstein in the NYT has obituary of gossip blogger Nikki Finke. This piece is notable for some tidbits on how the gossip trade works. Another journalist recalls: "I was saying that I had gotten this great tip and that I had a really reputable first source but not a second... And she basically said you call five people, plant a seed in their brain and wait a day for someone else to have heard it." What's great about this is that gossip is reporting news on human behavior, so spreading a plausible rumor can actually make it true. This would not work as well for, say, reports on progress in nuclear fusion, but if it's a tip about an upcoming movie release or a high-profile divorce, the rumor can become self-fulfilling. This dynamic applies in other places, like startup acquisitions, funding rounds (here's an example that is only technically fictional; here's a real-world one) , and bank runs.
- Mia Sato and James Vincent in The Verge look into CNET's use of AI for writing articles. A good story to read between the lines of, especially if you're anxious about your job being automated away any time soon. The AI-written articles in question were usually fairly generic ones, like daily posts on mortgage rates. There just isn't that much to say about shifts a few basis points, and "what does refinancing cost today?" is exactly the kind of question for which a natural-language answer should come from a language model rather than a person. But the key anxiety is here: "Those who knew of the AI tool and its uses say that the workflow was so unclear, they sometimes couldn’t distinguish between AI-written stories and articles written by colleagues." Search as an interface has accidentally created a jobs program for people who are able and willing to write English-language text on fairly boring but monetizable topics, which has been a nice accidental subsidy for writers. (I benefited early in my career, and I know at least one person who was cranking out just-for-SEO articles at a day job in order to pay rent while writing a novel.) Losing that has some obvious bottom-line benefits, but there's a cost as well. Though perhaps some of these writers will get day jobs as six-figure prompt engineers instead.
- Andrew Walker at Yet Another Value Blog has a fun look at what would happen if you tagged along with the investments of the world's richest people. The main answer: you'd benefit a lot from Buffett and Gates, and would otherwise end up betting on every big trend—Japan in the 80s, tech in the 90s, commodities in the 2000s—right at the worst possible time. It's a relief to get statistical evidence for what good investors always need to remember: no matter how good a year you're having, someone out there made more money than you specifically because they made a lucky bet with an irresponsible amount of leverage. But they won't do well forever.
- And in The Diff's educational newsletter, Capital Gains, this weeks' explainer is on the Alchian-Allen effect, which shows why luxuries are cheaper when everything is expensive. You can sign up for the newsletter for free here. Upcoming pieces include: explaining the "pod shop" model for hedge funds; how to think about book value; and the first ten minutes, hours, and years of getting up to speed on a stock.
- The China Model: Political Meritocracy and the Limits of Democracy: this book is partly a comparison of the American and Chinese political systems, but it’s also a comparison between the existing systems of China and Singapore and a hypothetically more Confucian alternative. It's an interesting book, and the author does a good job of trying to view things from the perspective of someone with no strong priors about which political system works best. But one way to read it is that there's more similarity between systems than either side would like to admit, since all governments face some of the same political selection pressures. China has voting (of limited real-world impact) for extremely local positions, and pays attention to popular opinion for pragmatic reasons. Meanwhile, the US has indirectly and accidentally adapted elements of political meritocracy: the range of debate is partly set by employees at big tech companies, who are hired in a way that at least attempts to measure their skills rather than their popularity; prestige media, also an institution that doesn't aim for a very democratic hiring process, also plays an important role in this. So one conclusion is that the models for running a modern country are not set points on a one-dimensional continuum, but a series of partly-overlapping fuzzy blobs in a high-dimensional space.
A Word From Our Sponsors
Built by investors, for investors, Tegus is a one-stop investment intelligence and research platform that modernizes research processes.
Our ecosystem combines at-cost, on-demand expert calls with a 55K+ transcript library; quantitative financial workflows that streamline research across company disclosures, management presentations, earnings calls and filings; and 4K+ fully-drivable financial models and company benchmarking data, including every KPI and comparison that matters. Better research starts with Tegus.
For a limited time, The Diff readers can trial the Tegus platform for free at http://www.tegus.com/thediff/.
- Drop in any links or comments of interest to Diff readers.
- We're spending some time looking at the defense tech space in the coming months. Readers with industry experience and/or strong opinions are encouraged to reach out. A particularly pressing question: if more defense spending moves to software, does this mean software-first challengers win, or that get-the-contracts-first incumbents still dominate and see higher margins?
Companies in the Diff network are actively seeking talent! If you're interested in exploring growth opportunities at unique companies, please reach out. Some top current roles:
- An early-stage startup aiming to reduce labor costs by over 80% in a $100bn+ industry is looking for a part-time technical advisor with robotics experience; this has the potential to evolve into a full-time role. (NYC)
- A company building ML-powered tools to accelerate developer productivity is looking for a mathematician. (Washington DC area)
- A firm using NLP and other ML tools to give retail and institutional investors access to custom-tailored portfolios is looking for a data engineer. (NYC)
- A crypto proprietary trading firm is actively seeking systematic-oriented traders with crypto experience—ideally someone with experience across a variety of exchanges and tokens. (Remote)
- A consulting firm that helps investors use alternative data to inform their views is looking for analysts with experience working with data and a fundamental research bent. (Remote)
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.