Democratizing Complicated Financial Products is Inevitable, but Fraught

In this issue:

• Democratizing Complicated Financial Products is Inevitable, but Fraught—The core problem is that financial products are mostly born as transactions between professionals who expect counterparties to be taken advantage of, and if they don't shake off that cultural norm before retail investors show up, retail gets hurt. But eventually, the equilibrium is that a transparent, low-fee option is the norm.
• Startup Fraud—Faking is less common when you don't know what to fake.
• Automation—What does automating more risk assessments say about the space of possible risks?
• Motivations—Some hackers are in it for money, but some are in it for fun, and the fun-loving ones at least help give society more immunity to bigger hacks later on.
• Ads—Generative ads make ad platforms even more of a black box.
• Utilization—Companies with scale-driven economics get to create their own M&A pipeline by continuously pushing down the price of their products.

The Diff June 2nd 2025 12:23 • Click to play audio

Democratizing Complicated Financial Products is Inevitable, but Fraught

The fundamental problem with regulating the average person's investing decisions is that individual investors are optimizing for basically three things:

1. They want to get good returns, for some definition of "good" that might range from "as high as possible" to "whatever ensures I never lose money in a given year.”
2. They want products to be simple, both in terms of what they're buying and how easy it is to buy it.
3. And, they don't want annoying know-it-alls who nag them about risk or tell them they're not allowed to do responsible things like buy GameStop at $420 a share or execute multi-legged options trades or exploit a bug in Robinhood's risk management system to get infinite leverage.

Obviously, it's pretty hard to reconcile simplicity, flexibility, and not either blowing all your money on bad trades or ending up losing half of your money to fees. (A fee of 1.86% annually on an investment with an 8% gross return is, over forty years, enough to reduce the final value of the investment by half in comparison to an 8% pre-fee return.^[1]

There's competition driving lower fees, but that competition doesn't take the economics 101 form where prices rapidly reach a price equals marginal cost equilibrium, because that econ 101 model is based on perfect competition. On the other hand, the financial sector is a many-layered lasagna alternating between the fiercest imaginable perfect competition (fill a 100-share market order for a large-cap stock) and ludicrously complex monopolistic economics (banks turning long-lasting real estate investments into a future stream of mortgage origination and credit card interchange revenue, or hedge funds trying to time the market in particular strategies by scooping up teams and locking them into long-term deferred compensation arrangements after a strategy blows up).

How this plays out in asset gathering is that the equilibrium is not just what the fee for this product would be if it were a commodity priced in order to give a just-barely-adequate return to whoever operated it. Instead, there are multiple equilibria, because a one basis point reduction in fees that comes directly out of the marketing budget can make the asset manager worse off over time. But there are many sales channels out there that respond to marketing spending in different ways. So there can be multiple equilibria, determined by weird zones of elasticity that are themselves determined by the marketing approach one layer up. One set of financial advisors might see their biggest value-add being tax planning and other kinds of personal finance advice, so for them going with a cheap index fund is ideal. But a wealth manager at a big bank will potentially have a menu of products that include a) exclusive opportunities the bank only wants to offer to a subset of their customers, and b) higher-fee stuff that justifies all of the costs of acquiring and serving their average customers. For them, a lot of the economics come down to the question of how willing the compliance department is to let something from category B sound like it's coming from category A.

For a while, one of the defining features of the asset management industry was that this was a nice equilibrium for advisors, since they could get a lot of revenue from their customers. But it also created an incentive to defect, and just offer something so transparently cheaper than the rest of the product offering menu that it was ludicrous not to take it—this was a winning move in that there seem to be more recently-minted passive investing billionaires (Larry Fink and Dimensional's cofounder David Booth) than among active managers with publicly accessible vehicles (there are plenty of hedge fund and PE billionaires, but among the self-made billionaires who got that way actively managing funds, the impressive results skew more to bonds (Gross, Gundlach), while the equity fortunes were mostly made earlier.^[2]

