Complex Systems with Patrick McKenzie (patio11)

Patrick McKenzie (@patio11) is joined by Alex Elliott (@alexpotato), who has spent his career in SRE and operations roles at financial firms ranging from ten-person fintechs to global investment banks, to discuss what the back office actually does all day. They cover the life cycle of a trade: execution, clearing, settlement, and the constant reconciliation between counterparties who don't quite agree on what happened. Alex recounts the Friday afternoon when twenty million shares went out to the street instead of five, and Patrick explains what happens between 4:15 and 4:30 when one bank owes another $36 million and the person who owns that process is on vacation.

–
Full transcript available here: https://www.complexsystemspodcast.com/two-banks-cant-agree-on-fourth-grade-math/

–
Presenting Sponsors: Mercury, MongoDB & Granola

Complex Systems is presented by Mercury—radically better banking for founders. Mercury's new feature Command brings an LLM directly into your banking interface, so checking balances, finding invoices, or sending a wire is as easy as asking. Apply online in minutes at https://mercury.com/. 

If meetings consistently leave you with hazy action items and lost context, Granola handles the transcription so you can actually participate and gives you searchable notes afterward. Try it free at granola.ai/complexsystems with code COMPLEXSYSTEMS

What's the point of building faster with AI if your database can't keep up? MongoDB's native data model mirrors the language LLMs already speak. Ship at the speed of AI while staying ACID compliant at Fortune 500 scale. Start building at https://mongodb.com/ai.

–

Links:

• Alex’s website: https://alexpotato.com/ 

–

Timestamps:

(00:00) Intro
(01:48) The life cycle of a trade
(04:37) Why you don't know who you traded with
(06:01) Hiding a large block trade
(10:04) Information walls inside financial firms
(12:49) What the middle office does: managing positions
(15:41) Calendars, time zones, and other sources of gibbering madness
(19:37) War story: twenty million shares out the door
(22:45) Sponsors: Mercury | Granola
(25:50) Ops as an iterated game
(28:53) Busting trades and who has the authority
(33:38) Sponsor: MongoDB
(34:29) Inside the Knight Capital collapse
(35:49) Agreements, obligations, and who you have to call
(36:57) Reconciliation, FTP files, and the humans behind them
(39:16) Deadlines, cutoffs, and the $36 million phone call
(44:26) How two banks fail to agree on fourth-grade math
(48:54) When the golden record goes down
(51:55) Finding the one person who can sell the position
(54:09) How much business do we do with Fidelity?
(56:38) Prime brokerage and extending credit
(01:01:03) Smart flow, dumb flow, and payment for order flow
(01:02:45) Wrap

Patrick McKenzie (patio11) is joined by Manish Goregaokar, a senior software engineer and member of the Rust security response team, to discuss what LLMs do to the cost side of running a sophisticated con. They walk through a scam built for programmers: a plausible company, a few rounds of interviews, and an NDA signed via "sign in with email" that quietly hands an attacker the identity account everything else resets against. A version of it hit at least five people in the Rust community within days of Manish publishing on the subject, and Patrick describes being targeted himself by someone posing as a Bloomberg reporter who wanted him on a Zoom call with a codec problem. They also cover money mules and the three stages of laundering, the $25 million Hong Kong transfer authorized on a fully deepfaked video call, and why the only durable individual defense is refusing to trust any communication you did not initiate.

–
Full transcript available here: https://www.complexsystemspodcast.com/what-llms-do-to-the-economics-of-scams-with-manish-goregaokar/

–
Presenting Sponsors: Mercury, MongoDB & Granola

Complex Systems is presented by Mercury—radically better banking for founders. Mercury's new feature Command brings an LLM directly into your banking interface, so checking balances, finding invoices, or sending a wire is as easy as asking. Apply online in minutes at https://mercury.com/. 

If meetings consistently leave you with hazy action items and lost context, Granola handles the transcription so you can actually participate and gives you searchable notes afterward. Try it free at granola.ai/complexsystems with code COMPLEXSYSTEMS

What's the point of building faster with AI if your database can't keep up? MongoDB's native data model mirrors the language LLMs already speak. Ship at the speed of AI while staying ACID compliant at Fortune 500 scale. Start building at https://mongodb.com/ai.

