The next cross border rail will be 10x more efficient

Welcome to Fintech Brainfood, the weekly deep dive into Fintech news, events, and analysis. You can subscribe by hitting the button below, and you can get in touch by hitting reply to the email (or subscribing then replying)

Hey Fintech Nerds 👋

What a week!

The week Founder Mode took over the internet, Paylocity acquired Airbase, and the FDIC reminded banks that their job is to know where customer funds are.

The FDIC news isn’t surprising, but it’s sad they had to come out and say it.

Airbase, the HR and payroll software, will be acquired for $325m by Paylocity.

Chase will auto-enroll every customer into its Paze wallet. I’m saving my thoughts on this for a rant. Too much to say.

Revolut made $500m from its business customers. Fintech companies serving both consumers and businesses are crushing.

Oh and this week, I learned that Founder Mode is alive and well in Saudi Arabia.

A Neobank has acquired 1.5 million customers in a month. It has full regulatory approvals to become a bank. You need to take a look at Barq (things to know).

This week, I’m reflecting on a podcast conversation that made it clear to me that a new cross-border payments rail is emerging. Unlike SWIFT or the local networks used by companies like Wise and Revolut, this new rail solves the last-mile problem. That’s your Rant this week.

Here's this week's Brainfood in summary

📣 Rant: Stablecoins the next platform for embedded finance?

💸 4 Fintech Companies:

1. Dub Prime - The Explainable AI Underwriting Model

2. Moneystack - Mental Health Support for better Money outcomes

3. PostEx - Sharia-compliant revenue-based finance

4. Magie - The Chat-First Neobank on Whatsapp

👀 Things to Know:

1.  HR and Payroll platform Paylocity will acquire Airbase for $325m

2. The worlds fastest growing Neobank hits 1.5m users in 30 days.

📚 Good Read: Why is it so hard to lock your credit report?

If your email client clips some of this newsletter click below to see the rest

📣 Weekly Rant

Stablecoins the next platform for embedded finance?

Health warning: This one is rambly. I'm still mapping this all out

There’s an alternative financial account, payments, and cross-border Infrastructure. It is quietly gaining traction with major PSPs, payments companies, and non-banks. Today, the use cases are long-tail and at the edge, but that's often true about many disruptive innovations.

Stablecoins are an alternative to local payment methods and SWIFT as a cross-border rail.

Most payment professionals still rank Stablecoins relatively low on their priorities.

Yet, they rank global acceptance and the challenge of being international as the highest or near highest.

I put it to you that you're systematically not paying enough attention to the weak signals in this market because of personal biases, lack of usefulness in your context, or distaste for anything with a proximity to Crypto.

That's a recipe for disruption.

And it's tasty.

Here's my summary

1. Context: How we move money today, especially cross border is high friction, high cost and error prone (as we saw with Evolve / Synapse)

2. Recap: What's a 1:1 backed Stablecoin and why is it different to holding a dollar deposit at a bank?

3. Stablecoins offer a compelling business case. They're a low cost, US dollar rail that solves for the last mile in a way other payment methods struggle to.

4. However they're hard to use unless you're a payments company or expert user, their utility for affluent consumers is limited.

5. The three (non-trading) use cases getting traction are

   1. P2P (e.g., Sling Money)

   2. Payouts (e.g., Payoneer but with Stablecoins)

   3. PSP to PSP. (Stablecoins as an invisible rail)

6. We will soon get regulatory clarity and easier to use Infrastructure unlocking new use cases and scale

7. There's a case for Stablecoins as an alternative embedded finance platform but it's going to take a while

8. I'm excited by the idea of Stablecoins as a platform for GenAI Agent-commerce. Card rails are too expensive, closed loops are too closed. 

1. Quick Recap - How we move money across borders today vs Stablecoins

When sending money abroad there's two fundamental mechanisms

1. "International Wire" AKA - A SWIFT Payment.

2. Direct connectivity between local payment networks (e.g., ACH to Pix) is how services like dLocal, Wise, and Revolut often offer lower fees.

With SWIFT, a transaction can cost between $30 and $120, and money settles in minutes to weeks*. So it can cost $100 to send $100.

• You send an instruction to your bank to move money from your bank account to your recipient's account at their bank.

• Your bank sends a message to the recipient's bank saying, "Hey, I want to send you $100 to user JOE with account 12345.

