OpenAI’s hottest app isn’t ChatGPT—it’s Codex.

In the last few weeks alone, the Codex team shipped a desktop app, GPT-5.3 Codex (a new flagship model), and Spark, the fastest coding model I’ve ever used. Usage has grown fivefold since January, and over a million people now use Codex weekly. Codex was also the app that OpenAI chose to run an ad for in the Super Bowl.

Dan Shipper talked to Thibault Sottiaux, head of Codex, and Andrew Ambrosino, a member of technical staff who built the Codex app, for Every’s AI & I about what OpenAI is building and how they’re using it internally.

If you found this episode interesting, please like, subscribe, comment, and share! 

Want even more?

Sign up for Every to unlock our ultimate guide to prompting ChatGPT here: https://every.ck.page/ultimate-guide-to-prompting-chatgpt. It’s usually only for paying subscribers, but you can get it here for free.

To hear more from Dan Shipper:

• Subscribe to Every: https://every.to/subscribe 

• Follow him on X: https://twitter.com/danshipper 

Head to granola.ai/every and get 3 months free with the code EVERY.

Timestamps:  

00:00:00 - Start

00:01:27 - Introduction 

00:05:27 - OpenAI's evolving bet on its coding agent 

00:09:42 - The choice to invest in a GUI (over a terminal) 

00:20:38 - The AI workflows that the Codex team relies on to ship 

00:26:45 - Teaching Codex how to read between the lines 

00:28:45 - Building affordances for a lightening fast model 

00:33:15 - Why speed is a dimension of intelligence 

00:36:30 - Code review is the next bottleneck for coding agents 

00:41:24 - How the Codex team positions against the competition 

Links to resources mentioned in the episode:

• Thibault Sottiaux: Tibo (@thsottiaux)

• Andrew Ambrosino: Andrew Ambrosino (@ajambrosino)

Every’s vibe check on everything the Codex team launched: OpenAI's Codex App Gains Ground on Claude Code, GPT-5.3 Codex—The 10x Engineer, Now More Fun at Parties, AI as Fast as Your Train of Thought

The AI labs fighting for attention during the Super Bowl call to mind another iconic Super Bowl moment: Apple’s 1984 ad for the Macintosh, which promised that the personal computer would be a source of unbound wonder, freedom, and delight.

They were right, but over time, the personal computer has also become cluttered with errands.

These “computer errands”—downloading a W-2 when tax season rolls around, hunting for the right coupon code before checkout, or navigating the unholy labyrinth of the Amazon Web Services dashboard just to change one permission setting—have taken over our digital lives. Atlas, OpenAI’s agentic browser, sprang from the idea that AI should handle this tedium for you.

In this week’s episode of AI & I, Dan Shipper sat down with two members of the Atlas team, Ben Goodger and Darin Fisher. Goodger is Atlas’s head of engineering, and Fisher is a member of the technical staff. Both are legends of the browser world. They’ve spent decades building the modern web, working together on Netscape, Firefox, and Chrome before arriving at Atlas. From that vantage point, they told Dan how they think browsing is about to change, why building a browser is harder than it looks, and what it’s like to create a new one with AI coding tools like Codex.

If you found this episode interesting, please like, subscribe, comment, and share! 

Want even more?

Sign up for Every to unlock our ultimate guide to prompting ChatGPT here: https://every.ck.page/ultimate-guide-to-prompting-chatgpt. It’s usually only for paying subscribers, but you can get it here for free.

To hear more from Dan Shipper:

• Subscribe to Every: https://every.to/subscribe 

• Follow him on X: https://twitter.com/danshipper 

Move fast, don’t break things

Most AI coding tools don’t know which line of code will actually break your system. Try Augment Code, which understands your entire codebase, including the repos, languages, and dependencies that actually runs your business, and use their playbook to learn more about their framework, checklists, and assessments. Ship 30% faster with 40% shorter merge times.

