Core Memory

Pedro Franceschi taught himself to code when he was eight years old. At 12, he began receiving legal notices from Apple, asking him to stop hacking iPhones. By 14, he was making hundreds of thousands of dollars a year selling software and had his mom accompanying him on job interviews in his home city of Rio de Janeiro. Even among coding and hacking prodigies, Franceschi stands out.
Today, Franceschi is the co-founder and CEO of Brex, a financial technology company that was just acquired by Capital One for $5.15 billion. Franceschi is all of 29 years old now, so he’s done alright.
Brex led a new wave of companies that brought more modern financial tools first to start-ups and then to businesses of all sizes. Over the years, it’s had some ups and downs, and Franceschi has been remarkably open about Brex’s stumbles, his mental health struggles and about the areas where he thinks Brex got things very right.
Franceschi remains a hacker at heart and has been experimenting away with AI agents. He, in fact, says he’s running Brex – and his life – with a team of AI agents that read his e-mails and Slack messages, perform job recruiting tasks and schedule his day-to-day activities.
We get into all of this on the episode, charting Franceschi’s rise from hacking phenom to running a multi-billion-dollar company and discussing where he thinks AI and money are heading.
Do we have journalistic conflicts with this episode? Yes, we do. Brex has been the top sponsor of our podcast and video series. You can learn more about the depths of our relationship and what Brex can do for your business right here.
The podcast is also made possible by E1 Ventures, which backs the most ambitious founders and start-ups.



This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.corememory.com/subscribe

Baiju Bhatt is trying to pull an Elon Musk.
About 25 years ago, Musk sold his finance tech company PayPal and left dot-com life to get into rockets with the founding of SpaceX. Hardly anyone considered this a rational choice on Musk’s part. Space, after all, was where rich people went to blow their fortunes and fail.
For his part, Bhatt co-founded the investing service Robinhood in 2013 and has now decided to get into the space business as well via a start-up called Aetherflux. The company aims to build a network of solar panel-packed satellites that suck up sunshine and then beam it down to Earth via infrared lasers. Yes. Actual space lasers. What could go wrong?
The lasers would feed antennas and ground stations on Earth with energy. In theory, you could then direct power just about anywhere without needing to build a ton of infrastructure on the ground. Army convoys, data centers, etc. could just have electricity sent to them in remote areas.
Bhatt explains all of this in the episode and gets deep into his personal story. He also recounts starting and running Robinhood through its ups and downs, including being both beloved and despised.
Will the space lasers work? I dunno. It’s a lot. But we are fully in the era of trying new, bold ideas in Low Earth Orbit, and, well, I wrote a book predicting this very thing, and so am very much here for it.
The Core Memory podcast is on all major platforms and on our YouTube channel over here. If you enjoy the show, please leave a review and tell your friends.
This podcast is sponsored by Brex, the intelligent finance platform built to help companies spend smarter and move faster.
We run on Brex and so should you. Learn more about Brex right here.
The podcast is also made possible by E1 Ventures, which backs the most ambitious founders and start-ups.



This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.corememory.com/subscribe

