The Network State Podcast

Episode #15 — Jeremy Howard (July 16, 2025)

Overview

In this episode, host Balaji Srinivasan (A) sits down with Jeremy Howard (B), co-founder of Fast AI and Answer AI, to discuss democratizing AI education, Jeremy’s unconventional career from McKinsey to Silicon Valley, global talent arbitrage, the perils and promise of positive feedback loops in wealth and power, and the future of education and nationhood in the age of internet-native communities. The conversation weaves personal anecdotes, philosophy, education reform, political theory, and startup insights with a forward-looking lens on AI and global innovation.

Jeremy Howard’s Mission: Making AI Accessible

• Jeremy underscores Fast AI’s founding mission: democratize access to AI knowledge and tools to avoid centralization in the hands of a few rich labs or corporations (00:34).

  • “We basically decided to get AI into the hands of as many people as possible, including people with few resources... At that time, only five labs in the world...” — Jeremy Howard (00:59)

• Fast AI combined open courses, research, software, and advocacy to break the perceived compute and knowledge monopoly of giants like Google and OpenAI.

Fighting Centralization in AI

• Jeremy describes the prevailing belief in tech giants around 2015-2017: that only Google-scale resources enabled true AI breakthroughs.
  • “I remember Jeff Dean saying, there’s no point trying to do stuff with AI unless you’re at Google because only we have scale compute.” — Jeremy Howard (01:55)
  • Fast AI’s success in the DawnBench competition proved lean teams could outpace resource-rich labs through innovation and efficiency (02:26).

Memorable Moment (DawnBench Competition)

• Fast AI beat Intel and Google by shrinking images for 90% of the training, only upscaling at the end—a clever workaround big labs didn’t need or think to use (04:09).
• “We just wanted to say like, no, you don’t have to be a rich Google person to be successful.” — Jeremy Howard (02:45)

Leveraging Global “Dark Talent” with MOOCs

The “Dark Matter” of Global Talent (07:56)

• Balaji analogizes phones as “mobile telescopes” that can now find the world’s hidden ‘dark talent,’ similar to how the Hubble telescope detects dark matter.

• Jeremy shares powerful stories:

  • Reaching mass audiences in India, Nigeria, Africa with Fast AI’s MOOC.
  • “She said...‘I want to know if you think it’s ok for me to do AI.’...A couple of years later, she wrote to me from Google in Silicon Valley.” — Jeremy Howard (09:21, 10:37)

• Both agree on the transformative potential of accessible online education for those otherwise picking “coffee beans or whatever” (09:07).

Notable Quote

• “Teach a man to fish... or teach a man to recognize an image of a fish...” — Balaji Srinivasan (10:48)

• Jeremy’s gentle deflation of academic gatekeeping and celebration of talent from unexpected places is a constant theme.

• “There’s all this talent around the world; it drives me crazy that it’s not being used.” — Jeremy Howard (09:07)

Jeremy Howard’s Life Story & Career Trajectory

Unconventional Origins

• After philosophy studies cut short by a McKinsey gig at 19, Jeremy explains how consulting’s “human optimization” taught him skills beyond spreadsheets and programming (14:00).
  • Detailed breakdown of organizational mapping, influence tactics, and “getting organizations to do things” (17:47, 18:34).

Memorable Reflection (McKinsey)

• “We brought professional actors in to play the role of different types of clients... It was way more intense human optimization than I’d ever conceived of.” — Jeremy Howard (18:34)

Early Startups Pre-Silicon Valley

• FastMail (email, global reach) and Optimal Decisions (pricing optimization for insurance) as first entrepreneurial successes in Australia.
• Journey to Silicon Valley to co-launch Kaggle (data science competitions), skeptical of US VC appetite—proven wrong.

On Being Nonconformist

• Jeremy expresses regret about undervaluing his tech intuition and staying in consulting for 10 years.
  • “I was interested in stuff nobody else was interested in... I felt really upset with myself for being stupid, that they—everybody else can see it...” (14:00 - 17:13)

On Education: From Prussian Classroom to Internet Apprenticeship

The Problem with Industrial-Era Schooling (33:04)

• The standard K–12 system is rooted in Prussian ideals of conformity, designed to create “the same software in the heads” of all citizens.
• Balaji argues history is running in reverse: our future resembles pre-industrial apprenticeships more than post-war mass education.

A New Apprenticeship Model

• Discussion of online work, coding, and global meritocracy as a modern equivalent for early, practical skill development.