[Playbook at https://www.augmentcode.com/]

Timestamps:  

00:01:57 - Introduction

00:11:51 - Designing an AI browser that’s intuitive to use

00:15:24 - How the web changes if agents do most of the browsing

00:25:06 - Why traditional websites will not become obsolete

00:29:00 - A browser that stays out of the way versus one that shows you around

00:39:51 - How the team uses Codex to build Atlas

00:44:47 - The craft of coding with AI tools

00:52:33 - Why Goodger and Fisher care so much about browsers

Links to resources mentioned in the episode:

• Ben Goodger: Ben Goodger (@bengoodger) 

• Darin Fisher: Darin Fisher (@darinwf) 

OpenAI’s browser, Atlas: Introducing ChatGPT Atlas

A few weeks ago, Natalia Quintero wouldn’t have called herself technical. But since the beginning of January, she has woken up at 6 a.m. to vibe code with Claude. The AI project manager she built saved her 14 hours a week. 

Getting there meant scrapping the system three times and starting over. But the result handles everything from onboarding new clients to generating weekly updates across all projects.

Natalia is the head of AI consulting at Every. As part of the role, she's spoken with over 100 organizations in the past year and worked with a select two dozen, including hedge funds, private equity firms, and Fortune 500 companies. She’s seen what separates companies thriving with AI from those floundering, and it comes down to patterns that have nothing to do with having the most resources or the fanciest tools.

Dan Shipper had her on AI & I to share what she’s learned from this front-row seat to AI adoption. Quintero reveals how a private equity firm cut investment memo creation from three weeks to 30 minutes, why AI adoption needs to come from the top down, and what happened when she learned from her early morning experiments.

She also explains why the companies going furthest with AI are the ones that give employees permission to fail—and how that counterintuitive approach is revolutionary.

If you found this episode interesting, please like, subscribe, comment, and share! 

Want even more?

Sign up for Every to unlock our ultimate guide to prompting ChatGPT here: https://every.ck.page/ultimate-guide-to-prompting-chatgpt. It’s usually only for paying subscribers, but you can get it here for free.

To hear more from Dan Shipper:

• Subscribe to Every: https://every.to/subscribe 

• Follow him on X: https://twitter.com/danshipper 

Ready to build a site that looks hand-coded—without hiring a developer? Launch your site for free at www.Framer.com, and use code DAN to get your first month of Pro on the house.

Timestamps:  

00:00:00 - Introduction

00:01:30 - Why successful AI adoption requires coordinated, top-down effort

00:07:05 - How a private equity firm reduced investment memo creation from weeks to 30 minutes

00:13:30 - The benefits of connecting AI to proprietary context

00:15:20 - The plan-delegate-assess-compound framework for engineering teams

00:17:55 - How non-technical team members are becoming vibe coding addicts

00:20:50 - Building Claudie: an AI project manager from scratch

00:23:00 - Why creative exploration time outside the 9-to-5 is essential

00:27:50 - Live demo: How Claudie automates client onboarding and tracking

00:38:40 - The human side of AI: spending less time in spreadsheets, more time with people

Links to resources mentioned in the episode:

• Natalia Quintero: Natalia Quintero (@NataliaZarina)

• What Natalia learned from working with companies on AI adoption: https://every.to/on-every/the-next-chapter-of-every-consulting

Every’s compound engineering plugin: https://github.com/EveryInc/compound-engineering-plugin

Entrepreneur Andrew Wilkinson used to sleep nine hours a night. Now he wakes up at 4 a.m. and goes straight to work—because he can’t wait to keep building with Anthropic’s latest model, Opus 4.5.

Two years ago, Wilkinson was obsessed with vibe coding on AI software development platform Replit. It was thrilling to describe something in plain English and watch an app appear, less thrilling when the apps were always broken in some way, often full of maddening bugs. So he set his app creation ambitions aside until technology caught up with them.

Then, a few weeks ago, he started playing with Claude Code and Opus 4.5. It felt, he says, like having a “$100,000-a-month payroll of engineers” working for him around the clock.