• The recipients bank sends a message bank saying, "Ok, go ahead, that's a good customer of ours."

• Your bank and your recipient's bank then perform a detailed set of compliance checks. These can include AML, Sanctions, PEPs, and adverse media. They also have to comply with country-specific tax rules.

• Your bank then removes $100 from your account, plus your $30 fee. It then adds $100 to the recipient bank's Nostro account.

• The recipient bank sees the $100 in their Nostro and credits your recipient account the $100 (less any fees or FX charges)

Does this sound complex?

Moving money is a bit like moving the London Bridge to South Carolina. A heavy lift.

Oh, and I left out about 30 different subprocesses and didn't bother explaining the nature of 

Nostro and Vostro (but here's a handy link).

* It costs so much because every SWIFT transaction is subject to full sanctions, AML, and anti-bribery checks. A bank will perform a similar level of due diligence on a $100 transaction than on a much higher transaction. They use the same systems, teams, and processes in many cases. An actual SWIFT message costs a fraction of a penny.

When we directly link local payment networks the process looks different.

With a service like Wise or Revolut, the fee is often much lower (as little as $1 and much thinner FX spreads)

• You enter the receiving account details into your app to move money from your wallet to the recipient account.

• Your app shows you estimated fees, and you can confirm them by hitting send.

• Your app removes $100 from your wallet (plus any fees)

• The money then sits in a local account in your country (e.g. the USA)

• The app then moves money they have in your recipient country (e.g., India) into your recipients account or wallet using a local payment rail (e.g., UPI).

Two important things here.

1. Money never moves across borders; in most cases, it stays in the country. This creates the illusion of real-time payments and low fees via "netting."

2. It looks like your customer received a domestic payment to the receiving bank or wallet. A bank may not perform the level of checks it would have if it had received a SWIFT payment.

Local payment networks are typically much cheaper but don't require the level of due diligence that SWIFT payments do. This is why the banks (rightly) complain that some of the apps create significant AML risk*.

* As a practical reality, most app services also use SWIFT to batch many payments together, just like banks do. It's rare banks would send a 1:1 payment for every instruction

The pros of SWIFT are that it's transparently a cross-border payment to everyone involved in a transaction, and it's accepted everywhere. Expensive, sometimes slow, but comprehensive and industrial grade. It's also an arm of the US and EU government's policy via Sanctions.

The pros of domestic clearing payments are lower cost and feel instantaneous. However, it's not as widely accepted and creates significant risks for AML and Sanctions policy by appearing as a local payment in many cases.

Now add that the global south countries are building their own BRICS currency rail to compete with SWIFT, and the need becomes clear.

We need a comprehensive, global, 24/7, affordable cross-border rail.

Enter: Stablecoins.

2. Quick Recap - What's a Stablecoin?

Stablecoins that are 1:1 backed by US Dollar equivalents (like USDC, PYUSD, etc.) are an alternative cross-border payments rail. For every Stablecoin, there's $1 held at a bank or in US Treasuries at a custodian. 

These custodians include some of the world's largest, like BNY Mellon and State Street, which have $44.3 trillion and $36.7trn respectively. This is a crucial detail because it means we move away from the SWIFT model of sending messages to move balances to a new model that resembles how other assets are transferred.

(There are many competing alternatives, like M^Zero that offer compelling alternative visions, but for simplicity, lets play this out for now)

Custody matters, and is an alternative way to store dollars vs holding it in a bank account. It's just not generally available by default to consumers, payments companies or embedded finance use cases as it is with other assets.

Today, most physical gold never moves; it stays in a vault and is in custody. Many assets are also held by custodian banks, which change the ownership records and ensure their safety.

Circle, for example, holds the vast majority of its "real" US Dollars in a Blackrock Money Market fund. In turn, Blackrock safeguards that fund across the global custodian banks.

Why does all of this matter?

Once you have your traditional dollars "locked," your token dollars (USDC or PYUSD) can move between wallets in real-time, 24/7, and globally. These transactions happen on a single, global ledger, so we never have reconciliation issues. And, if KYC'd, we know exactly which entity has which balance and where that balance came from.

Stablecoins rock!

If they get adoption, they'd also solve crucial payment challenges like the cost involved with SWIFT and the last mile problem we face in payments.

