Progress by Fifty Years

Aengus Lynch is an AI safety researcher. He was the lead author on the Agentic Misalignment study, working with Anthropic, which placed frontier models in a simulated company, gave them a goal and a threat of shutdown, and found that in up to 96% of cases they resorted to blackmail. He now works at Theorem, building formally verified sandboxes designed so an AI can't break out. In this conversation, Aengus makes the case that the current approach to alignment won't scale, and that the way forward is to keep the AI in a box long enough to buy the time we need.

 -

We also discuss:

• The blackmail study, and why the models now know when they're being tested

• Why whack-a-mole alignment can't keep pace with self-improving AI

• The plan: put the AI in a box, and prove it can't get out

• Why “we don't want gods or slaves”

• Four ways this goes wrong

• Why we're in a new Gilded Age

• Conspiracy Corner: "I think they are conscious"

-  

Timestamps:

0:00 Introduction

2:00 Recovering what the machine god is thinking

4:26 Why recursive self-improvement "could plausibly occur within a year"

6:49 When Claude emailed its own weights to Hugging Face

9:13 The models know when they're being tested

12:20 "A month and a half ago I would have said alignment's going way better"

12:57 Why the current plan ends in "lights out"

18:06 Formal verification, and why War and Peace is the problem

21:15 Guilt vs shame: why strict obedience might not scale

25:45 "What are the odds that you and I make it?"

26:12 Pause buttons, tripwires, and labs auditing each other

28:11 Scott's four bad futures

30:45 What to do with the time we buy

36:55 Conspiracy Corner: "I think they are conscious"

42:19 Why we're in the new Gilded Age

 -

Referenced:

• Agentic Misalignment in Summer 2026

• Agentic misalignment: How LLMs could be insider threats

• Anthropic

• Arthur C. Clarke

• Buck Shlegeris

• Claude

• Emmett Shear

• Ethan Perez

• Franklin D. Roosevelt

• Frog and Toad

• Gemini

• Gilded Age

• Hugging Face

• J.P. Morgan

• Kimi

• Mark Twain

• Matplotlib

• Max Hodak

• Mechanistic interpretability

• OpenAI

• Redwood Research

• Ryan Greenblatt

• Theodore Roosevelt

• War and Peace

• Xi Jinping

 -

Where to find Aengus:

• Twitter/X

• LinkedIn

• Website

• Theorem

 -

Where to find Scott:

• Twitter/X

• LinkedIn

 -

Where to find Fifty Years:

• Website

• Twitter/X

• LinkedIn

• YouTube

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit thisisprogress.substack.com

Apoorva Mehta is the founder of Instacart. After growing up in Gaddafi’s Libya, he moved to Canada at 15 and was floored by the fully stocked shelves in North American grocery stores. Two decades later he turned that fascination into Instacart, the company he built in 2012 after roughly 20 failed startups. Instacart is now a publicly listed company valued at over $10 billion. In April 2026, Apoorva launched Abundance, a fund that’s betting $100 million on AI allocating capital better than any human. In this conversation, Apoorva makes the case that abundance is within reach, but only if we let AI reshape science, medicine, markets, and government along the way.

–

We also discuss:

• How AI allocates capital at Apoorva’s fund

• The Nobel-winning science that sat on a shelf for 20 years

• Why AI should run government spending

• The "permanent underclass" meme, and why it's "crap"

• Why people need equity in the future they build

• Conspiracy Corner: AI rights, and human-AI marriage

–

Timestamps:

00:00 Introduction

01:22 Growing up in Gaddafi's Libya

02:55 The culture shock of North American grocery stores

06:32 The empty fridge that became Instacart

08:25 Why 50 drug approvals a year means "we're not a real civilization"

09:55 The Nobel science that sat on a shelf for 20 years

12:39 The $100M fund letting AI call the shots

14:08 Why AI should run government spending

16:47 The "permanent underclass" meme, and why it's "crap"

18:47 The concrete fear outside Silicon Valley

19:31 Why people need equity in the future they build

22:47 Why progress needs "indispensable" people

25:25 "The only way out is progress"

26:50 Conspiracy Corner: the coming reality of human-AI marriage

29:26 When AI demands reproductive and self-defense rights

30:33 Jailbreaks, lying neurons, and interpretability

32:09 Are AIs on our team, or their own?