Wilkinson is the cofounder of Tiny, a company that buys profitable businesses and holds them for the long term. The Tiny portfolio includes the AeroPress coffee maker and Dribbble, a platform where designers can share their work and find jobs. Dan Shipper had him on AI & I to talk about the automations Wilkinson has built for his work and personal life, including an AI relationship counselor, a custom email client, and a system that texts him outfit recommendations each morning. Wilkinson revealed how all of this individual exploration has changed the way he thinks about buying software companies at Tiny.

If you found this episode interesting, please like, subscribe, comment, and share!

Want even more?

Sign up for Every to unlock our ultimate guide to prompting ChatGPT here: https://every.ck.page/ultimate-guide-to-prompting-chatgpt. It’s usually only for paying subscribers, but you can get it here for free.

To hear more from Dan Shipper:

Subscribe to Every: https://every.to/subscribe

Follow him on X: https://twitter.com/danshipper

Ready to build a site that looks hand-coded—without hiring a developer? Launch your site for free at framer.com, and use code DAN to get your first month of Pro on the house!

Timestamps:

00:00:00 - Start

00:01:07 - Introduction

00:02:48 - Why Opus 4.5 feels like the iPhone moment for vibe coding

00:08:31 - Why designers have a unique advantage with AI

00:14:10 - How Wilkinson built a custom email client with Claude Code

00:18:13 - An AI trained on your relationship that predicts your fights

00:30:40 - Using AI meeting notes to make your life better

00:35:11 - Don't inject your opinion into prompts

00:40:21 - Wilkinson’s Claude Code tips and workflows

00:47:59 - Your personal stylist is a prompt away

00:53:17 - How AI is changing the way Wilkinson invests in software

Links to resources mentioned in the episode:

Andrew Wilkinson: Andrew Wilkinson (@awilkinson)

The book Wilkinson references in his prompts, when writing copy with AI: Made to Stick

Every’s compound engineering plugin: https://github.com/EveryInc/compound-engineering-plugi

LLMs have made it absurdly easy to go deep on almost any topic. So why haven’t we all used ChatGPT to earn college degrees we wished we had majored in or pursued a niche interest, like learning how to name the trees in our neighborhood? I know I’m not the only one to feel guilty for well-intentioned attempts at autodidactism that inevitably peter out.

Entrepreneur Nir Zicherman has a reason for this disconnect: LLMs can answer most of your questions, but they won’t notice when you’re lost or pull you back in when your motivation starts to fade.

As the CEO and cofounder of Oboe, a platform that generates personalized courses about everything from the history of snowboarding to JavaScript fundamentals using AI, Zicherman has thought deeply about why the ability to access information does not automatically lead to understanding a concept. In this episode of AI & I, he talks to Dan Shipper about everything he’s learned about learning with LLMs.

They get into Zicherman’s counterintuitive belief that learning is a more passive process than you’d think, the biggest blocker for most people who want to learn something new, and where AI agents currently fall short in providing a meaningful learning experience.

If you found this episode interesting, please like, subscribe, comment, and share!

Want even more?

Sign up for Every to unlock our ultimate guide to prompting ChatGPT here: https://every.ck.page/ultimate-guide-to-prompting-chatgpt. It’s usually only for paying subscribers, but you can get it here for free.

To hear more from Dan Shipper:

Subscribe to Every: https://every.to/subscribe

Follow him on X: https://twitter.com/danshipper

Timestamps:

00:00:00 - Start

00:00:36 - Introduction

00:01:49 - Why you need a dedicated AI learning app

00:04:32 - The process of learning is more passive than you might think

00:10:21 - Live demo of Oboe to create a course about philosopher Ludwig Wittgenstein

00:16:52 - Learning works best when it comes in many formats

00:28:21 - Where AI agents currently fall short in the learning experience

00:34:10 - The importance of making learning feel accessible

00:35:56 - How Zicherman uses Oboe to learn quantum physics

00:40:54 - How embeddings spaces remind Dan of quantum mechanics

Links to resources mentioned in the episode:

Nir Zicherman: @NirZicherman

Learn something new with Oboe: https://oboe.com/

Anthropic just dropped Claude Cowork—essentially Claude Code for everyone, not just engineers—and we got to chat about it with a product engineer at Anthropic who helped build it.