–

Links:

• Manish Goregaokar’s blog: https://manishearth.github.io/  
• Manish on X: https://x.com/ManishEarth

–

Timestamps:

(00:00) Intro
(00:27) The future of the con is already here
(01:40) Scams are run like businesses
(03:30) A scam aimed at programmers
(05:12) Why single sign-on is the one ring to rule them all
(06:50) Persistent threats against individuals
(08:58) What LLMs change about cost
(11:40) Manual scams and business email compromise
(14:30) How money movement constrains scammers
(17:39) The industry that helps you pay ransoms
(19:32) Bespoke attacks on individual targets
(21:41) The Rust community gets targeted
(23:02) Supply chain attacks and long-lived impersonation
(24:58) Money mules and the three stages of laundering
(28:35) Why law enforcement struggles with internet crime
(30:01) Sponsors: Mercury │ Granola
(33:03) Patriotic hackers and the post-Soviet talent pool
(35:27) What individuals can do
(36:27) Only trust communication you initiate
(38:21) Wire fraud in real estate closings
(41:07) Why legitimate institutions look like scams
(45:22) Who eats the loss
(47:21) Sponsor: MongoDB
(48:12) Why banks can't just block new payees
(50:45) Verification protocols and family passwords
(53:07) Urgency and secrecy
(54:44) Systemic fixes and who pays
(57:55) Open weights and the coming wave
(01:02:17) Wrap

In this episode, Patrick McKenzie (patio11) is joined by Misha Gurevich, CEO of Aerolamp, and Vivian Belenky, its chief scientist and a researcher at Columbia, to discuss why UVC, a wavelength of light that inactivates airborne pathogens but is absorbed harmlessly by the dead outer layer of human skin, remains almost entirely undeployed. They explain the physics and safety case for 222 nanometers, the economics of installation, and preliminary South African trial results showing 90% suppression of tuberculosis transmission in hospital wards. 

–
Full transcript available here: https://www.complexsystemspodcast.com/the-economics-of-putting-germicidal-light-in-every-room/

–
Presenting Sponsors: Mercury & MongoDB

Complex Systems is presented by Mercury—radically better banking for founders. Mercury's new feature Command brings an LLM directly into your banking interface, so checking balances, finding invoices, or sending a wire is as easy as asking. Apply online in minutes at https://mercury.com/. 

What's the point of building faster with AI if your database can't keep up? MongoDB's native data model mirrors the language LLMs already speak. Ship at the speed of AI while staying ACID compliant at Fortune 500 scale. Start building at https://mongodb.com/ai.

–

Links:

• Aerolamp: https://aerolamp.net/ 

–

Timestamps:

(00:00) Intro
(00:58) What far UVC does to pathogens
(03:25) Why it's safe for skin and eyes
(09:24) What a typical deployment looks like
(11:21) Why isn't this everywhere already?
(13:55) Where the rollout should be prioritized
(17:31) The economics per room and per building
(20:56) Lamp lifetime and maintenance
(21:48) When does the evidence become undeniable?
(23:24) Tuberculosis and long-term care as early proof points
(27:16) Pandemic math and kneecapping the next outbreak
(30:16) Sponsors: Mercury | MongoDB
(32:49) R0, K-factor, and why the built environment beats opt-in
(34:23) Building codes, ASHRAE 241, and competing technologies
(39:11) The deployment curve and the awareness problem
(42:33) Scaling manufacturing: krypton chloride versus LEDs
(45:14) Price versus cost, and where the inflection point is
(47:35) Should you put one in your house?
(51:51) What would make this fail?
(56:09) The hygiene hypothesis objection
(01:01:13) The precautionary principle and reasoning under uncertainty
(01:06:04) Where the first million dollars should go
(01:10:49) The business case and the heckler's veto
(01:14:51) Where to find Aerolamp
(01:15:52) Wrap

Patrick McKenzie (patio11) is joined by Garrison Lovely, journalist and author of Obsolete: The AI Industry's Trillion-Dollar Race to Replace Us and How to Stop It, to map the three-sided debate over AI risk and why the arguments keep talking past each other. 

