Mar 20, 2014 · 21m · mad
Sean Gourley, Quid // Data Driven #25 // March 2014 (Hosted by FirstMark Capital)
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
At a FirstMark Capital event, Sean Gourley explores the hardware limitations of human cognition and how high-speed computer algorithms are increasingly overtaking financial markets, digital media, and web infrastructure. He highlights systemic risks, automated narrative manipulation, and proposes frameworks for augmented human-algorithm interaction.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Matt holds 4.6% of the talking time here. How this is scored →
Matt as informed peer 0.3 Guest teaching 3.0 Guest disagreement 1.2 Matt pushing back 0.2
speaking balance: gold is Matt, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 20:55 Polite premise rejection on prediction
Sean gently rejects the audience member's premise that tools are used to predict government actions, clarifying that powerful actors use them to actively shape outcomes instead.
Hardest push from Matt ▶ 16:50 Host steering guest to explain companyMatt Turck steps in after the presentation to explicitly guide Sean to talk about Quid and what he is building, ensuring the audience learns about his actual product.
Biggest teaching moment ▶ 9:50 Economic analysis of internet data filteringSean educates the audience by breaking down the contrast between Facebook charging 1.20 dollars per user in compute versus Google's 12 to 13 dollars, highlighting how cheap data filtering influences platform incentives.
Matt holds his own ▶ 16:50 Host initiating Q and A moderationMatt Turck demonstrates host authority by transitionally moderating the stage, thanking Sean for the talk, and prompting him to cover Quid's core technology.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Matt as informed peer | Guest teaching | Guest disagreement | Matt pushing back | Why |
|---|---|---|---|---|---|---|
| Measuring Cognitive Speed and Limits of Strategic Thinking | 0 | 3 | 1 | 0 | Sean delivers a solo keynote presentation detailing fMRI studies on chess players to illustrate cognitive speed limits. Because this is a monologue, host scores are strictly zero. | |
| Market Volatility, the Flash Crash, and Time Scale Ecosystems | 0 | 3 | 1 | 0 | Sean explains market volatility and high-frequency trading algorithms, citing the Knight Capital software failure. Host scores remain zero as this is part of the presentation monologue. | |
| Machine-Readable News, Fed Reports, and Fake Tweet Crashing Markets | 0 | 3 | 1 | 0 | Sean discusses how algorithms process machine-readable news and leave the market during Fed announcements or fake tweets. Host scores are zero due to the monologue format. | |
| Non-Human Web Traffic and Real-Time Ad Exchanges | 0 | 4 | 1 | 0 | Sean details ad exchange auctions and analyzes the infrastructure cost per user for Facebook versus Google. As a pure presentation segment, host scores are set to zero. | |
| Social Platform Design Grid: History vs. Identity | 0 | 3 | 1 | 0 | Sean breaks down social platform design along axes of history and identity, contrasting Facebook, Reddit, 4chan, and Snapchat. Host scores remain zero for this monologue section. | |
| Propaganda Bots and Automated Narrative Manipulation | 2 | 2 | 2 | 1 | Matt Turck opens the Q and A by inviting Sean to explain Quid and asking standard follow-up questions about its data ingestion. Sean gently reframes an audience member's premise from predicting to shaping world events. |
Statements from this episode (16)
Human reaction speed tops out at 650 milliseconds for expert chess players
“For an amateur chess player it's about 900 milliseconds for the brain to register whether or not the king is in danger. For an expert chess player, it does increase. It does, you know, get faster, but it tops out at 650 milliseconds.65 of a second.”
High-frequency algorithms execute over 60% of US equity trades
“U.S. Equities market of all completed trades is now somewhere north of 60%, done by high frequency algorithms.”
Firms spent $300M on fiber cables cutting NY-London latency by 5ms
“Ships now laying new fiber optic cables between New York and London to show five milliseconds of round trip time at the cost of three hundred million dollars.”
Gourley: Algorithms can briefly wipe $1 trillion from financial markets
“Sometimes the algorithms get together and decide collectively to wipe about a trillion dollars of market capitalization off the global financial markets only to return it about 15 minutes later with nothing really happening in the world around that.”
Gourley: Fast financial algorithms misbehave because speed requires simplicity
“So the algorithms that we've put in place of humans, although quicker, sometimes misbehave because they're not that, that smart, and they're not that smart because they've got to be quick.”
Gourley: Knight Capital lost $440M in 45 minutes from a rogue algorithm
“The case in point of Knight Capital a couple of years ago they released an algorithm onto the market. It was never supposed to be on the market. It was an algorithm that was meant for testing. It somehow makes its way off the staging machine and into the real …”
Gourley: Trading algorithms exit markets during Fed report releases
“They're all pre-programmed. To get out of the market when a Fed report comes out about the unemployment rates because they don't know how to read it.”
A fake AP tweet about White House explosions wiped $200B
“And a fake news report from AP Newswire comes out, says there's a bomb in the White House, and then this happens. The algorithms leave the market because they go, I don't really know what's going on here. I know something to do with the bomb happened, and prob…”
Gourley: Algorithms that read the news will inevitably start writing it
“Now, of course, if algorithms can start to read the news, they'll, surely, they'll start writing the news.”
Sean Gourley: 61% of internet traffic is non-human
“61% of the internet traffic is now non-human. It looks something like this. And it's increasing. Last year it was at 51%.”
In 2014, Google generated $30 per user annually versus Facebook's $6
“Google makes about 30 bucks a year per user. Facebook is doing a little bit worse. They're at six dollars per user”
Facebook's infrastructure cost $1.20 per user annually versus $13 for Google
“It costs about buck 20 to run a user on Facebook for a year. That'll vary a little bit depending on which country you're in in the world. It costs 12 bucks, 13 dollars, for a user on Google.”
Gourley: Google includes Wikipedia on page one for 85% of common searches
“Google actually returns Wikipedia 85% of the time with the most common searches on the first page.”
Gourley predicted consumers would eventually pay subscriptions for algorithmic control
“I think we're going to see subscriptions. I think we're going to see people coming out and saying, you know what, I want a little more control of this algorithm, the information that it gives me, the friends that it chooses.”
DARPA and intelligence agencies fund automated narrative manipulation research
“The KGB's just put money into this. IAPA and DARPA have got money going into this stuff.”
Quid works with PR and government clients to shape narratives
“So we work a lot of PR and marketing Drivequid to kind of shape the stories, and they try and not so much predict what the message is gonna be, but be actually the ones that are making it. And that of course has the government clients doing that stuff as well.”