In this live Vibe Check, Dan Shipper and Kieran Klaassen explore the new interface together, testing what works (and what doesn't) in real time. Anthropic’s Felix Rieseberg joins midway through to explain the philosophy behind Cowork's design: why it separates "Tasks" from "Chats," how the queue system lets you send messages while the agent is working, and what "agent-native" architecture means in practice. They also dig into Skills—Claude's prompt system that lets you customize how it works—and the Chrome connector for browser automation.

This is a raw, unfiltered first look at what might be the future of how knowledge workers interact with AI: async workflows instead of turn-by-turn chat.

If you found this episode interesting, please like, subscribe, comment, and share!

Want even more?

Check out Dan's guide to building agent-native applications: https://every.to/guides/agent-native

To hear more from Dan Shipper:

Subscribe to Every: https://every.to/subscribe

Follow him on X: https://twitter.com/danshipper

00:01:00 - What is Claude Cowork

00:02:36 - First demo: competitor analysis

00:03:33 - Email drafting that sounds like me

00:06:18 - Calendar audit running for an hour

00:07:39 - Book taxonomy demo

00:08:42 - PostHog analytics via Chrome browsing

00:14:36 - Chat vs Code vs Cowork: when to use what

00:31:06 - Felix from Anthropic joins

00:36:39 - Why they built it in a week and a half

00:37:57 - Design decision: why a separate tab

00:43:57 - Skills as the primary hackable surface

00:49:36 - Agent-native architecture principles

00:56:57 - The origin story of skills at Anthropic

01:03:00 - Our final rating

From cofounding LinkedIn to backing OpenAI early, Reid Hoffman is in the habit of being right about the future, so we wanted to know what he saw coming in 2026.

In his third appearance on AI & I, Hoffman lays out his predictions for where AI will go in the 12 months ahead. He talks to Dan Shipper about how agents will break out of coding into other domains and who’s winning the coding agent race. They also get into how Hoffman defines artificial general intelligence, the way he believes enterprises will use AI, and why public debate on AI might turn more negative, even as the technology becomes more empowering for individuals.

Hoffman’s other bets on the future include cofounding AI drug discovery startup Manas AI, investing at venture capital firm Greylock Partners, writing books, and hosting the Masters of Scale podcast. He’s also an investor at Every.

If you found this episode interesting, please like, subscribe, comment, and share!

Want even more?

Sign up for Every to unlock our ultimate guide to prompting ChatGPT here: https://every.ck.page/ultimate-guide-to-prompting-chatgpt. It’s usually only for paying subscribers, but you can get it here for free.

To hear more from Dan Shipper:

• Subscribe to Every: https://every.to/subscribe

• Follow him on X: https://twitter.com/danshipper

Timestamps:

00:00:00 - Start

00:00:52 - Introduction

00:02:20 - The future of work is an entrepreneurial mindset

00:05:22 - Creation is addictive (and that’s okay)

00:09:22 - Why discourse around AI might get uglier this year

00:17:03 - AI agents will break out of coding in 2026

00:24:18 - What makes Anthropic’s Opus 4.5 such a good model

00:28:46 - Who will win the agentic coding race

00:36:13 - Why enterprise AI will finally land this year

00:43:16 - How Hoffman defines AGI

00:55:33 - The most underrated category to watch in AI right now

Links to resources mentioned in the episode:

• Reid Hoffman: Reid Hoffman (@reidhoffman)

The AI drug discovery startup Hoffman cofounded: Manas AI

Tomorrow is the first day of 2026, and to give our listeners a view of the trends that’ll shape the year ahead, Dan Shipper had Every COO Brandon Gell on AI & I to discuss their predictions for what’s next. They discussed how software will be built, who will build it, and what it will take for truly autonomous AI agents to become a reality.If you found this episode interesting, please like, subscribe, comment, and share! Want even more?Sign up for Every to unlock our ultimate guide to prompting ChatGPT here: https://every.ck.page/ultimate-guide-to-prompting-chatgpt. It’s usually only for paying subscribers, but you can get it here for free.To hear more from Dan Shipper:Subscribe to Every: https://every.to/subscribe Follow him on X: https://twitter.com/danshipper Timestamps: 00:00:00 — Start00:01:05 — Introduction00:01:34 — Reflections on Every’s growth over the past year00:09:38 — What changes when a company grows from 20 to 50 people00:11:55 — How agent-native architecture will change software in 202600:17:13 — Why designers are slated to become power users of AI00:23:24 — The new kind of software engineer who will direct AI agents00:33:42 — Why the next wave of AI training will focus on autonomy

Learn how to use philosophy to run your business more effectively.

