This issue of The Diff is brought to you by our sponsor, Daloopa.
Longreads
- Jamie Rumbelow and Henry Dashwood have a post-mortem on their proptech startup, Tract. They "were driven by a mix of moral outrage and technological optimism," which is definitely a way to ramp up the variance—the problem they were addressing was that land in the UK is worth orders of magnitude more once it's legal to build on it, and that making this happen is at least partly a problem of organizing the relevant information. But it's also a problem of how the industry is structured and what the relevant incentives are, and it turned out that this information was not nearly as accessible through webscraping and LLMs. One takeaway from this piece is that starting a startup is partly a meta-optimization problem—the real growth only begins when you've found a KPI you know you can and should push up as fast as possible.
- From Quanta: intelligence evolved independently in mammals and birds. It's clearly a useful trait to have, but brains are expensive, and there seem to be a few different cases where they lead to step-function improvements—coordinating social groups, being able to outplan prey, being able to use simple tools, and eventually being able to build a technological civilization that can recursively self-improve. It seems at least logically possible that intelligence could have been selected against much earlier in our species' history (and given how long life has existed on earth, perhaps humans are just the first species lucky enough to get smart enough to pay the calorie cost of our big brains).
- Chris Gillett reviews 9/11's impact on infrastructure: a good look at how a single shock ripples through a complex system, and all of the second-order effects that show up—planes are grounded, so there's a scramble for cars; there's immense demand for news, so news sites can't handle the traffic; there's also a lot of demand to reroute supply chains, but markets are closed and aren't providing price signals; people want cash, but some banks have closed their branches. It's heartening to remember that things recovered as quickly as they did, when so many backup plans failed at exactly the same time as what they were backups for.
- In Foreign Affairs, the case for the US to work with allies to isolate China. This was written before the tariffs hit, but is a lot more relevant now that China and the EU are negotiating trade deals. Economic power is partly a simple formula multiplying GDP per capita by total population, so any time economies are mostly converging, the most powerful countries will be the biggest ones. As the piece notes, the UK's relative decline is instructive, especially since commentators at the time saw it coming.
- Liberty Street Economics on why credit card rates are so high. The short answer is a delight for finance fans: the performance of unsecured consumer loans to dubious borrowers is strongly correlated with the economic cycle, and there isn't a good direct way to hedge. It's not so much that the credit card companies need to make those high returns on average, it's that when they're losing money on unpaid credit cards, they're losing money on everything else, too.
- In Capital Gains this week: why markets crash faster than they rise, and why some of the S&P 500's best days are during its worst months. The Diff is not really in the business of timing the market, but this turned out to be well-timed: the S&P was down about 11% through the beginning of April when it was published, then had its 8th-best day on record by the close.
You're on the free list for The Diff. This week, paying subscribers read about why markets being excessively responsive to news during highly uncertain periods is actually an efficient market in liquidity ($), why it takes more than four years to lose reserve currency status ($), and why white-collar jobs may follow the blue-collar pattern in the next recovery ($). Upgrade today for more upbeat coverage!
Books
Fight: Inside the Wildest Battle for the White House: all through the last few months of the 2024 campaign, you could almost hear the knives being sharpened, and this book is one of the places they've come out. Business books are not the only genre where you get the best stories from the failures because everyone involved has free time, lots of gossip, and a burning desire to prove that it wasn't their fault.
The book opens with a sort of parable about the nature of modern political power. It's the night of the debate, and all across the country various people in their 70s and 80s are gearing up to watch TV alone in order to decide the fate of their party. There are adults in the room, yes, but they're eligible for Social Security and watching the same mass media you are. (At one point, Nancy Pelosi sends the wrong person a text, which is at one level understandable since she'd spent 80% of her life without a smartphone. But at another level, not so understandable that this is how important decisions get made.)
One level down, the campaign did have plenty of competent, get-things-done operators, and purely in terms of production values they did a really excellent job. But Biden's circle also did an excellent job of managing around his cognitive decline—they were a lot better than Team Trump at keeping their aging leader away from microphones when he was feeling improvisational, at least. It's very good to have organizational competency, like the Vice President's communications director, who carried around a spreadsheet of the nearest Republican-appointed judge in every location they'd visit in case the President literally keeled over during the campaign and Harris had to be hastily sworn in with as few constitutional crises as possible. It demonstrates a lot of competence at that specific task, and it's easy to imagine such people being very helpful to all sorts of different regimes.