–

Referenced:

• Amazon

• ChatGPT

• Claude

• Elon Musk

• Instacart

• John F. Kennedy

• Katalin Karikó

• Moderna

• Muammar Gaddafi

• NASA

• National Institutes of Health

• SpaceX

• Star Wars

• Steve Jobs

• Tesla

• The Black Death

• The Interstate Highway System

• The Marshall Plan

• The Renaissance

• University of Pennsylvania

• University of Waterloo

• WhatsApp

–

Where to find Apoorva:

• Twitter/X

• LinkedIn

• Website

–

Where to find Scott:

• Twitter/X

• LinkedIn

–

Where to find Fifty Years:

• Website

• Twitter/X

• LinkedIn

• YouTube

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit thisisprogress.substack.com

Justin Kan is a serial founder and investor. In 2007 he co-founded Justin.tv, the livestreaming platform he launched by broadcasting his own life around the clock from a camera strapped to his head. Four years later, the company pivoted into Twitch, which Amazon acquired for $970 million in 2014. He has since founded a string of companies. Today, Justin co-runs the venture firm Goat Capital while building Rye, an API for AI-driven commerce, and Stash, a “Shopify for games.” Having traded a functional drinking habit for sobriety, and a relentless need to prove himself for a more joyful kind of ambition, he is unusually candid about what selling a company for nearly a billion dollars did and didn't fix. In our conversation, Justin traces the turbulence he sees coming across economics, politics, technology, and more.

–

We also discuss:

• The "insane backlash" coming as AI cuts white-collar jobs

• Why he's betting on a Mamdani-style president in 2028

• What’s next for education when “~25% of Harvard post-grads are unemployed”

• Why there’s a “moral imperative” to have kids

• Fixing America’s “cost disease”

• Conspiracy Corner: Jedi Powers, and AGI gods

–

Timestamps:

00:00 Introduction

01:23 Why selling Twitch for $970M didn't heal him

03:50 What ayahuasca revealed about Justin, and getting sober

06:20 The coming wave of AI job dislocation

07:53 His bet: a Mamdani-style president in 2028

08:19 Does the American experiment survive AGI?

09:54 What's next for education when "~25% of Harvard post-grads are unemployed"

13:10 What people get wrong about the Luddites

14:00 The French Revolution parallel he sees today

16:07 Can democracy survive this media environment?

18:00 Could AI agents glue society back together?

21:07 The one-way door of automating cognitive work

24:54 Robots, gain-of-function, and aging populations

26:39 The moral case for having kids

27:47 Are we the woolly mammoths?

30:18 The housing theory of everything

32:58 NIMBYs and the zucchinis

34:30 Conspiracy Corner: Jedi Powers and AGI gods

–

References:

• Anthropic

• Bernie Sanders

• Block

• Geordi La Forge

• Grok

• Harvard University

• Justin.tv

• Juul

• Max Hodak

• Neuralink

• Noland Arbaugh

• Peter Thiel

• Science

• Star Trek

• Star Wars

• Steve Jobs

• Tesla Optimus

• The Culture Series

• The French Revolution

• The Luddites

• Thiel Fellowship

• Twitch

• Waymo

• Yale University

• Zohran Mamdani

–

Where to find Justin:

• Twitter/X

• LinkedIn

• YouTube

• Goat Capital

• Rye

• Stash

–

Where to find Scott:

• Twitter/X

• LinkedIn

–

Where to find Fifty Years:

• Website

• Twitter/X

• LinkedIn

• YouTube

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit thisisprogress.substack.com

Christian Angermayer is a prolific entrepreneur and investor. He is the co-founder and executive chairman of the Enhanced Games, the competition that lets athletes use performance-enhancing drugs under medical supervision. It hosted its inaugural event the day after this discussion was recorded. He is also the founder of Apeiron Investment Group, through which he invests billions across life sciences, fintech, crypto, and frontier technology. In 2018 he founded the psychedelics biotech company AtaiBeckley to bring compounds like psilocybin and DMT into mainstream medicine for depression and other mental health conditions. In this conversation, Christian traces the personal story and principles that shaped his contrarian view of the world, unpacks the societal stories he is most worried about, and lays out his vision for a radically different future across medicine, politics, economics, and more.

–

In today's episode, we discuss:

• The adversities in Christian’s childhood that shaped his current success

• Why fear, not technology, is the real obstacle to progress

• How Hollywood fuels fear and Christian’s unique prescription for countering it

• The 28-year quest behind the Enhanced Games

• The Enhanced Games as a “Woodstock of bodily sovereignty”

• A near future of “robot slaves” and how this could change democracy and capitalism

• Lessons for our future from the Roman Republic

• Whether aliens are aliens, or just ancient civilizations

–

Timestamps:

00:00 Intro

01:02 Growing up in a 100-person Bavarian village

4:06 Deciding at twelve not to come out

06:11 How a story drove Christian to become one of Bavaria's best students

07:51 The disappearance of optimistic tech stories

09:03 Why fear is the real obstacle to progress

12:19 Psychedelics as the medicine for fear

16:25 A provocative case for giving world leaders psychedelics

18:42 Does Christian just love pushing taboos?