They then turn to what cheap cognition does to surveillance that already exists: FinCEN receives roughly 4 million suspicious activity reports a year and reads almost none of them, ICE agents run about a million queries against that database annually, and every podcast ever recorded is now transcribable for approximately nothing. The conversation covers capabilities denialism, ablated open-weights models, the fraud supply chain, and why AI is also unusually good at writing the Regulation E letter that gets your bank to fix your problem.

–
Full transcript available here: https://www.complexsystemspodcast.com/cheap-cognition-and-the-end-of-practical-obscurity-with-garrison-lovely/

–
Presenting Sponsors: Mercury, MongoDB & Chainguard

Complex Systems is presented by Mercury—radically better banking for founders. Mercury's new feature Command brings an LLM directly into your banking interface, so checking balances, finding invoices, or sending a wire is as easy as asking. Apply online in minutes at https://mercury.com/. 

What's the point of building faster with AI if your database can't keep up? MongoDB's native data model mirrors the language LLMs already speak. Ship at the speed of AI while staying ACID compliant at Fortune 500 scale. Start building at https://mongodb.com/ai.

If attackers are using AI to weaponize code faster than any team can review it, your scanners won't save you. Chainguard builds libraries and container images from source, verified all the way down, with near-zero CVEs and zero malware. Build safely at https://www.chainguard.dev/. 

–

Links:

• Obsolete: The AI Industry’s Trillion Dollar Race to Replace Us―and How to Stop It: https://www.amazon.com/Obsolete-Power-Profit-Machine-Superintelligence/dp/1682196305 

–

Timestamps:
(00:00) Preview
(00:43) Intro
(01:51) The three-sided debate over AI risk
(05:32) Power, politics, and the tech backlash
(09:49) Capabilities denialism and AI tells
(14:54) The obsoleting machine
(17:09) The Turing test is dead
(18:56) Languages for free, then software engineering
(22:37) Sponsors: Mercury | MongoDB
(25:09) How high up the stack do the models decide?
(29:11) Surveillance and cheap cognition
(32:38) Podcasts, FinCEN, and the end of practical obscurity
(38:35) Section 702 and the data broker loophole
(39:35) Sponsor: Chainguard
(40:55) Section 702 and the data broker loophole (cont’d)
(47:23) Institutional friction and a million ICE queries
(53:29) Security through obscurity no longer works
(54:54) Scams, fraud, and ablated models
(1:00:20) Personal utility versus societal backlash
(1:02:16) AI as a tool for redress
(1:06:34) State capacity and regulating what you understand
(1:12:37) Tobacco, nuclear, and AlphaFold: strangle it or steer it
(1:18:37) Where to find Garrison and the book
(1:20:32) Wrap

In this episode, Patrick McKenzie (patio11) is joined by Leila Clark, founder of Stardrift and formerly a software engineer at Jane Street, to discuss her essay "What Are You Getting Paid In?"  They cover how a Jane Street manager staffed the firm's least lucrative corner by paying people in culture, why Ken Griffin is worth roughly fifty Taylor Swifts, and how a longtime Google product manager quietly ends up with Grammy-winner money. The conversation ranges from academia's brutal tournament structure and YC as a script to founderdom to the one currency Taylor Swift holds that Ken Griffin's $50 billion buys only awkwardly: a restaurant reservation anywhere in New York.

–
Full transcript available here: https://www.complexsystemspodcast.com/what-youre-actually-getting-paid-in-with-leila-clark/

–
Presenting Sponsors: Mercury, Chainguard & MongoDB 

Complex Systems is presented by Mercury—radically better banking for founders. Mercury's new feature Command brings an LLM directly into your banking interface, so checking balances, finding invoices, or sending a wire is as easy as asking. Apply online in minutes at https://mercury.com/.

If attackers are using AI to weaponize code faster than any team can review it, your scanners won't save you. Chainguard builds libraries and container images from source, verified all the way down, with near-zero CVEs and zero malware. Build safely at https://www.chainguard.dev/.

What's the point of building faster with AI if your database can't keep up? MongoDB's native data model mirrors the language LLMs already speak. Ship at the speed of AI while staying ACID compliant at Fortune 500 scale. Start building at https://mongodb.com/ai.