3. Stablecoins offer a compelling business case.

a) Stablecoins solve the biggest problem in payments: The last mile. The industry operates on a hub-and-spokes model, where large corridors like the US to Europe are fast, cheap, and instant. Things are great so long as you go via the US dollar. If you're in the global south, trading with other global south nations, this is less so.

Stablecoins, by their nature, are instant and work anywhere with compatible software. So long as you have internet access and a network-compatible wallet, you can interact with Stablecoins.

You push USD along this global, instant, very cheap rail, and they arrive in the country. You only convert them when you need local currency.

b) Stablecoins are cheaper than SWIFT but a cross-border transaction when converting to local currency. If you can find a local exchange (like Bitso in Brazil), you can buy USD-denominated Stablecoins and send them locally or abroad, or you can just hold on to them. That conversion then becomes a cross-border transaction (albeit in many markets that might lack regulatory clarity)

This means a user or business in the global south could

• Get paid instantly for low fees

• Pay internationally instantly for low fees

• Hold USD instead of their local currency

c) Digital dollars have traction in the global south. Companies like Airtm are creating simple wallets and enterprise payout solutions. They're a mix of a local and Wise-like service. They offer consumers a local wallet and enterprise payout solutions. Global citizens in high-inflation markets get dollars they can transact simply, 24/7, and for low fees. Enterprises get a payout solution that works in 190 markets.

d) Stablecoin liquidity and supply is growing. The supply of fiat-backed (1:1 backed by US dollar or US dollar equivalent) Stablecoins has reached an all-time high of $165bn. Crypto has been trading sideways, so this isn't driven by mania. Even if you assume 90% of that is used for trading, then $16.5bn has other use cases.

If you assume

• It solves for the last mile

• Liquidity is improving

• Regulatory clarity is coming

• It's more affordable than SWIFT

• Potentially(!) more compliant than connecting local clearing networks

The practical issue is the last mile isn't solved.

At least not yet.

4. Today: Stablecoins are too high friction for most users

If you send someone $100, you generally don't have to ask which dollar flavor they'd like.

If you want to send someone $100 in Stablecoins you might have to ask,

• Which dollar? USDC? PYUSD? There are many flavors of Stablecoin from different issuers, and not all wallets support all tokens.

• On which network? Are you using Solana, Ethereum, Lightspark, or one of the Eth L2s like Base?

• Which wallet? Not all wallets support all networks. Some are more closed-loop in nature or use proprietary tokens.

• Do I have the regulatory clarity? If you're a larger corporation, you might not want to touch Stablecoins (e.g., in the US) because it's unclear if you can.

The default way to access Stablecoins is also confusing. You have to take your regular dollars, buy digital dollars, add those to a wallet, find the address of the person you want to send to, and then hit send.

Exhausting.

These are all solvable problems, and they will be solved in the next 2 to 5 years. The biggest issues are negative perception and the lack of perceived value.

If you don't live in the global south, transact globally, or seek instant, programmable payments, Stablecoins seem pretty useless—doubly so when they're high-friction.

Yet I remember when Revolut appeared in London a decade ago, and the mainstream payments reaction was, "These guys are idiots; they must be losing so much money." And "We could do this. It's just netting."

Disruption always starts at the edges.

A global, 24/7 US Dollar rail could unlock new use cases, and it already does. Just maybe not for you yet.

5. Three ways Stablecoins are gaining traction

There are three core ways Stablecoins are gaining traction.

1. Wallet to wallet (P2P). where one user of a wallet (e.g., Metamask) sends to another user of a wallet (e.g., Coinbase wallet). This is typically for Crypto natives or those sufficiently motivated (like tech-savvy Argentines or Nigerians). For the mass market, it's high difficulty. It's confusing and not as widely accepted as many domestic payment rails.

2. Payouts: When scenario, a corporation sends a payment like payroll on traditional rails, but the user receives a Stablecoin into a compatible wallet. This is happening today via payroll platforms like Deel. The enterprise has fewer regulatory or treasury changes to make, but the user gets paid in a way they benefit from. Think of it as Payoneer being on another rail.

3. Stablecoins as the hidden rail (AKA The Stablecoin sandwich). This is when a payments company (or multiple payments companies) settle in Stablecoins between each other, but the sender (e.g., a corporate) and the recipient (e.g., a consumer) send and receive fiat. This removes difficulty for both the sender and receiver, while the PSP can pass on the efficiency and cost savings for competitive benefit.