Notable Quotes

• “I know a lot of kids...ready to go to university when they’re like 11 or 12. And adults all try to stop them.” — Jeremy Howard (38:22)
• “Not every kid learns everything at the same speed...I want to help them all have the opportunity to have that excitement of feeling like they’re achieving their potential...” — Jeremy Howard (38:51)

Practical Parenting Example

• Jeremy describes letting his 9-year-old daughter self-direct learning, pushing back against excessive emphasis on “good behavior” and conformity (40:59–41:25).
  • “I don’t want her to be somebody in society who just fits in and does what she’s told.” — Jeremy Howard (41:38)

Positive Feedback Loops: Power, Wealth, and Societal Stability

The Peril of Positive Feedback Loops

• Jeremy’s key concern is how unchecked positive feedback loops—where wealth and power beget more of themselves—threaten democracy and stability (53:26).
  • “Power gets begets power. It doesn’t matter if I’m five generations away from Genghis Khan. What matters is I’m the king.” — Jeremy Howard (59:48)
  • “That’s the natural state of things. The natural state of things is for there to be one incredibly wealthy and powerful person that is there because of the power of positive feedback loops.” — Jeremy Howard (55:41)

Counterpoints: Decay, Arbitrage, & Social Mobility

• Balaji argues inherited power exhausts itself (“shirt sleeves to shirt sleeves”), and new talent will always arise through “arbitrage” (56:45, 62:42).
• The debate leads to AI, surveillance, and the risk of new forms of “global dictatorship” enabled by technology.

The Global Stakes: Network States, China, and the End of American Empire

The Internet vs. China (48:01)

• Balaji advances a provocative theory: the 21st-century geopolitical contest is “China vs. The Internet.”
• The Internet’s unique feature: true peer-to-peer property rights, rule of law, and a global digital polity—via blockchains (50:21).

On American Decline and “Network School” Vision

• US has become self-isolating; other countries (Malaysia, Singapore, Australia) are now competing for global tech talent through startup and nomad visas (27:57).
• Global startup communities enabled by visa innovation and digital infrastructure—the seed of “network states.”

Memorable Exchange

• “Hopefully [the future] won’t be versus...Unfortunately Xi Jinping has moved into a power vacuum...Prior to that, actually, China was much more of a democracy than people realized.” — Jeremy Howard (51:26)
• “The world is recentralizing around China...and the Internet has the people, the values, and the language.” — Balaji Srinivasan (48:01, 50:20)

New Models for Online Learning and Startups

“Learn-a-thon” and Cohort Binging (24:36)

• Fast AI and Balaji’s own MOOCs encounter common pitfalls: high dropout rates, aspirational start/finish gaps. Solutions discussed:
  • Intensive “binge” weekends ("learn-a-thon") to maximize course completion rates.
  • Peer encouragement, global fellowships.

Notable Quote

• “It drives me crazy because so many people tell me...‘I started your course, I meant to finish...You could just put aside one weekend and binge it, get it done.’” — Jeremy Howard (24:58)

Solve It: A New Way to Code and Learn AI (70:06)

• Jeremy previews Fast AI’s new course and platform, Solve It: teaching problem-solving with stepwise, iterative coding integrating AI, inspired by Polya’s heuristics and the Lean Startup approach.
  • “You’re constantly in control of the AI; you never get into that situation where the AI is kind of controlling you.” — Jeremy Howard (71:42)
  • “We’re hoping...to create like a thousand new startups from that course.” — Jeremy Howard (72:26)

Meta-Rationality, Empathy, and the Heart of Social Behavior

• The dialogue closes on how empathy is foundational, not sentimental, for successful societies, markets, and individuals (42:02-44:13).

  • “Altruism is not weakness. Altruism is strength. These are the people that survived.” — Jeremy Howard (44:13)

• The risk of destabilizing “stable equilibria” is highlighted, with warnings about technology’s power to accelerate societal upheaval (44:13–46:05).

Key Timestamps & Segments

• 00:34 – Jeremy’s founding story & vision for Fast AI
• 02:26 – Fast AI’s surprise win against Google & Intel (DawnBench)
• 06:44–11:17 – MOOCs, “dark talent,” global stories from Nigeria, Bangladesh, and beyond
• 14:00–19:51 – Jeremy’s time at McKinsey, organisational dynamics, sales, human “graph traversal”
• 33:04–38:22 – Systemic critique of K–12 education; history, conformism, and the Prussian model
• 38:22–42:02 – How to truly personalize education for gifted, passionate kids
• 53:26–62:47 – Wealth, power, and the trap of positive feedback; debate on inheritance and decay
• 48:01–50:20 – Macro theory: “China vs. The Internet” and future of global governance
• 70:06–72:45 – Solve It: Fast AI’s new approach to learning coding and AI collaboratively with AI
• 42:02–44:13 – Evolutionary arguments for altruism and the danger of breaking stable societal equilibria

Tone Highlights

• The conversation is pragmatic yet idealistic; passionate, occasionally skeptical, but always focused on constructive progress.
• Jeremy’s warmth, humility, and sense of mission are matched by Balaji’s pattern-spotting, playful analogies, and macro-perspective.
• Frequent mutual respect and genuine moments of mutual discovery (e.g., shared excitement over global talent, networked learning, new experiments in governance).

Summary

This episode is a master class in the intersection of technology, education, power, and societal structure—grounded by Jeremy Howard’s remarkable life story and Balaji’s visionary theorizing. If you care about the future of AI, education, or global citizenship, it’s an insightful and inspiring listen.