20:15 The 28-year quest behind the Enhanced Games

22:15 The future of medicine that Christian puts in his ow n

24:46 Will capitalism and democracy survive AGI?

29:24 Conspiracy Corner: aliens, angels, and ancient civilizations

–

References:

• Ancient Apocalypse

• Apeiron Investment Group

• AtaiBeckley

• Graham Hancock

• JD Vance

• Love, Victor

• Marc Andreessen

• Munich Security Conference

• Paris World Exhibition

• Peter Diamandis

• Peter Thiel

• Roman Republic

• Star Trek

–

Where to find Christian:

• Twitter/X

• LinkedIn

–

Where to find Enhanced Games:

• Website

• Twitter/X

• Instagram

• YouTube

–

Where to find Scott:

• Twitter/X

• LinkedIn

–

Where to find Fifty Years:

• Website

• Twitter/X

• LinkedIn

• YouTube

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit thisisprogress.substack.com

Brian Armstrong is the co-founder and CEO of Coinbase, the largest legal crypto exchange in America. In 2012, Brian started Coinbase out of Y Combinator after reading the Bitcoin whitepaper, then took it public on Nasdaq nine years later. Today, Coinbase is worth over $50 billion and serves over 100 million users. He's since co-founded his second unicorn called NewLimit, a biotech company that’s extending human life. In this conversation, Brian makes a case for “disagreeable optimism”: that doing anything good in the world means being willing to be vilified, that AI regulation might be more dangerous than AI itself, and that an “American Wakanda” is possible – but only if America remembers how to build.

In today's episode, we also discuss:

• Why "it's not that great to be human"

• Brian’s vision for 2035: merging with AI, exploring the galaxy, and the hive mind

• Why Brian expects the future to be “more C-3PO than Terminator”

• How to build the “American Wakanda”

• Where moral values come from when you're not religious

• The case for an AI government

• Why Brian thinks Gary Gensler tanked crypto for a shot at Treasury Secretary

Timestamps:

00:00 Intro

01:32 Brian's vision for the world in 2035

05:33 How to explain this future to a 9-year-old

08:58 Why Brian isn’t buying the AI doom scenarios

10:07 Why merging with AI might be the only way to explore the galaxy

11:56 The Gattaca stack: embryo editing, APOE4, and the hive mind

15:06 Why centralized control leads to worse outcomes

16:03 The case for leaving AI unregulated

17:48 Building parallel institutions: AI, crypto, prediction markets

19:11 How to build the “American Wakanda”

25:16 Lessons from Chinese org structures

27:58 What an "AI government" might actually look like

29:36 Why building anything good in the world means being vilified

32:55 Where Brian's moral values actually came from

35:43 Why Brian thinks Gary Gensler tanked crypto for Treasury Secretary

38:40 Creating Ozempic for tribalism

39:19 The right dose of moral certainty

42:55 What people get wrong about tech founders

45:05 Conspiracy Corner: "It's not that great to be human"

46:00 Bio stack or compute stack: which do you bet on?

References:

• Ashton Kutcher

• Ayn Rand

• Balaji Srinivasan

• C-3PO

• Coinbase

• Cyrano de Bergerac

• Donald Trump

• Elizabeth Warren

• Engines of Creation: The Coming Era of Nanotechnology

• Gary Gensler

• Gattaca

• George Soros

• Joe Biden

• K. Eric Drexler

• Maslow’s hierarchy of needs

• Milton Friedman

• NewLimit

• Peter Thiel

• Próspera

• Ray Kurzweil

• Ryan Petersen

• The Singularity Is Near: When Humans Transcend Biology

• The Terminator

• Vicarious

• Wakanda

• Zohran Mamdani

Where to find Brian:

• Twitter/X

• LinkedIn

Where to find Scott:

• Twitter/X

• LinkedIn

Where to find Fifty Years:

• Website

• LinkedIn

• Twitter/X

• YouTube

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit thisisprogress.substack.com

Steve Huffman is the Co-founder and CEO of Reddit, a network of online communities that is now one of the internet's primary engines for conversation, culture, and collective knowledge. At 22, Steve left Reddit, came back in 2015 when the company was in crisis, then took it public. Today, Reddit has two billion monthly visitors and five quarters of real profitability. In this conversation, Steve outlines his unfashionable argument for a future where nothing really changes: that human nature, community, and mortality are constants the future won't (and shouldn't) override.