In the pre-indexing days, when funds were smaller and markets less efficient, one of the most glaring market inefficiencies was how many talented people were working for management fees instead of carry. In the 80s, middle-class savers could and did pay Peter Lynch to pore over the financials of hundreds of S&Ls in order to pick the best ones for them! You could pay Mario Gabelli himself to read John Malone's latest Form 10 and figure out which side of the spinoff to own! You could pay Bill Gross to harangue his traders so they could bully their counterparties into letting them sweep the street.^[3]

One of the things that stands out from rereading The Snowball is that the divide between a public company and a private one was differently fuzzy than it is today. Today, the fuzziness comes from the fact that if a sufficiently high net worth investor wants to own some SpaceX or Databricks, there are people they can call to accomplish this^[4], whereas it would have been harder to buy into a private growth company in the 1950s. When Buffett was trading stuff like AT&T or Jones & Laughlin Steel in his fund's portfolio, he was doing something that looks recognizable to a typical investor today. But when he was buying streetcar stocks or tiny insurance companies, he was doing things like placing classified ads to ask people to sell literal stock certificates, or looking up old boards of directors in order to identify which towns would be a good place to go door-to-door hunting down stock.

There's a lot more access to equities like that than there was in his day. If you're looking at a company of similar scale to Sanborn Maps (~$50m in market cap, adjusted for inflation), there will almost certainly be a ValueInvestorsClub message board, and a few tweets with the relevant cashtag. It won't be hard to track down other shareholders, if you're committed. In this sense, access is much more democratized. We've hollowed out the middle of assets where you can trade, but it's hard—now we either have assets where being able to trade them is a boolean (ISDA, accredited investor status, willingness to use crypto), but we have fewer asset classes where trading is possible but time-consuming. And even when we do, it creates demand for some easily-accessible vehicle whose pitch is that it offers exposure to some asset class that's otherwise hard to reach.

When you're bridging these two worlds, the world of individual savers who want transparent access to understandable investments, and the OTC world of people whose moral standard is something like "what's the optimal pace at which to rip this person off?" you necessarily experience friction. Some of this can be surmounted: commercial real estate, for example, is rife with zero-sum opportunities, but the longer someone wants to work in that space, the more likely they are to play an iterated game.

There are some transactions that are naturally adversarial, and it's hard to regulate them out of existence—some of these transactions are quite enticing, and almost all of them are opt-out. When a financial product is fully commoditized, that pushes the adversarial element down towards zero, so for many financial interactions—saving through a tax-advantaged account, indexing, using a credit card and paying off the balance every month—a typical member of the middle class straightforwardly wins. But that's just not the end state of consumer finance for people outside of the 0.1%, as evidenced by the fact that this mix wasn't always the historical default. Until the mid-twentieth century, equities were just not seen as an accessible investment option—and the transition to equities being a normal asset class was full of scams and wildly overpriced financial products.

It's hard for an asset to be accessible to retail investments before professionals are involved (crypto might be the only meaningful counterexample). And the intra-professional norm is that trades happen between adversaries, each of whom is happy to take advantage of the other side. When financial democratization goes badly, it's often because this mercenary attitude persists after getting access to a larger, and more naive, cohort of counterparties. But that's typically temporary. At the current scale of alternative assets, it's just implausible to imagine middle-class savers shifting enough of their assets to saturate the available opportunity set. Which means it behooves their counterparties to think a few moves ahead and not rip everyone off right away.

Markets tend to resist segmentation: they don't like to remain either all-amateur or all-professional for long. When professionals are setting prices, the market tends to be a boring way to discount the known present—which makes it a cost-effective way to bet on the unknown future. In the other direction, markets that are mostly populated by retail investors typically don't clear, or experience wild price swings, because there's no one around in the business of warehousing risk. So these markets tend to bleed into one another, that transition is never easy.

