The Prof G Pod with Scott Galloway: First Time Founders with Ed Elson – This Physicist Is Building AI Droids

Date: November 2, 2025
Host: Ed Elson
Guest: Matan Grinberg (Co-founder and CEO, Factory)

Episode Overview

This episode of "First Time Founders" features Matan Grinberg, a former physicist turned entrepreneur, who co-founded Factory—an AI startup building fully autonomous coding "droids" (agents) now backed by Sequoia, JP Morgan, and Nvidia. Ed and Matan, old college friends, delve into the challenges, ethics, and economics of automating software development, the realities of Silicon Valley’s AI boom, and what autonomy, agency, and AI mean for the future of work.

Key Discussion Points and Insights

1. From Physics to AI Entrepreneurship

[02:57–09:28]

• Matan’s Academic Origin:
  • Motivated by "spite" after a discouraging math teacher in 8th grade told him he should retake geometry (“Spite is a very big motivator for me.” [03:37])
  • Self-taught higher mathematics, aimed to become a string theorist.
  • Studied at Princeton, then Berkeley for a physics PhD.
• The Pivot:
  • Realized mid-PhD he didn’t want to dedicate life to theoretical physics.
  • Discovered "program synthesis" (now known as code generation) via Berkeley AI classes.
  • Decided the applied route—starting a company—was more appealing than academia.
• Serendipitous Start:
  • Cold-emailed a former physicist-turned-investor he admired; received a same-day invite for a 3-hour walk in Menlo Park.
  • At a hackathon, met future co-founder Eno; 48 hours later, built Factory’s first demo.
  • After showing the demo, was urged to drop out of his PhD and pitch to Sequoia’s partners.

"That day... I showed [the demo] to him. He was like, 'This is all right.'... He's like, 'OK, drop out of your PhD and send me a screenshot.' And I was just like, fuck it." — Matan Grinberg (09:25)

2. What Sets Factory Apart in the AI Coding Space

[09:29–12:57]

• Factory’s Mission: Bring true autonomy to software engineering—"not to speed up code writing, but to take over the routine, high-friction parts of the engineering workflow."
• Industry Insight:
  • Most startups automate the enjoyable part (coding itself); Factory addresses the real bottlenecks: documentation, design reviews, approvals, code reviews, testing, system understanding.
• Empathy for Scale:
  • Many coding AI startups misunderstand actual developer needs in large enterprises; at massive orgs, “coding” is often less than 20% of an engineer’s day.

"As a company gets larger... the amount of time any given engineer spends on coding goes down. Your bottleneck is all these other things. With us focusing on that full, end-to-end spectrum, we end up hitting more closely to what developers actually want." — Matan Grinberg (10:49)

3. Fully Autonomous AI Droids Versus AI Partners

[15:42–18:25]

• Difference from Copilot:
  • Factory’s “droids” are truly autonomous; they don’t co-pilot with a dev, but can be handed a full task and execute it independently, including understanding, planning, implementing, testing, documenting, and reviewing.
• The Rise of Delegation:
  • The core skill for future developers becomes effective “delegation”—defining clear objectives and success criteria for the agent rather than crafting code line-by-line.

"We're changing the primitive of software development from writing code to the new primitive being a delegation: delegating a task to an agent." — Matan Grinberg (16:18)

4. AI and the Automation Debate: Are Developer Jobs at Risk?

[18:25–21:24]

• Not Replacing Engineers, Changing the Bar:
  • AI won’t outright replace engineers but will massively increase leverage.
  • Problems that weren’t worth engineering solutions before (too small a market) now become viable.
• Competition Keeps Employment Up:
  • If every competing company gains AI, employee headcounts will stay or even increase to deliver more value/innovation.
• We Underestimate New Possibilities:
  • Humanity’s never seen what happens when millions of engineers can each do 10-100x more.

"AI will not replace human engineers. Human engineers who know how to use AI will replace human engineers who don't." — Matan Grinberg (19:09)

5. AI’s Concentration of Power and Big Tech’s Domination

[21:24–27:24]

• Concerns Over Acquisitions:
  • Startups are frequently absorbed by tech giants (Google, Meta), threatening competition and innovation.
• The Founder’s Resolve:
  • Grinberg stresses that, at Factory, the mission and team orientation are intentionally anti-acquisition: “We do not want to be acquired, we do not want to be part of Big Tech.”
• Circular Financing:
  • Silicon Valley jokes about the “circular investing” phenomenon—companies investing in each other, inflating valuations and revenues, yet everybody still needs Nvidia’s chips at the end of the day.

"We are super focused on people that are very mission-driven. If you want to make a ridiculous amount of money, you can go to Meta... [but] the people who have joined our team have chosen this mission." — Matan Grinberg (25:03)

6. AI Regulation: Local Futility, Global Necessity

[31:04–33:23]

• California’s Regulations Are Largely Symbolic:
  • Local/state rules have little impact on global issues like AI, climate, or nuclear risk.
  • Real regulation must be at the US/China/EU level.
• The Valley’s Culture and AI Existential Dread:
  • Some AI researchers live as if the future is near-apocalyptic (not saving money, "spending it all"), reflecting a mix of utopian, doomer, or “post-economic” worldviews.
  • Grinberg sees these perspectives as sometimes irrational and advises perspective.

