← Back to list
Lex Fridman PodcastPodcast24 Apr 2020Source: lexfridman.comHost: Lex Fridman

#91 – Jack Dorsey: Square, Cryptocurrency, and Artificial Intelligence

In plain words

Jack Dorsey (Twitter and Square CEO) talks about Square's success, Bitcoin, and AI risks. He says Square's trick isn't tech but trust: instead of checking credit scores, they trust merchants first and monitor each transaction, boosting approval rates from 30% to 99%. He loves Bitcoin as the internet's native currency because no one controls it. He worries AI outsources self-awareness, like Google knowing you better than yourself. Key mentions: Square (approval rate surge), Bitcoin (called the internet's native currency), Google (flagged for data centralization).

AI SummaryAI-generated · may contain errors · verify against the original

Jack Dorsey discussed topics including Square, cryptocurrency, and artificial intelligence on the Lex Fridman Podcast. Key takeaways include: Square is committed to empowering small businesses through technology and promoting financial inclusion; cryptocurrencies (such as Bitcoin) are seen as intern

~10 min full read · 8 sections
Deep Analysis

This Issue at a Glance

Jack Dorsey (co-founder and CEO of Twitter and Square) focused on Square’s mission, cryptocurrency convictions, and AI risks in the Lex Fridman podcast. Core judgment: Square attributes its 99% merchant application approval rate to a mindset shift from "distrust" to "trust + verification," rather than purely technological advantages—a rare contrarian framework in the fintech space.


I. Square's Core: Rebuilding Financial Access Through "Trust"

Jack Dorsey believes the biggest problem in the financial industry is the default mindset of "distrust."

When Square was founded, it discovered that a large number of small merchants were shut out by banks due to poor credit records or a lack of FICO scores—"checking credit just to accept payments is inherently absurd." Traditional financial institutions use credit checks as the sole tool to "understand authenticity, intent, and future behavior," but Square chose a different direction: shifting from "prove you are trustworthy" to "trust first, then verify."

The specific approach is:

  • Downgrading the monitoring granularity from the merchant level to the transaction level—analyzing each transaction in real time rather than pre-assessing the merchant's overall risk
  • Once a merchant enters the system, as long as their behavior is "high-integrity, trustworthy, and not anomalous," service continues; manual review is triggered only when suspicious patterns emerge
  • Result: The industry average approval rate is only 30–40%, while Square has raised it to 99%

This shift relies on the real-time intervention of data science (later evolving into machine learning/deep learning). Dorsey emphasized: "We built a data science team in our first year because the essence of the business is not making hardware, but getting more people into the system."

Cooperation with banks and card networks follows a gradual trust strategy: starting with a pilot of 500 people, then progressively demonstrating results—1,000 people → 10,000 people → 50,000 people—convincing regulators with data rather than promises.


2. Bitcoin: The Internet's Native Currency and the Beauty of "No Owner"

Dorsey believes Bitcoin is "the native currency the internet deserves," with its greatest value lying in the absence of a single controller.

Core arguments:

  • Decentralized design: The Bitcoin whitepaper is "one of the most important computer science works of the past 20-30 years, a poem." Published under a pseudonym (rather than anonymously), it preserves a humanized identity while sacrificing ego, making it "an extremely powerful statement."
  • Eliminating cross-border barriers: Square's biggest bottleneck is currently its inability to "operate like an internet company" — entering each new market requires partnering with local banks and navigating different regulatory environments. As a unified currency, Bitcoin allows for "launching products in all global markets at once."
  • Unpredictability: "No one knows what will happen in 10-20 years, and that's the beauty — everyone entering the ecosystem can change its direction in small or massive ways."

Dorsey specifically notes that currency is a foundational layer, and its greater impact lies in "changing how society operates." He illustrates this with a startup case in Ethiopia: the biggest problem for the local ride-hailing platform is not user experience, but that "passengers cannot conveniently pay the company, and the company cannot conveniently pay drivers" — the lack of payment infrastructure directly stifles innovation.

Interpretation of the pseudonym (Satoshi Nakamoto): Dorsey argues that a pseudonym is smarter than anonymity — "Anonymity is dropping something and walking away; a pseudonym builds a temporary but recognizable identity, allowing the community to construct narratives around it (e.g., the concept of Satoshi's), while maintaining genuine self-sacrifice."


3. AI Risks: Explainability, Outsourced Self-Awareness, and Centralization

Dorsey’s concerns about AI center on three levels, rather than existential threats.

1. Black-Box Decision-Making and Explainability

"We are building countless black boxes that cannot explain their decisions, yet we increasingly trust them—from loans to content recommendations, from driving to healthcare." He argues that explainability is a key gap in general intelligence—"being able to explain 'why' is true intelligence."

2. Outsourcing Self-Awareness

Citing Yuval Noah Harari’s view: Google knows your preferences better than you do. For a generation that has only known the internet, self-awareness is being "offloaded" to algorithmic systems—from "Should I stand?" to "Which doctor should I choose?" to "Who should I date?" Dorsey believes that the key is not to reject algorithms, but to remain "aware"—knowing that one is handing over decision-making to a system that cannot explain itself.