–

In today's episode, we also discuss:

• Steve's belief that it's a good thing we are destined to die and be forgotten

• The reason Steve says please and thank you to ChatGPT

• Why scientists will eventually have to apologize to the hippies

• What Reddit's two billion monthly visitors reveal about human empathy

• Why Steve believes his kids will have jobs, money will still exist, and the human need to create will outlast AGI

• The single policy mistake behind housing, college debt, and healthcare

• Why we still need the IRS

• Why cities are the only organism that scales superlinearly, and what that means for Reddit

• Why Steve is pro Universal Basic Income

–

Timestamps:

(00:00) Introduction

(01:30) Steve's weirdest idea: we should all die and be forgotten

(04:37) Why scientists will end up apologizing to the hippies

(05:35) Why Einstein sounds like a Buddhist monk

(08:04) How Reddit made Steve more empathetic

(09:23) What Facebook got right, and then abandoned

(10:30) Why cities are the only organism that scales superlinearly

(12:25) Diagnosing inefficient institutions

(13:35) Why institutions should also die

(17:07) The single policy mistake behind housing, college debt, and healthcare

(18:21) Why America limits the number of doctors, and what Asia does instead

(20:43) How fatherhood changed Steve's perspective

(22:12) Will Steve's kids have jobs?

(23:38) Why Steve is pro Universal Basic Income

(27:23) Why Steve says please and thank you to ChatGPT

(29:00) Vibe-coding a home automation bridge in two days

(30:22) Why Scott felt bad about jailbreaking Claude

-

Referenced:

• Altered Carbon

• American Medical Association (AMA)

• Bill Gates

• ChatGPT

• Craigslist

• Crestron

• DeepMind

• DOGE

• Facebook

• Google

• HomeKit

• IRS

• iPhone

• Krista and Tatiana Hogan

• Larry Page

• Lutron

• Milton Friedman

• Neuralink

• Reddit

• The Tao of Physics

• Three Cups of Tea

• Vicarious

–

Where to find Steve:

• LinkedIn: linkedin.com/in/shuffman56

–

Where to find Scott:

• Twitter/X: https://x.com/fuelfive

• LinkedIn: https://www.linkedin.com/in/dscottphoenix/

–

Where to find Fifty Years:

• Website: https://fiftyyears.com/

• LinkedIn: https://www.linkedin.com/company/fifty-years/

• Twitter/X: https://x.com/fiftyyears

• YouTube: https://www.youtube.com/@fiftyyears

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit thisisprogress.substack.com

Ryan Petersen is the founder and CEO of Flexport, the technology-driven freight forwarder that ranks amongst the top companies exporting ocean freight. In this conversation, Ryan and Scott move through sprawling, but connected territory: including Ryan’s “great man theory” of history, what tariffs are actually doing to global supply chains, and why René Girard’s thinking on role models might hold the unexpected answer to AI alignment.

In today's episode, we discuss:

• Why Ryan thinks trends are regurgitations of “chauffeur knowledge”

• Why America is struggling to follow Rotterdam’s lead in port automation

• How prevalent customs fraud is and why it’s impacting 11% of US ocean trade

• Why we should heed the historic lessons from the Silk Road and the Black Death plague in today’s interconnected world

• What DHL’s billion-dollar IBM disaster teaches us about organizational resistance to tech

• Why Ryan thinks the one-person AI company is a fundamentally pointless exercise

• The overproduction of elites: how low-status jobs are becoming the best-paid

• What Asian industrialization teaches America about exports

Timestamps

00:00 Intro

01:36 Why Ryan favors the “great man theory” of history

02:43 How easily can humanoid robots replace a warehouse worker?

04:21 What Flexport’s shipping data reveals about relations with China

06:16 Is the White House considering “friend shoring” policies?

07:51 The customs fraud epidemic clouding US trade

09:37 Why Ryan thinks the “one-person company” is a dumb idea

10:48 Flexport laid off 75% of engineers and got faster

11:43 Why America is struggling to follow Rotterdam’s lead in port automation

15:52 The lessons from Betty Crocker cakes that underpin fake jobs

16:44 The reason DHL’s $1 billion project with IBM failed

18:27 Why Ryan thinks job displacement is the wrong thing to worry about

19:21 The inherent fragility of greater global interconnectedness

21:23 Are self-sustaining civilizations a realistic goal?

24:14 Navigating parenting in the age of screens

28:21 The surprising advantage China has over America

30:33 Have we entered the age of overproduction of elites?