We have tracked down the man and dog of the hour.
Paul Conyngham and his dog Rosie gained worldwide attention over the past week for breaking new medical ground. Using a variety of artificial intelligence tools, Conyngham – and some doctors and scientists in Australia – managed to create a personalized (petalized?) cancer treatment for Rosie that appears to be working.
The story resonated with the public for a couple of big reasons. First off, Conyngham has no real science or biology background. He’s a longtime AI researcher who used things like ChatGPT, Gemini and Grok to give him a plan for how to attack Rosie’s untreatable cancer and then how to craft and shape a unique mRNA shot for his pup. This exercise demonstrated the powers of AI technology to aid all of us with extra knowledge and skills and just how far bio-tech has come in terms of new cancer therapies.
Most people have had their hearts warmed by the tale of Paul and Rosie. Dude’s dog is dying. Dude goes to great lengths to try and solve the problem. Dude and his dog seem to mark a major moment for AI and medicine.
Some other people on the internet, however, are less excited by the story. They argue that the AI tools did very little here and that the science isn’t terribly conclusive or ground-breaking. Companies like Moderna and BioNTech already have personalized cancer vaccine data in trials, and it looks good. Who cares if we did the same thing for a dog? Rosie has also been treated with chemotherapy drugs, so we don’t even know if the mRNA technology is really the thing shrinking her tumors. And so on.
You can find some of the major criticisms here and here.
Some of the pushback may be valid, although Conyngham isn’t having it – as you’ll hear in the episode. It also sort of misses the point of this story.
After talking to Conyngham, it’s clear enough to me that he used AI in some profound ways here and that what was done with Rosie is symbolic of a huge shift in medicine. Regulators better get ready because the tools now exist for people to do rather daring experimentation on their pets and themselves. People in dire circumstances and with some means are going to be pushing the boundaries of what’s possible on a regular basis.
Paul and Rosie hit a nerve because their journey bundled up some massive technological and societal shifts into a tidy narrative.
Anyway, come listen to Paul and have a peek at Rosie.
The Core Memory podcast is on all major platforms and on our YouTube channel over here. If you enjoy the show, please leave a review and tell your friends.
This podcast is sponsored by Brex, the intelligent finance platform built to help companies spend smarter and move faster.
We run on Brex and so should you. Learn more about Brex right here.
The podcast is also made possible by E1 Ventures, which backs the most ambitious founders and start-ups.



This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.corememory.com/subscribe

The mainstream media says almost nothing about induced pluripotent stem cells (iPSCs). So, you’re lucky that we’re here to help.
These cells with a clunky name hold the promise of being able to reverse the aging process across our bodies. Put rather bluntly, your old, wine-soaked liver could become like your twenty-something, Jell-O-shot-soaked liver. Your aging neurons could fire like they once did. And your tired heart could be fresh and loving again.
Billions of dollars have been funneled toward trying to figure out how to push iPSCs into our organs safely and effectively. We have not cracked the code yet, but there are signs that scientists are getting closer.
Nabiha Saklayen, the co-founder and CEO of Cellino Bio, is an iPSC whiz and joined the podcast this week to bring us all up to speed on the technology. She covers how iPSCs work, their history and the state of iPSC treatments around the world.
Her company is trying to take iPSCs, which have largely been made by hand, and mass produce them to accelerate experimentation and hopefully therapies and to reduce costs around this fascinating technology.
The Core Memory podcast is on all major platforms and on our YouTube channel over here. If you enjoy the show, please leave a review and tell your friends.
This podcast is sponsored by Brex, the intelligent finance platform built to help companies spend smarter and move faster.
We run on Brex and so should you. Learn more about Brex right here.
The podcast is also made possible by E1 Ventures, which backs the most ambitious founders and start-ups.


This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.corememory.com/subscribe

Will Wilson paints a bleak picture for where we’re heading with code written by AIs.
He thinks the world will fill with poorly written code that no one understands and that software bugs will proliferate through critical systems. Your airplane that has gotten safer and safer with each passing decade will be running on code that no one has really checked all that well. Which would be bad.
What’s more, Wilson fears that humans will lose their software writing skills over time as AI takes on more and more tasks. We’ll become dumber as a whole. Which would also be bad.
Wilson is a mathematician turned start-up founder who built the company Antithesis in a bid to modernize software testing techniques and help humans write better code.
In this episode, we get into his life story, his fears around AI software and what he thinks we should do to make massive improvements to the code that underlies everything.
The Core Memory podcast is on all major platforms and on our YouTube channel over here. If you enjoy the show, please leave a review and tell your friends.
This podcast is sponsored by Brex, the intelligent finance platform built to help companies spend smarter and move faster.
We run on Brex and so should you. Learn more about Brex right here.
The podcast is also made possible by E1 Ventures, which backs the most ambitious founders and start-ups.



This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.corememory.com/subscribe