–

Links:

• What are you getting paid in: https://www.approachwithalacrity.com/p/what-are-you-getting-paid-in 

–
Timestamps:
(00:00) Intro
(01:23) What are you getting paid in?
(03:16) Scripts, teacher figures, and hitting capitalism
(08:18) The academia trap
(11:08) Paying people in culture: Jane Street's back office
(14:03) Ken Griffin vs. Taylor Swift
(16:53) Learning about finance by osmosis (or not)
(20:44) Sponsors: Mercury | Chainguard
(23:46) Musician money vs. Google money
(30:03) Keeping up with the Joneses and the meritocratic ladder
(33:12) YC as a script to founderdom
(36:19) What entrepreneurship pays you in
(39:12) Gratitude, culture, and quirky preferences
(40:11) Sponsor: MongoDB
(43:48) Does winning this script look like winning to you?
(48:29) Don't end the week with nothing
(51:06) Fame and living in bubbles
(56:57) Restaurant reservations as a currency
(1:02:57) Where to find Leila
(01:03:35) Wrap

Patrick McKenzie is joined by returning guest Clara Collier, editor-in-chief of Asterisk Magazine, to discuss how working writers and reporters actually use LLMs. Patrick walks through the machinery behind his recent SPLC reporting, including "parallel construction" with LLMs: when sources can't go on the record, models can surface public confirmation of the same facts in unguarded podcast interviews, press releases, and prepared remarks. They also cover why the best smoking gun sat unnoticed on the coalition's own Twitter account, the David O. Selznick theory of the corporate memo, and where hardball politics in a democracy ends and crimes begin.

–
Full transcript available here: https://www.complexsystemspodcast.com/llms-and-writers-with-clara-collier-of-asterisk-magazine/

–
Presenting Sponsors: Mercury, MongoDB & Chainguard

Complex Systems is presented by Mercury—radically better banking for founders. Mercury's new feature Command brings an LLM directly into your banking interface, so checking balances, finding invoices, or sending a wire is as easy as asking. Apply online in minutes at https://mercury.com/. 

What's the point of building faster with AI if your database can't keep up? MongoDB's native data model mirrors the language LLMs already speak. Ship at the speed of AI while staying ACID compliant at Fortune 500 scale. Start building at https://mongodb.com/ai.

If attackers are using AI to weaponize code faster than any team can review it, your scanners won't save you. Chainguard builds libraries and container images from source, verified all the way down, with near-zero CVEs and zero malware. Build safely at https://www.chainguard.dev/. 

–

Links:

• Complex Systems on YouTube: https://www.youtube.com/@patio11podcast 
• Asterisk Magazine: https://asteriskmag.com/ 

–

Timestamps:

(00:00) Preview

(00:51) Intro

(01:16) How legacy media feels about LLMs

(05:02) Writing as thinking: why the byline matters

(08:51) LLMs as feedback partners and simulated readers

(12:39) The future of the corporate memo

(15:58) Sponsor: Mercury | MongoDB

(18:30) Memos, de-skilling, and thinking on paper

(20:02) David O. Selznick's memos and the MrBeast production doc

(23:28) Can models introspect on their own work?

(30:06) Sponsor: Chainguard

(31:27) The SPLC investigation

(37:07) Chains of evidence and protecting sources

(45:20) Parallel construction with LLMs

(49:41) Mining podcasts for unguarded admissions

(59:35) Deepen your strengths, don't just patch weaknesses

(1:03:34) The smoking gun the LLM missed

(1:04:48) Hardball politics versus crimes

(1:10:48) Why a payments newsletter reported a political story

(1:16:19) Reporting stories that will be weaponized

(1:20:24) The tech right and internal company politics

(1:23:44) Debanking discourse and drawing distinctions

(1:30:21) Hate speech, social sanctions, and owning decisions

(1:34:51) Wrap

In this episode of Complex Systems, Patrick McKenzie (patio11) is joined by Justin Kuiper, a longtime writer for MatPat's Game Theory family of channels and now creator of Proof Positive, to discuss the microeconomics of YouTube. They break down how creators actually get paid — from $3–$20 CPMs and the leaky funnel where a million views yields perhaps 50,000 actual ad views, to sponsor reads, Super Chats, and MrBeast selling chocolate bars as his own advertiser of last resort. Along the way, they explore how a successful channel becomes a firm that absorbs specialist labor from less successful peers, why the algorithm gives every upload a trial by attention, and what parasocial spending has in common with Mark Twain's speaking circuit.