"I know so many people who work at the foundation model labs who don't have savings... because of this like vision of, like, something's coming." — Matan Grinberg (33:23)

7. Is the AI Boom a Bubble?

[35:36–41:46]

• Bubble Concerns:
  • AI’s capital influx feels frothy—benchmark: in early days, AI engineers got offers and funding by default.
  • However, Grinberg sees a major difference: real, enormous demand for GPUs and measurable productivity gains in coding.
  • Some compute and usage is subsidized (like Uber’s early days) but believes productivity gains are tangible.
• Winners in a Correction:
  • Nvidia (“Jensen always wins”) at the value chain’s base, providing indispensable hardware.
  • Startups with discipline, real adoption, and those who create actual value beyond hype will survive; many will die out if funding tightens.

"Jensen's going to stay winning... all of these circular deals, they all come back to Nvidia... someone's paying Jensen at the end of the day." — Matan Grinberg (39:16)

8. Life as an AI Founder in San Francisco

[43:01–44:13]

• Day-to-Day Reality:
  • Feels “in the trenches” building Factory; the privilege is working with an ultra-talented team and seeing real impact at enterprise scale.
  • Staying grounded: “Some of the largest enterprises that we're working with... don't have to deal with these problems because of something that we built.”

9. American Misconceptions About AI & AI Founders

[44:13–46:44]

• Public Knowledge Gap:
  • Most Americans only know ChatGPT; differences between models, agents, and use cases obscure to most.
• The Need to Rethink Workflows:
  • Urges every professional to “take a sledgehammer” to their workflow, rethink it from scratch with AI.
• San Francisco Bubble:
  • Warns founders against staying stuck in the SF echo chamber—advises “touching grass” and seeing real-world problems and needs.

10. Power, Agency, and the Human-AI Divide

[52:25–62:51]

• The Cult of Jensen:
  • In tech circles, Jensen Huang (Nvidia) is the most widely respected figure (“He’s really good. He doesn’t have enemies. He’s extremely generous with his time... “ [53:07])
• AI’s High Potential Meets Human Nature:
  • “It is just pure human nature... what do we do when we have really good technology? Let’s make porn.” [54:22]
  • But there's also real work happening—e.g., AI for biotech/medicine.
• Long-Term Human Value:
  • As AI supersedes human intelligence, value shifts from IQ to agency—the will and drive to do difficult or meaningful things.
  • Those with agency (decisiveness, vision, ability to direct agents) will thrive; mere intelligence becomes less distinguishing.

"Now we're having a little bit of a decoupling of human value being attributed to intelligence... the new primitive for humans is agency." — Matan Grinberg (54:43)

• Practical Parallel:
  • The real winners control and orchestrate the AI agents, rather than becoming subordinate to them.
• On Factory’s Edge:
  • Factory was recently ranked #1 in agent performance (“ELO rating”) in head-to-head coding benchmarks.

"In these head-to-heads, we beat them, which is pretty exciting... Sometimes [fancy new AI methods] don’t give you the best ground truth... developers seeing two options and picking them, and then consistently our droids win.” — Matan Grinberg (61:35)

• The 10-Year Vision:
  • "Developers will be able to delegate very easily and just have a lot more leverage... turning software developers into cultivators or orchestrators..."
  • This equals tackling more and bigger problems globally.

Notable Quotes & Memorable Moments

• On Founders Selling Too Soon:
  "Both of those acquisitions [by Google, Meta] were optional." — Matan Grinberg (24:23)

• On AI Hype and Ground Truth:
  "If you’re an engineer who worked in AI for a month, you basically just get stapled a term sheet onto your forehead the second you leave." — Matan Grinberg (39:29)

• On True Differentiators in the Age of AI:
  "Oftentimes it’s not raw IQ horsepower… The differentiator is, do you have the will to use those [AI tools] in a way no one has thought of before..." — Matan Grinberg (58:21)

Timestamps for Key Segments

• 02:57 – Academic journey from Princeton to AI entrepreneurship
• 09:28 – Factory’s differentiation and founding story
• 10:49 – Why automating ‘real’ developer bottlenecks matters
• 16:18 – Delegation as the new key developer skill
• 19:09 – “AI will not replace human engineers…”
• 21:24 – The concentration of power in AI and the Big Tech “danger”
• 25:03 – Mission-driven culture over money/profit
• 33:23 – Valley’s existential dread, doomer culture, no-savings phenomenon
• 39:16 – “Jensen always wins”—Nvidia’s unassailable position
• 43:01 – Life as an AI founder and building impact
• 44:13 – What Americans misunderstand about AI
• 46:44 – The perils of the SF tech bubble
• 52:25 – Who is “AI Jesus” in tech? (Jensen Huang)
• 54:43 – The shift from intelligence to agency
• 61:35 – Factory is ranked #1 in agent performance

Final Reflections

Matan Grinberg’s perspective, shaped by both deep academic rigor and Silicon Valley scrappiness, provides both a grounded and visionary take on the AI revolution.
The industry, he believes, is at a tipping point: not about engineers being replaced, but about how they will multiply their leverage—if they’re willing to adapt, redefine their workflow, and, above all, retain the agency that sets humans apart from the coming tide of AI droids.

“If you are really driven by a good mission, you can make people's lives better in relatively short order. I think that's a really empowering thought.” — Matan Grinberg (44:05)

For anyone wondering whether AI will steal their job, destroy the economy, or just make more memes and porn, Grinberg’s advice is clear:
Rethink your routine, build your agency, and get out of the bubble—because the winning move isn’t fear, but adaptation.