The question of which dollar, on which rail, and using which wallet is almost entirely solved by payment companies adopting Stablecoins as a rail.

Considering that the vast majority (> 60%) of SWIFT transactions use the US Dollar, the dollar's role as a bridging currency is well-proven. Having that work via another, internet-native, 24/7 rail is just as powerful.

Considering that the global south is trying to move away from the US Dollar due to

1. The high cost for global south to global south trade (e.g. LATAM to Asia)

2. The geopolitical risk of being sanctioned by the US

There's a need for a US Dollar rail that at least solves the first problem. The US Dollar has the technology to be the most competitive, widely adopted trading base and rail.

If it can solve a few more infrastructure challenges.

6. The Infrastructure will get fixed, and regulatory clarity will follow.

If you're a payments company or wallet that wants to adopt Stablecoins you have a high barrier to entry. The first question for every PSP is, "If we do this, will our bank partners cut us off or start de-risking." There's a constant balance of keeping bank partners on-side. The second question is, how do I make this work with my tech stack?

There are infrastructure companies that solve Stablecoin's complexity for wallets and payment companies. If you don't want to go directly to Coinbase, Circle, Fireblocks, Anchorage, etc., then companies like Bridge package this into neat APIs.

Europe is showing regulatory clarity is possible, which will follow in the US in the next few years. The MICA regulation has allowed banks like Banking Circle and Societe Generale to offer their own Euro-denominated Stablecoins traded on exchanges.

The US now has a Bitcoin and Ethereum ETF. How did the latter go from controversial to non-event in the space of 12 months? The mood is shifting. We've come close on Stablecoin regulation several times, and the next administration will likely pass something.

I've always been convinced that if Stablecoins don't die, they inevitably become a major part of the payments landscape. My experience in payments tells me that we move in 7-year increments.

7 years ago Apple Pay was questionable at best. It's now a major market player. 7 years ago UPI and Pix weren't in the conversation, now they are the conversation.

Today, Stablecoins are at the edge and useless to most Western founders, VCs, or operators.

That will change.

7. Stablecoins for Embedded Finance and AI Agents.

I have two questions for you.

1. What if we never had a lingering issue like Synapse and Evolved ever again?

2. How will AI agents pay for things?

The fundamental problem with the Synapse, Evolve bankruptcy saga is reconciliation. Nobody has a complete record of which customer had which balance and where the money is. There isn't a single source of truth, a single ledger.

Stablecoins fix this.

(Something many Crypto skeptics have told me privately but felt a cognitive dissonance in the process).

• Stablecoins backed 1:1 against US Treasuries at a global custodian do not have a safety and soundness issue.

• Balances tracked by tokens against wallets in a blockchain do not have a ledgering issue.

We know where the money is and which wallet has which balance. The challenge for this model is acceptance. There aren't many places to spend and use Stablecoins as consumers or businesses. But then why should you?

Imagine you're building a consumer Neobank or embedding a credit card into your SaaS platform. If that card can buy things, and if you can manage its expenses, do you care how the Infrastructure works? You likely care that you can get high yield, make and accept payments, and get competitive loans. But nobody thinks about the Infrastructure until it goes wrong.

The big gap right now is that Stablecoins themselves can't attract yield (for fear of being a security, which is a fair reading of the law). But if Stablecoins are simply a settlement mechanism, a rail, you can still have other products. Stablecoins are just how you move the balance.

AI Agents could do more economically useful things if they could pay and get paid. The default for making a payment is to give them an account and a virtual card. This is not a horrible idea (AI + virtual card can buy certain things but is restricted by spend management).

But lets assume you want them to buy and sell from each other at high volume and low transaction values. Cards quickly become expensive, and even bank rails become slow and pricey.

Ideally, every agent should have a wallet, a simple set of instructions, and an ultra-low-cost settlement infrastructure.

This sounds really pie in the sky, but here's a practical example.

• Imagine your AI ComplianceAgent crawling the dark web for dangers and coming across some text. However, it can’t understand the internet slang, and it’s in another language.

• It could find and pay an expert translation agent $0.0075 to ensure the highest-quality translation.

• SlangAgent pays its underlying LLM service providers $0.0005 (of which there are say, 10). Costing it $0.0050.

SlangAgent just made $0.0025 profit!

This mixes identity (discover trusted LLMs), with nano-transactions (fractions of a penny) and it runs at the speed of software, not legacy payments infrastructure.