3. Centralization of Data and Algorithms

"The scariest thing about AI is who owns the data and who can manipulate it." He argues that data is currently concentrated in the hands of a few large companies, and society needs better answers—regarding data ownership, sharing mechanisms, and how to make algorithms serve broader needs.

Assessment of the "Detection vs. Generation" race: Dorsey believes that "the speed of creating technology far outpaces the speed of detecting it"—especially in finance, where deepfakes could be used to fabricate fake identities (driver’s licenses, passports) to bypass risk control systems. The solution is not to pursue perfect detection, but to "accelerate the system’s evolution speed, quickly absorbing correct/incorrect signals."


4. UBI and Job Displacement: Andrew Yang's Framework Is Correct

Dorsey explicitly endorses Andrew Yang's assessment of automation's impact and argues that UBI is "long overdue."

Specific observations:

  • Trends already visible in Square's business: Cashier positions are being replaced by mobile ordering, self-service kiosks, Amazon Go, etc.—"There are a huge number of retail cashiers, yet no real answer for how they can transition."
  • UBI is not an incentive to quit, but a survival floor: "$1,000 a month is not enough to make you quit your job, but it allows you to stop worrying about daily survival and focus on 'what to do next and what skills are needed.'"
  • Difference from the Industrial Revolution: It is not automation itself, but "speed and concentration"—data and algorithms are concentrated in the hands of a very small number of companies, making the transition window shorter and power more centralized.

Dorsey adds: The concept of UBI dates back to Martin Luther King Jr. and even earlier. The specific amount ($1,000) can be adjusted, but "we should start implementing it and observe the results."


5. Personal Philosophy: Death as a Tool, Discomfort as a Learning Path

Dorsey treats death as a "tool used multiple times daily" rather than an object of fear.

  • "Death is a transformation and a tool—it makes me feel the importance of every moment. I only have this many sunrises and sunsets to see."
  • He contemplates death multiple times a day for priority judgment: "I only have one hour—is this really how I want to spend it?"

On diet (one meal a day): Dorsey learned about it from a Wim Hof podcast and began trying it, with the core motivation being "to make himself uncomfortable"—"I learn the most when I am most uncomfortable, because that's when everything surfaces." He draws a parallel to his childhood experience of overcoming a stutter: joining the speech club was "the most uncomfortable thing imaginable," but it gave him a profound understanding of the power of communication.

On the meaning of life: It boils down to a sense of connection—being aware of one's own existence, connecting with others, seeing the things one helped build being used, and "being part of something larger than oneself."


Mentioned Positions

Position Guest Stance Key Data
Square Bullish (core business) Merchant application approval rate improved from 30-40% to 99%; founder committed 28% of his wealth (approximately $1 billion in equity) to COVID-19 relief
Bitcoin Strongly bullish Described as "internet-native currency" and "one of the most important computer science works of the past 20-30 years"
Twitter Not discussed in depth This episode focuses on Square; Twitter engineering topics to be covered later
Amazon Go Risk warning (job displacement) Cited as an example of cashier positions being replaced
Google Risk warning (data centralization) "Google knows your preferences better than you do yourself"

Judgments Worth Remembering

1. The "Trust + Verify" Framework (Dorsey): The financial industry defaults to distrust; Square reversed this to "trust first, then verify"—lowering monitoring granularity from the merchant level to the transaction level, raising the approval rate from 30-40% to 99%. This is a rare counter-consensus move in the fintech space.

2. Bitcoin as "Internet-Native Currency" (Dorsey): Its greatest value is not the technology itself, but "having no single controller"—"No one knows what will happen in 10-20 years; everyone entering the ecosystem can change its direction." Square's current biggest bottleneck is "the inability to operate like an internet company," and Bitcoin can eliminate cross-border barriers.

3. Pseudonymity is Smarter than Anonymity (Dorsey): Satoshi Nakamoto's pseudonym "built a transient yet identifiable identity, allowing the community to build a narrative around it while maintaining true self-sacrifice"—a unique interpretation of identity strategies in crypto culture.

4. AI's Biggest Risk is Outsourcing Self-Awareness (Dorsey, citing Harari): For a generation that has only known the internet, Google knows their preferences better than they do themselves—"The key is not to reject algorithms, but to stay aware, knowing that you are handing over decision-making to a system that cannot explain itself."

5. The Speed of Creating Technology Far Exceeds the Speed of Detecting It (Dorsey): In financial risk control, deepfakes could be used to fabricate false identities—the solution is not perfect detection, but "accelerating the system's evolution speed to quickly absorb correct/incorrect signals."

6. UBI is "Long Overdue" (Dorsey): Andrew Yang's framework is "100% correct"—not to incentivize quitting jobs, but to provide a survival baseline, giving those affected time to transition. Unlike the Industrial Revolution, the current challenge lies in "speed and centralization."

7. Death as a "Tool Used Multiple Times a Day" (Dorsey): Not fearing death, but using it to "feel the importance of every moment"—"I only have one hour; is this really how I want to spend it?" This is the ultimate framework for priority judgment.

8. Discomfort as a Learning Path (Dorsey): From eating one meal a day to overcoming a stutter, the core motivation is "making oneself uncomfortable"—"I learn the most when I am most uncomfortable, because that's when everything surfaces."