–
Full transcript available here: https://www.complexsystemspodcast.com/justin-kuiper-youtube/

–
Presenting Sponsors: Mercury, Chainguard & MongoDB 

Complex Systems is presented by Mercury—radically better banking for founders. Mercury's new feature Command brings an LLM directly into your banking interface, so checking balances, finding invoices, or sending a wire is as easy as asking. Apply online in minutes at https://mercury.com/. 

If attackers are using AI to weaponize code faster than any team can review it, your scanners won't save you. Chainguard builds libraries and container images from source, verified all the way down, with near-zero CVEs and zero malware. Build safely at https://www.chainguard.dev/. 

What's the point of building faster with AI if your database can't keep up? MongoDB's native data model mirrors the language LLMs already speak. Ship at the speed of AI while staying ACID compliant at Fortune 500 scale. Start building at https://mongodb.com/ai.

–

Links:

• Proof Positive on YouTube: https://www.youtube.com/channel/UCns_C7cPAguFSv2YdltaOsg 

–

Timestamps:

(00:00) Preview

(00:45) How creators make money on YouTube

(04:06) CPMs and the international ad market

(06:36) From solo creator to firm: working for MatPat

(09:28) Sponsors: Mercury | Chainguard

(12:30) From solo creator to firm: working for MatPat (cont’d)

(13:49) YouTube as a farm league for other industries

(17:15) Why now is the best time to start: discovery and universal basic attention

(20:47) Power users and how recommendations work

(24:23) Brand safety, rabbit holes, and the money laundering short

(27:38) Fads, timing, and the day-two piece

(35:52) The production function: scripts, shot lists, and editing labor

(42:14) The aesthetics of authenticity

(46:58) Sponsor: MongoDB

(47:47) Why more people should make videos

(53:52) Content marketing, sponsor reads, and remnant inventory

(01:00:24) Chocolate bars, paint sets, and creator products

(01:05:00) Parasocial relationships and the market in status

(01:16:43) Where to find Justin: Proof Positive
(01:18:58) Wrap

Patrick McKenzie (patio11) reads his 2023 essay "Deposit Franchises as Natural Hedges," written seven weeks into that year's banking crisis, making the case that deposit franchises are a natural hedge against interest rate risk (one regional banks were quietly encouraged to sell off by loading up on agency MBS). He walks through why "sweat and smiles" deposits were assumed to be sticky enough to fund long-duration assets, why that assumption broke down for retail and sophisticated depositors alike once rates rose, and how the resulting losses moved through bank balance sheets. Patrick closes with two years of hindsight: what the essay got right and wrong about how bad it would get.

–
Full transcript available here: https://www.complexsystemspodcast.com/deposit-franchises/

–
Presenting Sponsors: Mercury, MongoDB & Chainguard

Complex Systems is presented by Mercury—radically better banking for founders. Mercury's new feature Command brings an LLM directly into your banking interface, so checking balances, finding invoices, or sending a wire is as easy as asking. Apply online in minutes at https://mercury.com/. 

What's the point of building faster with AI if your database can't keep up? MongoDB's native data model mirrors the language LLMs already speak. Ship at the speed of AI while staying ACID compliant at Fortune 500 scale. Start building at https://mongodb.com/ai.

If attackers are using AI to weaponize code faster than any team can review it, your scanners won't save you. Chainguard builds libraries and container images from source, verified all the way down, with near-zero CVEs and zero malware. Build safely at https://www.chainguard.dev/. 

