Core Memory

On January 5th, famed AI researcher Jerry Tworek stunned world+dog by announcing his departure from OpenAI. A few days later, he hopped over to the Core Memory podcast studio for his not-so-formal exit interview.
Tworek joined OpenAI in 2019 when the research lab was a research lab and had about thirty employees. He went on to work on many of OpenAI’s most consequential products, including the company’s reasoning technology, which ushered in a new era for the entire AI field. (Yes, Tworek worked on Q* before it was Strawberry before it was o1.)
Both Kylie and I have been longtime Tworek fans. He’s smart, funny and never really sought the limelight despite his massive contributions.
In the episode, Tworek reveals that he found it hard to keep doing high-risk, pioneering work at OpenAI as the company shifted toward what Tworek describes as more conservative ways. He, in fact, thinks the large AI companies have become conservative as a whole and that there might be bigger, better ideas to be found elsewhere.


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

Biohacking has gone through a lot of different phases. Implanting an NFC chip in your hand is old school and having a blood boy is passé. Among Silicon Valley’s 20-somethings, all the cool kids have a peptide stack.
Jasmine Sun joins us this week to chat about all things peptides. She was previously a product manager at Substack, but now she writes about San Francisco culture on her own Substack. Jasmine recently published a deep dive in The New York Times about the trendy injectable and deets on the Chinese peptide rave (which you first read about from our new writer, Kylie Robison, last month).
If you want to be like Wolverine, don’t do drugs. Subscribe to our newsletter and podcast instead. Our words are made of adamantium.

Do you feel old yet? We do.
In this episode, we get into all the important bits: What are peptides, why are they Chinese, and how is RFK Jr. involved? This is not medical advice, but if you do inject some peptides after this episode, tag us.
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.
Our show is sponsored by Brex. It builds finance tech that makes expensing and accounting for things like peptides super easy, if your company is cool with such things.
Like thousands of ambitious, innovative companies, we run on Brex so we can spend smarter and move faster. And you can too. Learn more at www.brex.com/corememory
The podcast is also made possible by E1 Ventures, which backs the most ambitious founders (probably some peptide users) 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 a guest host and some breaking news for this episode.
Eryney Marrogi, the scientist and soon-to-be doctor who writes for us now and again, has taken over the pod studio to interview Richard Fuisz. Earlier today, Marrogi broke a story on Fuisz’s company Nonfiction Labs, which has developed technology that could make it possible to use magnets to better control how cancer therapies are doled out in the body.
The two big brains get into Nonfiction’s technology and into Fuisz’s rather prolific work at the cutting-edge of the biotech field. The conversation goes into how biotech actually gets built, competition with China and Fuisz’s family legacy of invention (his grandfather was the prolific inventor featured in “Bad Blood”).
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.
Our show is sponsored by Brex, the intelligent finance platform. Like thousands of ambitious, innovative companies, we run on Brex so we can spend smarter and move faster. And you can too. Learn more at www.brex.com/corememory
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 been talking about computer-aided drug discovery for well more than a decade. It used to be the case that start-ups pitched their ability to use “machine learning” to hunt for new, promising therapies. Now we call machine learning “artificial intelligence” and have a new class of start-ups claiming big science breakthroughs.
One of these new wave start-ups is Chai Discovery and its founders Josh Meier and Jack Dent join the podcast this week. (The Core Memory podcast is available on all major platforms and on our YouTube pod channel.) The company was founded in 2024 and is backed by OpenAI, Menlo Ventures+Anthropic, Thrive and others. (Chai is already a unicorn.) It published a number of notable accomplishments this last year, including using its own AI model to churn out promising antibody designs at an unprecedented clip.
The first couple iterations of machine learning-aided drug discovery companies came and went without tremendous success. Chai and Nabla Bio are two of the buzziest members of this new era of AI companies. Their models really do seem to be harnessing the advances in AI to hit on potential drug targets and designs in rather profound ways. Bio-tech, in fact, seems like the place where AI may make the most stunning scientific advances first.
In this episode, we get into Chai’s intellectual roots as a research project within Facebook/Meta and how the company has gone after building its models. We also try to provide a realistic picture of the current state of AI drug discovery.
The implications of the work done by Chai, Nabla and others are far reaching. If we’re able to come up with new drug designs at this accelerated rate, we will need major changes around how drugs are tested and put through trials. The current drug testing and FDA approval system is simply not set up to move as quickly as bio-tech appears to be going.
This will be our last episode for the year, and we’re taking a tiny break between posting the next one as the Core Memory crew has a little time off. Thank you so, so much to all of you who have listened to the show in our first year. We hope you’ve enjoyed it and learned some things along the way.
Our show is sponsored by Brex, the intelligent finance platform. Like thousands of ambitious, innovative companies, we run on Brex so we can spend smarter and move faster. And you can too. Learn more at www.brex.com/corememory
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

Well, here we are. It’s brain uploading time.
As we’ve just reported, famed neuroscientist Sebastian Seung has created a new start-up called Memazing. The company has set out to build digital brains in software that are based upon the maps of animal brains. Memazing is, in effect, seeking to reverse engineer how animal brains work and to use this information to bring to life a new form of computerized intelligence.
This work could lead to, say, more energy efficient AI systems that are modeled on real brains. It could help with aligning AI systems with human intelligence. And it could be a major step toward creating emulations of full human brains and perhaps, one day, making minds uploadable.
We get into all of this with Seung on this week’s podcast. We also explore his decades of neuroscience work dedicated to building connectomes, or ultra-detailed schematics of animal brains and all their neurons and synapses.
Seung is brilliant and fascinating. Listen and/or watch for yourself.
The Core Memory podcast is available on all major platforms and our YouTube channel.
Our show is sponsored by Brex, the intelligent finance platform. Like thousands of ambitious, innovative companies, we run on Brex so we can spend smarter and move faster. And you can too. Learn more at www.brex.com/corememory
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