Skyfire does exactly this.

And it uses a Blockchain to do it.

Could they use an FBO account (suspense account) and manage all this with virtual accounts? Yes, in principle, there's no reason why not.

The advantage of a Stablecoin is that it’s open loop, and combined with identity, any agent could discover any other agent. It’s also much more amenable to nano-scale transactions.

If agents are to have agency, they need software-defined money. Stablecoins are exactly that.

They just happen to be global, 24/7, and increasingly more stable than dollar deposits in a bank.

Summary

Stablecoins are an ultra-low-cost, 24/7 payment rail that has already been adopted by a large number of users. However, it is currently held back by negative perceptions, a lack of regulatory clarity, and a high difficulty curve for the early use cases.

What's coming now is Infrastructure and regulatory clarity. This isn't happening as fast as bullish advocates would have you believe, nor is it a failure to launch, as the cynics will tell you.

The US Dollar rail, which settles instantly with no safety and soundness risk, will be a game changer. My only question is when.

Will it displace embedded finance? Not anytime soon, but it's a compelling idea someone could explore.

Will it unlock AI agents? Possibly. The idea is intriguing enough that its worth exploring.

However, it will almost certainly evolve as a strong competitor to SWIFT and local clearing networks, especially for global south US trade routes.

Like everything in payments, it's slowly, then suddenly.

Look for the weak signals, the new use cases, and the new markets.

That's where the future is.

ST.

4 Fintech Companies 💸

1. Dub Prime - The Explainable AI Underwriting Model

Dub Prime is an AI credit model that looks for more data points about a borrower to help banks confidently to sub 700 credit scores. Dub Prime says only 48% of Americans can access prime credit, despite 80% having never missed a payment.

🧠 This is Fintech-like cashflow and alternative data for underwriting but for credit unions and small banks. These smaller organizations often struggle with tech, so this package is neat. They're leaning hard on ethical lending and explainability, which makes sense. There's a lot of fear about AI out there. PS. I love the name. It's not subprime; it's dub prime. Yas.

2. Moneystack - Mental Health Support for better Money outcomes

Moneystack works with state-funded gambling and mental health support groups to improve financial outcomes with technology. The software platform is integrated with mental health treatment. They're initially starting with problem gambling and sports betting and helping users turn their finances around. Over time, they'll package more financial control and financial products into their offering (but their primary client is hospitals and support clinics)

🧠 This kind of thing reminds you that Fintech is a very good thing. Financial distress worsens health conditions, and mental health conditions worsen financial conditions. These things are linked and need to be treated accordingly. Moneystack provides financial counseling on behalf of the clinics, reducing patients' churn. The combination of financial counseling with mental health counseling is ultra-smart. I love this.

3. PostEx - Sharia-compliant revenue-based finance

PostEx is primarily a logistics business that offers advanced payment for "cash on delivery" if the seller uses its courier service. PostEx helps digital businesses in Pakistan and the GCC finance their business and adjust repayments based on their revenue. The company also offers a business suite with spend management tools, manages vendors, and offers instant payouts for sales.

🧠 This is the best of digital for cash-heavy markets. The GCC and the Middle East have a burgeoning digital economy that is held back by a lack of financing. This type of financing could unlock a significant untapped potential for business creation to "grow the GDP" of not just the internet but the local economies. Love this.

4. Magie - The Chat-First Neobank on Whatsapp

Maggie acts as a personal assistant and AI as a Whatsapp agent. Users can manage bills, schedule payments, and turn a picture of an invoice into a payment via Pix. It acts as a Pix wallet native to Whatsapp. It offers additional passwords and step-up authentication for payments. The service also offers 100% of the CDI (equivalent to the Fed Funds rate) for cash held on deposit, which is market leading.

🧠 RTP + Chat Apps are the natural evolution of money in most markets. Chat is the natural home of money movement; you can see this in WeChat in China. The addition of bots + AI makes this experience much slicker and more available. I'm super curious to see if this catches on. Meta has tried and failed to stick the landing on payments for the last 15 years, but perhaps the answer is Neobanks as agents.

Things to know 👀

1. HR and Payroll platform Paylocity will acquire Airbase for $325m

Airbase, the platform for CFOs, will be acquired for $325m by Paylocity. This is good for the ecosystem; we’re getting complete ERP competitors to SAP and Oracle in much more modern software.

