Core Memory

Andrew Hsu has been something of a legend for most of his life.

In 2007, The Seattle Times published a story documenting Hsu’s graduation from the University of Washington. He was only 16 and had just picked up three degrees in neurobiology, biochemistry and chemistry.

But stories of Hsu’s academic feats had already been circulating for years. He’d won science contests, written an award-winning autobiography and started a foundation to help children in need as an adolescent. Hsu’s family hails from Taiwan, and the young man often found himself being interviewed on TV and touring the country to tell his story.

After graduating from college, Hsu pursued a PhD at Stanford before dropping out and using some Thiel Fellowship money to start an ed-tech company called Airy Labs.

That company struggled, but Hsu’s latest venture – Speak – has been booming. It’s an AI-powered language tutor that enjoyed immense success first in South Korea and then beyond. It’s been valued at more than $1 billion after a $78 million funding round closed near the end of last year.

In this episode, Hsu graciously tolerates my child prodigy questions and then gets into how he hit on AI and language before it was cool and how people can learn better.

This episode was made possible by E1 Ventures, backers of bold people and bold ideas.

This week on the Core Memory podcast – we fix American science and advance civilization.

We were joined by Anastasia Gamick and Adam Marbleston from Convergent Research. They’ve spent the last few years pioneering a new model of science funding centered on FROs or Focused Research Organizations. And FROs take a little bit of explaining.

Convergent Research has backing from Eric Schmidt, James Fickel (a fantastic patron of science) and others and tries to fund small groups of people chasing very big ideas. In essence, Convergent wants to support things that help open up new fields of science and technology, and it funds folks whose ideas might be too expensive for a university lab and/or not obviously commercial enough for typical venture capital. Gamick and Marblestone argue that the FRO model fills a crucial gap in US science funding.

Convergent tends to put $30 million to $50 million into what look like quasi-start-ups and gives them five to six years to build their thing. To date, it has backed around a dozen efforts with a pretty heavy emphasis on the bio-tech and neuroscience fields.

Recently, Convergent also put out the Gap Map, which is a well-researched exploration of all the things that it thinks the world still needs to develop. Go ahead. Poke around.

In this episode, we break down FROs, science funding, the US vs. China, brains and much more.

Our show is made possible by the fine people at E1 Ventures. No cap table is complete without E1.

This week’s guest is Kian Sadeghi, the founder and CEO of Nucleus.

Sadeghi has everything you want in a controversialish bio-tech CEO. He’s a college dropout, a Thiel Fellow and a “wild child,” as one Nucleus investor told me. He’s also trying to uplevel the consumer DNA testing game by poring over entire human genomes with every test instead of just looking at snippets of DNA as companies like 23andMe and Ancestry have done for many years.

Nucleus charges about $500 for its mainstream health test aimed at adults. It promises to give you insights about a wide variety of health conditions, including your likely disposition toward things like mental health issues, cancers and rare genetic diseases. You can use the information to inform your lifestyle choices and to compare your DNA traits with those of your potential baby making partner to see if you’re a good baby making fit. (You can go here to see how Sadeghi uses this information on dates.)

The company also has a new, far more expensive service ($5,000) aimed at parents going through the IVF process to help them select embryos with certain traits. This type of service is quickly becoming all the rage, as we noted in our recent video on Orchid, which you should absolutely watch because it’s awesome. (Orchid contends that it does a much deeper dive on the embryo DNA than does Nucleus. I gave Sadeghi a chance to respond to some of this in the podcast.)

Sadeghi has been controversialish because he’s made big claims about Nucleus’s ability to discern things like someone’s IQ from DNA and because he’s been an aggressive marketer in a bio-tech field that tends more toward conservatism - lest one become the next jailed blood testing start-up CEO. He’s also been way more outspoken about the rather obvious direction we’re heading toward where people will be picking the desired traits of their future kids and where sex may well just become a purely recreational event as society moves toward IVF and artificial wombs for the majority of its new human production.

What’s clear enough is that the first wave of consumer genetic testing companies arrived many years ago when DNA tests were rarer and more expensive, and we’re now seeing them be usurped by a new crop of services that really take advantage of the massive decreases in sequencing costs. In short, we can test more of your DNA more cheaply than ever before, and we have much better data and software to analyze the DNA now.