–

Links:

• Deposit franchises as natural hedges: https://www.bitsaboutmoney.com/archive/deposit-franchises-as-natural-hedges/ 

–

Timestamps:

(00:00) Intro
(03:48) Natural hedges
(07:38) Deposit franchises as an asset
(10:53) The value of a deposit franchise increases with interest rates
(11:30) A brief aside about deposit beta
(14:58) Sponsors: Mercury | MongoDB
(17:30) A brief aside about deposit beta
(18:05) The deposit franchise as a hedge
(21:52) Regional banks were instructed to load up on agency MBS
(25:09) Why did the hedge bust?
(29:42) Sponsor: Chainguard
(31:03) Further bad news: the problem is bigger than MBS
(34:01) It is no longer February
(34:51) “Why didn’t the hedger hedge?!”
(35:37) So what do we do now?
(37:59) Postscript
(39:45) Wrap

Patrick McKenzie (patio11) reads his 2021 essay "Payments in Japan," tracing how Japanese consumers navigate a landscape with dozens of competing payment methods at once: credit cards, electronic money, QR-code super apps, convenience-store cash vouchers, and bank transfers. Along the way he covers the JFTC's campaign to force credit card networks to disclose interchange rates, how Rakuten and 7-Eleven each bought a bank to solve a payments problem blocking their core business, why PayPay's subsidized 2018 launch let it run away with the QR code market, and why konbini payments remain popular despite a user experience frozen in the late 1990s.

–
Full transcript available here: https://www.complexsystemspodcast.com/japanpayments/

–
Presenting Sponsors: Mercury & MongoDB

Complex Systems is presented by Mercury—radically better banking for founders. Mercury's new feature Command brings an LLM directly into your banking interface, so checking balances, finding invoices, or sending a wire is as easy as asking. Apply online in minutes at https://mercury.com/. 

What's the point of building faster with AI if your database can't keep up? MongoDB's native data model mirrors the language LLMs already speak. Ship at the speed of AI while staying ACID compliant at Fortune 500 scale. Start building at https://mongodb.com/ai.

–

Links:

• Payments in Japan: https://www.bitsaboutmoney.com/archive/payments-in-japan/ 
• An Introduction to Japanese Society: https://www.amazon.co.jp/Introduction-Japanese-Society-Yoshio-Sugimoto/dp/1107626676/ 
•  Use transit cards on your iPhone or Apple Watch in Japan: https://support.apple.com/en-us/120474 

–
Timestamps:
(00:00) Intro
(02:44) Credit cards
(10:40) Payment method heterogeneity
(12:57) Cash
(14:57) Sponsors: Mercury + MongoDB
(17:29) Cash (cont’d)
(19:58) Electronic money systems
(22:13) App-based payments
(28:27) Convenience store payments
(31:27) Bank transfers
(34:03) Ambitions thwarted
(34:30) Wrap

Patrick McKenzie reads from his 2022 Bits About Money essay on mortgages, making the case that a mortgage is best understood as a manufactured product, not a simple loan between a bank and a customer. He walks through the assembly line behind every home loan, the loan officer and back-office staff who build the 700-page document. Then he traces the supply chain it gets sold into, where GSEs insure against non-payment risk, servicers buy the right to collect monthly checks, and pension funds and other private capital end up holding the economic exposure, because they want it more than banks do.
–
Full transcript available here: https://www.complexsystemspodcast.com/mortgages/
–
Presenting Sponsors: Mercury & Granola 

Complex Systems is presented by Mercury—radically better banking for founders. Mercury offers the best wire experience anywhere: fast, reliable, and free for domestic U.S. wires, so you can stay focused on growing your business. Apply online in minutes at mercury.com.

If meetings consistently leave you with hazy action items and lost context, Granola handles the transcription so you can actually participate and gives you searchable notes afterward. Try it free at granola.ai/complexsystems with code COMPLEXSYSTEMS

–

Links:

• Mortgages are a manufactured product: https://www.bitsaboutmoney.com/archive/mortgages-are-a-manufactured-product/ 
• The 30-Year Mortgage is an Intrinsically Toxic Product: https://byrnehobart.medium.com/the-30-year-mortgage-is-an-intrinsically-toxic-product-200c901746a 
• Michael Lewis’ The Big Short: https://www.amazon.com/Big-Short-Inside-Doomsday-Machine/dp/0393338827 

–

Timestamps:

(00:00) Intro

(02:26) Mortgages are a manufactured product

(04:19) Who manufactures mortgages?

(07:08) Who buys mortgages?

(07:42) The risk of non-payment

(10:08) Sponsor: Mercury | Granola

(14:35) The risk of failing to service a mortgage correctly

(17:51) Every other risk you could imagine, of which there are many

(24:04) Scratching the tip of the iceberg

(25:10) More about flow meters

(26:24) Wrap