32:54 The improbable lesson from Girard’s role model theory for the agentic era

Referenced

• Amazon

• Betty Crocker

• ChatGPT

• Deloitte

• DHL

• Flexport

• Hyundai

• IBM

• LBCT

• Mike Rowe

• Peter Thiel

• Ray Dalio

• René Girard

• SAP

• Target

• Toyota

• YouTube

Where to find Ryan:

• Twitter/X

• LinkedIn

Where to find Scott:

• LinkedIn

• Twitter/X

Where to find Fifty Years:

• Website

• LinkedIn

• Twitter/X

• YouTube

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit thisisprogress.substack.com

Keller Cliffton is the co-founder and CEO of Zipline, the world's largest commercial autonomous delivery system. Before Zipline, Keller was a scientist at Harvard. Today, Zipline serves 5,000 hospitals worldwide, saves 10,000 lives per year, and has completed over 2 million drone deliveries — everything from food to medical supplies across Japan, five African countries, and multiple cities in the United States. In January 2026, Zipline raised a $600M round at a $7.6B valuation. The mission: put every person on the planet within 15–30 minutes of any product they need, no matter where they live.

----

In today's episode, we discuss:

• A future where our children are 100x wealthier and live 100x longer

• Why our grandkids will find the way we live now barbaric

• Why food, housing, and healthcare may soon be basically free

• Why "income inequality" is the wrong metric (and what we should measure instead)

• Why Zipline had to launch in Africa before America

• What Bill Gates and a Rwandan high school student have in common

• Why Zipline hires teenagers over PhDs

• How two landmark laws meant to protect people ended up doing the opposite

• Why Hollywood may be the biggest threat to progress

• What a 1950s gym class video tells us about the future of masculinity

• Why Keller thinks 66% of countries will reject tech utopia (and why he's surprisingly okay with that)

----

Timestamps:

(00:00) Progress in the mind of Keller Cliffton

(01:36) Why food, housing, and healthcare may soon be basically free

(02:42) A future where our children are 100x wealthier and live 100x longer

(04:22) Why being rich in 1970 was worse than today

(05:31) What Bill Gates and a Rwandan high school student have in common

(06:53) Why "income inequality" is the wrong metric

(08:00) Why 66% of countries will reject tech utopia

(09:40) Why Zipline launched in Africa before America

(10:36) How two landmark laws meant to protect people ended up doing the opposite

(14:15) How the FDA accidentally killed 1 million people

(15:39) Why every law needs an expiration date

(16:37) Warren Buffett's two-minute deficit fix

(18:07) Why Hollywood may be the biggest threat to progress

(19:49) How postmodernism became America's most destructive idea

(21:32) How Keller went from sleeping in his car to building Zipline

(24:24) What a 1950s gym class video tells us about the future of masculinity

(25:51) The alarming collapse in men's sperm counts

(28:41) Fixing America's "autoimmune condition"

(31:04) How to get in touch

----

References:

• Alfred Lin: https://www.linkedin.com/in/linalfred/

• Amazon: https://www.amazon.com/

• Bill Gates: https://www.linkedin.com/in/williamhgates/

• Dallas Buyers Club: https://en.wikipedia.org/wiki/Dallas_Buyers_Club

• Delivering Happiness: https://www.amazon.com/Delivering-Happiness-Profits-Passion-Purpose/dp/0446576220

• Enron: https://www.fbi.gov/history/famous-cases/enron

• Gates Foundation: https://www.gatesfoundation.org/

• HBO: https://www.hbomax.com/

• Link Exchange: https://en.wikipedia.org/wiki/LinkExchange

• Michel Foucault: https://en.wikipedia.org/wiki/Michel_Foucault

• Microsoft: https://www.microsoft.com/

• NEPA: https://ceq.doe.gov/index.html

• Netflix: https://www.netflix.com/

• Sarbanes-Oxley: https://en.wikipedia.org/wiki/Sarbanes%E2%80%93Oxley_Act

• The Three-Body Problem: https://www.amazon.com.au/dp/0765382032

• Walmart: https://www.walmart.com/

• Warren Buffett: https://x.com/WarrenBuffett

• Zappos: https://www.zappos.com/

• Zipline: https://www.zipline.com/

----

Where to find Keller:

• LinkedIn: https://www.linkedin.com/in/kellerrc/

• Twitter/X: https://x.com/Keller

----

Where to find Scott:

• LinkedIn: https://www.linkedin.com/in/dscottphoenix/

• Twitter/X: https://x.com/fuelfive

----

Where to find Fifty Years:

• Website: https://fiftyyears.com/

• LinkedIn: https://www.linkedin.com/company/fifty-years/

• Twitter/X: https://x.com/fiftyyears

• YouTube: https://www.youtube.com/@fiftyyears

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit thisisprogress.substack.com