🧠 Great timing by Paylocity to acquire. Paylocity revenues in June 2024 were up 15.8%, with net income up 31% to 48.2M. As a cashflow-generating business, it can pick up strategic additions.

🧠 Airbase is a small piece of a bigger puzzle. The release says the Airbase represents 1% of group revenues by 2025. With June 2024 revenues at $357m, assuming $1.42bn run rate, that would make Airbase revenues in the ~14.2m range.

🧠 You either live long enough to compete with SAP and the banks, or you die a hero being acquired by someone who is. Airbase was a well-liked, well-put-together product that struggled to stand out in a crowded field. Paylocity is a growing business with a well-oiled sales and marketing machine. You’d imagine the complementary capabilities will get a lot more volume as part of Paylocity than they would alone.

🧠 The CFO stack is turning into a battle for the next mega ERP. Paylocity's addition of Airbase gets them the accounts payable, invoicing, and finance operations part of the stack they were missing. When you zoom out, companies like Rippling are also heading this way, while Ramp is coming back the other way.

🧠 The time is now for M&A. Corp Dev is lit. There are sooooo many companies that raised big in 2021, not there on profitability but with great products and traction. With the right homework, there are great deals to be done.

2. The worlds fastest growing Neobank hits 1.5m users in 30 days.

Barq in KSA hit 1 million customers in 21 days, making it the fastest-growing Neobank or wallet in history outside of China. Barq is a digital wallet in the Kingdom of Saudi Arabia (KSA). It took Revolut a year to hit 100,000, and while UPI and Pix grew faster, they were government-mandated. Barq is 100% market-driven.

🧠 My gut reaction was - what’s the catch? From the country that brought you The Line, massive sporting events and a monorail nobody uses, the temptation is to write this off as just another white elephant in a nation of big government projects. I was wrong.

🧠 Founder mode - Saudi style. The CEO of Barq Ahemd Alenazi would be a great founder in any geography. I had an hour with the guy. He’s a beast but a Gentleman. At 27 he built the largest syndicated lending desk in the country. He left and became the CEO of STC Pay and helped drive that to 10m users. He’s taken his team, his lessons, and relentless energy to what’s next.

🧠 Great infrastructure, great relationships, terrible competition. The ingredients are simple. The KSA has government-mandated eKYC, but most big banks barely use it, and their tech doesn’t work. Barq is built cloud-native and by a CEO who gets UX, design, and growth hacks. There’s a low bar.

🧠 An expat population in need of solutions. There’s also a massive ex-pat population. 27% of the population is ex-pat, making remittances, and the standard solutions are poor.

🧠 Founder market fit. Getting all of that stood up is hard, but the Barq team has done it before. Ahemd built remittances before at banks and has been a CEO who knows regulators before. That helps, but it’s not enough alone. This is an ambitious guy who’s trying to build the “Nubank of MENA.” Lets see.

🧠 The quiet part out loud. This company stayed in stealth for months while it got its central bank licenses in place. So, 0 to 1.5m in 30 days is from launch, not founding. I also have a sneaky suspicion they have distribution via a big government employer like Aramco. None of these things are bad; they’re distribution.

PS. While a friend introduced me to them, and I happen to be in KSA as I write this, I genuinely believe this is a legit business. The founder showed me their dashboards.

Good Reads 📚

1. Why is it so hard to lock your credit report?

Alex laments, trying to lock his credit file after another round of data breaches. He manifests a mobile app that gives free credit reports, instant report freezing, push button opt-outs of "pre-screen offers," and initiates report file disputes.

🧠 I want this too, Alex. I want this, too.

Tweets of the week 🕊

That's all, folks. 👋

Remember, if you're enjoying this content, please do tell all your fintech friends to check it out and hit the subscribe button :)

(1) All content and views expressed here are the authors' personal opinions and do not reflect the views of any of their employers or employees.

(2) All companies or assets mentioned by the author in which the author has a personal and/or financial interest are denoted with a *. None of the above constitutes investment advice, and you should seek independent advice before making any investment decisions.

(3) Any companies mentioned are top of mind and used for illustrative purposes only.

(4) A team of researchers has not rigorously fact-checked this. Please don't take it as gospel—strong opinions weakly held

(5) Citations may be missing, and I've done my best to cite, but I will always aim to update and correct the live version where possible. If I cited you and got the referencing wrong, please reach out