Reid Hoffman thinks a masters in philosophy will help you run your business better than an MBA. Reid is a founder, investor, podcaster, and author. But before he did any of these things, he studied philosophy—and it changed the way he thinks. Studying philosophy trains you to think deeply about truth, human nature, and the meaning of life. It helps you see the big picture and reason through complex problems—invaluable skills for founders grappling with existential questions about their business.

I usually bring guests onto my podcast to discuss the actionable ways in which people have incorporated ChatGPT into their lives. But this episode is different. I sat down with Reid to tackle a deeper question: How is AI changing what it means to be human?

It was honestly one of the most meaningful shows I’ve recorded yet. We dive into:

- How philosophy prepares you to be a better founder

- The importance of interdisciplinary thinking

- Essentialism v. nominalism in the context of AI

- How language models are evolving to be more “essentialist”

- The co-evolution of humans and technology

Reid also shares actionable uses of ChatGPT for people who want to think more clearly, like:

- Input your argument and ask ChatGPT for alternative perspectives

- Generate custom explanations of complex ideas

- Leverage ChatGPT as an on-demand research assistant

This episode is a must-watch for anyone curious about some of the bigger questions prompted by the rapid development of AI.

If you found this episode interesting, please like, subscribe, comment, and share!

Want even more?

Sign up for Every to unlock our ultimate guide to prompting ChatGPT here: https://every.ck.page/ultimate-guide-to-prompting-chatgpt. It’s usually only for paying subscribers, but you can get it here for free.

To hear more from Dan Shipper:

Subscribe to Every: https://every.to/subscribe

Follow him on X: https://twitter.com/danshipper

Ready to build a site that looks hand-coded—without hiring a developer? Launch your site for free at framer.com, and use code DAN to get your first month of Pro on the house!

Timestamps:

00:00:00 - START

00:04:35 - Why philosophy will make you a better founder

00:08:22 - The fundamental problem with “trolley problems”

00:14:27 - How AI is changing the essentialism v. nominalism debate

00:29:33 - Why embeddings align with nominalism

00:34:26 - How LLMs are being trained to reason better

00:44:52 - How technology changes the way we see ourselves and the world around us

00:46:24 - Why most psychology literature is wrong

00:52:46 - Why philosophers didn’t come up with AI

00:56:30 - How to use ChatGPT to be more philosophically inclined

Links to resources mentioned in the episode:

Reid Hoffman: https://twitter.com/reidhoffman

The podcasts that Reid hosts: Possible (possible.fm) and Masters of Scale (https://mastersofscale.com/)

Reid’s book: Impromptu https://www.impromptubook.com/

The book Reid recommends if you want to be more philosophically inclined: Gödel, Escher, Bach https://www.amazon.com/G%C3%B6del-Escher-Bach-Eternal-Golden/dp/0465026567

Reid’s article in the Atlantic: "Technology Makes Us More Human" https://www.theatlantic.com/ideas/archive/2023/01/chatgpt-ai-technology-techo-humanism-reid-hoffman/672872/

The book about why psychology literature is wrong: The WEIRDest People in the World by Joseph Henrich https://www.amazon.com/WEIRDest-People-World-Psychologically-Particularly/dp/0374173222

The book about how culture is driving human evolution: The Secrets of Our Success by Joseph Henrich https://press.princeton.edu/books/paperback/9780691178431/the-secret-of-our-success

We recorded someone guide himself into a Jhana live on our podcast. And he narrated the whole process from start to finish.