Sadeghi and I get into all of this on the podcast.

The show was made possible by the fine people at E1 Ventures. No cap table is complete without E1, or at least that’s what I tell my kids.

Enjoy!

The news here is that Paul Eremenko has a new start-up called P-1 AI.

Eremenko is billing P-1 as one of the first stabs at building an AI engineer. The company’s “Archie” AI can help with day-to-day engineering tasks today, and, if all goes according to plan, will be designing buildings, planes and rockets in the future. We, of course, getting into what Archie can do today and what it might do in the years to come in the pod. Spoiler alert: Eremenko thinks we get MUCH better spaceships.

Most of our time, though, was spent discussing Eremenko’s rather incredible life and career.

Born in Ukraine, Eremenko came to the US at 11 and went on to get aeronautics degrees from MIT and Caltech and then – just to show off - a law degree from Georgetown. He’s worked at DARPA and Google and as CTO of both Airbus and United Technologies. He also tried to turn hydrogen into a mainstream fuel source for commercial planes at Universal Hydrogen, although that venture did not pan out.

And so, we got into Eremenko’s life, aerospace and where AI is possibly taking us a species.

Eremenko’s dog Li made a special guest appearance and picked me as his favorite podcast host by the end of the show.

As ever, we thank the wonderful people at E1 Ventures for their support with the podcast

Enjoy!

More than a decade ago, someone I respect told me to go meet these young, Irish brothers - Patrick and John Collison. The brothers had started a small company called Stripe, and my friend assured me they were primed to accomplish big things.

The Collisons were working on payments, and I had no interest in payments, so my attention waned a bit as they described how Stripe functioned and what it would one day do. What was very clear, though, was that the brothers were bright - as in exceptionally bright - and focused and determined. I interview start-up founders for a living, and there’s been a handful of times where I knew for certain that the people in front of me would succeed at whatever they chose to do. This was one of those times.

This is a long way of saying that I have the utmost respect for the Collisons and try to take particular note when they and/or Stripe make big bets. They tend to have a pretty accurate window into the future.

Last year, Stripe bought Bridge for $1.1 billion. Bridge was a two-year-old start-up that had started out doing some NFT nonsense and then pivoted almost right away into stablecoins.

Going off the premise that the Collisons must have spent $1 billion on a very, very, very young company for a reason, we asked Bridge CEO Zach Abrams to come on the podcast to explain what Bridge does, what the hell stablecoins are and where the future of money is heading.

Abrams, thankfully, did not disappoint.

The short of it is that Bridge has made it much easier for companies and governments to move money internationally. SpaceX, for example, relies on Bridge to collect and process payments for its Starlink internet service in far off lands. The same goes for people sending and receiving remittances, which happens to be a massive part of our global economy.

We discuss all this in the show and then get weird. Abrams talks about AIs using credit cards to accomplish tasks out in the world and a future where an AI might end up as the wealthiest being on the planet and what that could mean for us humans and the economy.

This podcast was made possible with support from the fine people at E1 Ventures.

Enjoy!

Palmer Luckey has come on the Core Memory podcast today to deliver some full-on shocking news. (And top tips on raising children as well.)

As you’ll hear on the show, Luckey’s company Anduril has partnered with Meta to create a product for the U.S. military dubbed “Eagle Eye.” At its core, this product is meant to become the sci-fi style military helmet that you see depicted in movies but that does not actually exist in real life. It will have displays that place all kinds of information in front of soldiers’ faces by tapping into virtual and augmented reality technology and data feeds that will be pumped into the device.

Microsoft once owned the contract to make this type of product for the U.S. Army but had been struggling terribly to deliver anything useful. Anduril took over the $22 billion project earlier this year and will now pair its defense and tech expertise with Meta’s headset and VR/AR expertise to try and give the Army what it desires and modernize the U.S. military in the process. We go into “Eagle Eye,” the technology behind it and how and where it will be made in gory detail on the podcast.

This is all shocking for a bunch of reasons, but the lead shocker is that Luckey has agreed to work with Meta and Zuck at all.

Some context.

In 2014, Facebook acquired Luckey’s Oculus VR for $2 billion. In 2016, Facebook then fired Luckey more or less for being a Republican in public.