This narrowly-scoped competence extends to how they communicate their deliberations to outsiders. During the period where Democrats were saying publicly that they stood by Biden, they were privately trying to figure out who'd actually run. This was a plan Biden himself blew up by dropping out of the race with no notice and naming Harris as his preferred successor. One of the responses to this was a fairly bland Medium post from Obama expressing hope that the party would quickly coalesce around a new winner. Which, as it turns out, was a kind of brutal knife-twisting move, punishing Biden for being salty that Obama hadn't backed him in 2016 and targeting Harris for being too aligned with Biden. But this is all between-the-lines subtext of what otherwise sounds like polite, professional communication! It's interesting to compare that to the winning administration, where the President's senior advisor on improving government efficiency recently referred to the President's senior counselor on trade and manufacturing a "Peter Retarrdo." You just don't need a behind-the-scenes book or a family tree of grievances to know who hates whom on that side.
The rest of the book is, frankly, less fun. A paradox of observing politics is that the more invested you are in the outcome, because you have strong views and it's a close contest, the harder it is to relate to those median voters who decide the outcome. The country is very close to 50/50, and sometimes the whole outcome hinges on who got the last gaffe in: Trump had a rally at Madison Square Garden in order to 1) drive up his popular vote total because he was confident he could win, and possibly 2) get a bunch of commentators to make a pretty flimsy comparison to certain events held in the same place generations earlier (nobody other than a full-time scholar of 1930s politics would instantly associate Madison Square Garden with "Nazi rally" as opposed to "Billy Joel" or "The Knicks," and I suspect that part of the political playbook, on both sides, is to do or say things that will strike most voters as innocuous but put a small segment of the other side into paroxysms of conspiratorial thinking. The message that the other side is a little bit kooky hits hardest when it comes from that side). One of Trump's speakers, a comedian, made a joke about Puerto Rico being a "floating island of garbage," and Biden responded—while one of Harris' key speeches was being broadcast!—by saying that "The only garbage I see floating out there is" either "his supporter's" or "his supporters," with the interpretation of possessive or plural depending, more or less, on who the person listening had already decided they'd vote for. So, at the last possible minute, everyone got one last round of potential recrimination in.
The skills that are required to quickly consolidate power in a party that suddenly has a vacuum of it are correlated with, but not identical to, the skills required to lead that party to victory in an election. And the skills required to win an election are also correlated with, but not identical to, the ones required to run the country. But the more competitive each stage is, the more everyone will be selected for the immediately-useful skill even at the expense of what they'll need at the next step.
A Word From Our Sponsors

The funds moving fastest in this market aren’t copy/pasting numbers—they’re clicking.
Top hedge funds have shifted how they build and update models, source financials, and prepare for earnings calls. They’re using Daloopa to cut hours and days of manual work into seconds and minutes — and they’re not looking back. If you're still building and updating models from Qs and Ks by hand, you're already behind.
If your team is triple-checking numbers, they’re losing precious time. And if you're not using Daloopa, your competitors are gaining that edge.
Daloopa gives you 1000s of audit-ready, hard-to-get fundamental historical data to either build models or automate model updates all in one click. No more wasted time. Just faster, smarter work so you can do your job even better. The smartest teams are already in. Don’t get left behind.
Start your free account now and see why everyone’s switching to Daloopa.
Open Thread
- Drop in any links or comments of interest to Diff readers.
- Do institutions like political parties naturally decay unless someone brings them back, or are there any that have been designed to be self-healing without needing someone to periodically blow them up and reform them?
Diff Jobs
Companies in the Diff network are actively looking for talent. See a sampling of current open roles below:
- Well-funded, fast-moving team is looking for a full-stack engineer to help build the best AI-powered video editor for marketers. Tackle advanced media pipelines, LLM applications, and more. TypeScript/React expertise required. (Austin, Remote)
- A well-funded startup ($16M raised) that’s building the universal electronic cash system by taking stablecoin adoption from edge cases to the mainstream is looking for a senior full-stack engineer. (NYC, Singapore)
- A Google Ventures-backed startup founded by SpaceX engineers that’s building data infrastructure and tooling for hardware companies is looking for full-stack and front-end focused software engineers with 3+ years experience, ideally with data intensive products. (LA, Hybrid)
- An OpenAI backed startup that’s applying advanced reasoning techniques to reinvent investment analysis from first principles and build the IDE for financial research is looking for a data engineer with experience building robust data infrastructure and performant ETL pipelines that support intense analytical workloads. (NYC)
- A company building the new pension of the 21st century and enabling universal basic capital is looking for a mobile-focused engineer who has experience building wonderful iOS experiences. (NYC)
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.