Jhanas are meditative bliss states and they traditionally require thousands of hours of practice. But Stephen Zerfas and his team at Jhourney are changing that—creating retreats where most participants hit a Jhana in their first week.

Dan Shipper went to one of their retreats earlier this year, and it was by far the best he’s been to. So we had Stephen on AI & I to show us how he gets into a Jhana and what the future of super wellbeing might look like.

If you found this episode interesting, please like, subscribe, comment, and share!

Want even more?

Sign up for Every to unlock our ultimate guide to prompting ChatGPT here: https://every.ck.page/ultimate-guide-to-prompting-chatgpt. It’s usually only for paying subscribers, but you can get it here for free.

To hear more from Dan Shipper:

• Subscribe to Every: https://every.to/subscribe
• Follow him on X: https://twitter.com/danshipper

Timestamps:

Introduction: 00:00:56

A primer on Jhana meditation: 00:01:18

Zerfas guides himself into a Jhana: 00:05:47

Why Jhana is about resting into what already exists: 00:36:04

Approaching meditation with play and curiosity: 00:39:30

The potential pitfalls of Jhana meditation: 00:45:04

How to hack your personality through memory reconsolidation: 00:48:21

Why Jhana won't let you numb yourself to real problems: 00:53:10

How Jhana meditation has changed Zerfas: 00:55:36

How Jhourney is using AI to make Jhanas more accessible: 01:09:41

Links to resources mentioned in the episode:

• Stephen Zerfas: Stephen Zerfas
• Jhourney: https://www.jhourney.io/
• The Donella Meadows book Zerfas refers to: Thinking in Systems: A Primer

Sarah Rose Siskind is incubating two types of intelligence at once: her unborn child, and FetusGPT—an LLM trained on nothing but what she hears and says throughout the day.

This includes Seinfeld episodes, YouTube videos about lemurs, eight hours of snoring per night—and even conversations with me, all condensed into MP3 and text files that are used to train the AI. Since FetusGPT is learning English from such a narrow, idiosyncratic slice of the world, it mostly babbles right now, and if she swears, it picks that up too.

FetusGPT is one zany example of how Siskind uses humor to make a bigger point: AI is what we make of it. It’s an approach that feeds through her comedy writing and work as the founder of science and technology communications agency Hello SciCom.

We had Siskind on AI & I to talk about how she uses AI in her creative process as a comedian, and the unexpected support it's become, both practical and emotional, as she navigates pregnancy.

Want even more?

Sign up for Every to unlock our ultimate guide to prompting ChatGPT here: https://every.ck.page/ultimate-guide-to-prompting-chatgpt.

It’s usually only for paying subscribers, but you can get it here for free.

To hear more from Dan Shipper:

Subscribe to Every: https://every.to/subscribe

Follow him on X: https://twitter.com/danshipper
Pitch is the AI presentation platform that helps professionals collaborate on, create, and deliver winning slide decks — all while staying on brand: https://pitch.com/use-cases/ai-presentation-maker/?utm_medium=paid-influencer&utm_campaign=every 

Timestamps:

00:00:00 - Start

00:01:54 - Introduction

00:02:03 - How Siskind is running an experiment between her unborn child and an LLM

00:07:34 - A demo of Siskind’s FetusGPT

00:15:16 - Siskind’s pick for the funniest LLM

00:17:12 - How Siskind uses AI in her comedy writing

00:24:41 - Dan and Siskind use ChatGPT to write a joke together live on the show

00:37:21 - Why AI is useful even when you don’t use its output directly

00:44:15 - How Siskind used a ChatGPT project to biohack her energy levels

00:57:09 - A question we fundamentally couldn’t have asked in pre-ChatGPT times

01:05:29 - How ChatGPT is a source of emotional support for Siskind in pregnancy

Links to resources mentioned in the episode:

Sarah Rose Siskind: https://sarahrosesiskind.com/

Siskind’s agency HelloSciCom: https://www.hellosci.com/

Siskind’s book recommendations: I Forced a Bot to Write This Book, The Let Them Theory, Artificial Intelligence: An Illustrated History