Diff Jobs

Companies in the Diff network are actively looking for talent. See a sampling of current open roles below:

• 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 software engineers and a fundamental analyst. Experience at a Tiger Cub a plus. (NYC)
• A Google Ventures-backed startup founded by SpaceX engineers that’s building data infrastructure and tooling for hardware companies is looking for a product manager with 3+ years experience building product at high-growth enterprise SaaS businesses. Technical background preferred. (LA, Hybrid)
• YC-backed startup using AI to transform how companies quantify and optimize engineering productivity is hiring formidable full-stack and AI engineers. Experience with React + Typescript, Go, or Python on the ML side a plus. (SF)
• A multi-stage, fintech focused investment firm is looking for an investment associate to support thematic opportunity identification, diligence, and execution. Investing experience OR high-growth operating/investment banking/consulting experience and demonstrated interest in fintech required. (NYC, London)
• A hyper-growth startup (10x growth in 9 months) that’s turning customers’ sales and marketing data into revenue is looking for a head of deployments who is excited to work closely with customers to make the product work for them. Experience as a forward deployed engineer and leading enterprise deployments preferred. (SF, 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.

Elsewhere

Startup Fraud

The vague questions about why Builder.ai suddenly collapsed recently got more specific, after the revelation that they'd round-tripped revenue with another startup. It's a testament to the venture industry that this kind of literal accounting fraud is so rare, given that VCs invest in businesses at a scale where accounting fraud would be incredibly easy to pull off.

One of the things that makes venture uniquely resistant to accounting frauds is that it's hard to tell which numbers to fake. The more unique a company, the more there's a need for custom metrics, and the less the valuation is dependent on accounting profits. So small-scale fudging probably still happens, but wholesale fraud raises enough questions that it's not worth the effort.

Automation

Meta plans to offload more of its risk evaluation for new features to AI. One of the interesting things about AI is that it's hard to treat the output as a special case, whereas for human judgments it's always possible to imagine all-too-human misjudgements. Deferring more arguments to AI is a way to reduce the second-guessing inherent in edge-case moderation decisions. What this really does is allow Meta to choose how big a dose of reasoning to apply to a problem. Risk assessments are basically an iterated process of repeatedly asking "and then what might happen?" and reasoning models are getting increasingly good at following this kind of intellectual rabbithole as far as it can go.

Disclosure: Long META.

Motivations

The FT has a piece on the hacker consortium that targeted Marks & Spencer, noting that they're less interested in money than in causing chaos ($, FT). It's a good reminder to expect an increasingly long tail of inscrutable mayhem ($, Diff).

But it also raises a fun question: the threat model for hacking covers a range of outcomes, from minor digital vandalism to extortion to shutting down critical infrastructure. For the last category, sample sizes will be small, because state-sponsored attackers will tend to hold back some exploits and then use them all at once. So there's something hormetic about these comparatively minor hacking incidents: at least they remind everyone to stay on their toes.

Ads

Meta wants to use AI to automatically generate ads and then target them ($, WSJ). Making advertising more of a black box has been in the big platforms' interests in two ways: first, it means that they have more control over the data, and that they can keep users' data to themselves while still using it to target ads. And second, the harder it is to understand how the ads get in front of customers, the more pricing power the ad platform has. Generative ads are a natural extension of this: advertisers are even more locked in to whichever platform they use if the platform supplies the creative, too. And, of course, the biggest platforms will collect more data on how to make the best generative ads, so the opportunity cost of opting out of black-box platforms keeps rising.

Disclosure: Long META.

Utilization

DoorDash has been scooping up smaller delivery competitors. Food delivery tends towards commoditization, because the brand names customers care about are owned by the restaurants, not by the platform. ("I don't care if we have sushi or pizza for dinner, as long as it gets delivered by UberEats!") On the other hand, it delivers economies of scale because of the same black-box ad targeting Meta uses, as well as relentless math of network utilization: the more drivers and customers there are, the more of their time drivers spend on revenue-generating trips, and if the market's pricing for delivery is set by the lowest-cost provider, smaller companies get continuously priced out of having viable economics. But they still have customers, drivers, and restaurant relationships, and all of those are worth more when they're tied to a bigger network.