In the runup to the 2016 election, Luckey gave $9,000 to a group that put up a billboard depicting Hillary Clinton’s face – with an extra-large forehead – and the words “Too Big To Jail” underneath the face. This was during the Clinton e-mail controversy and came at a time when much of Silicon Valley had gone apoplectic about the idea of Donald Trump possibly becoming president.

Once the mainstream press figured out that Luckey had paid for the billboard, it went full hysteria mode and portrayed Luckey as some kind of hate-filled, fascist meme lord set on destroying the moral fabric of, er, politics and possibly the American Way of Life. Facebook decided it could not stomach the PR hit and pushed Luckey out of the company. Lawsuits and much vitriol between the two parties followed.

For Luckey, the whole incident was well beyond personal. Oculus and VR tech had been his life. Facebook stripped him of his true love, and the press and others turned Luckey into a pariah. The saga is captured wonderfully in Blake Harris’s The History of the Future where hindsight allows us to see how something relatively trivial – the billboard – morphed into an absurdist drama acted out by reporters and Facebook executives.

Luckey also made his feelings on the incident very clear in this historic performance in which he eviscerated professional remora Jason Calacanis.

But, you know, times change. Zuck and others in Silicon Valley have discovered their inner patriots and want to work on defense tech now. Luckey being buddies with Trump and Republicans is so okay that he can appear in a Meta press release. Public apologies have been made. And now, perhaps, soldiers can have their fancy helmets.

We spent two hours chatting with Luckey, and the Anduril/Meta deal is only a fraction of the discussion that also gets into Anduril’s manufacturing expansion, China (of course), AI and a host of other topics.

For more Core Memory pods, head here.

The episode was made possible by support from E1 Ventures.

Enjoy!

The facts are these: Peter Beck is the founder and CEO of Rocket Lab, and Rocket Lab is an absolute beast in the aerospace world. It has launched more than 60 times from spaceports in New Zealand and the US and is in the midst of creating a bigger, more powerful rocket to help it earn more business and compete more directly against SpaceX and others.

Beck and Rocket Lab also happen to be near and dear to my heart. I wrote a book about them and made a movie about them.

Beck has an incredible life story. He’s a self-taught rocket engineer who built a commercial space giant in New Zealand. None of this should really be possible. You’re supposed to have a PhD in aerospace and/or billions of dollars to be successful in the rocket game, and you’re supposed to build rockets in places that have some experience building rockets. Nonetheless, here we are. Rocket Lab sits alongside SpaceX as the obvious winners to date in the commercial rocket and commercial space games.

We’re thrilled that Beck gave us some time as he crunches away on preparing the Neutron rocket for its first launch.

This episode was made possible by the fine people at E1 Ventures. We thank them for their support.

Enjoy!

For more podcasts and the finest in sci-tech reporting, subscribe here.

Back when I first began covering technology in the early 2000s, my favorite thing to write about was open source software. I was young and idealistic, and the hardcore free software and open source zealots spoke to me. Code was meant to be by the people, for the people. Richard Stallman and Eben Moglen seemed like heroes. Microsoft and its proprietary code-fueled desktop monopoly seemed clearly evil. I enjoyed the energy and vitriol on both sides during the peak of these debates. Linux 4EVA!!, I would write on my all-too hard to use Debian machine.

The software religious wars kind of, sort of linger on but in much more muted forms than I remember.

Sad.

We, however, might have a proper tech revolutionary for this contemporary era in the form of Guillaume Verdon.

Some of you will know Verdon better as Beff Jezos, the X personality who built the e/acc or effective accelerationist movement into a countervailing force against the doomy, gloomy Effective Altruism movement, which, rather comically, managed to undermine itself without Verdon’s help by taking gobs of money from the anxiety-ridden villain SBF and wrapping itself in an uninspiring blanket of malaise.

Anyway, Verdon became and remains a thing both with e/acc and with his start-up Extropic, and the two are very much interlinked.

Extropic has shown early success with “thermodynamic computing.” It’s a form of computing that Verdon says harnesses the underlying properties of nature and probabilities in far better ways than traditional computers and in more practical ways (possibly) than quantum computers. Verdon used to work on quantum computers under Sergey Brin at Google, so he might even know what he’s talking about.

The revolutionary part of all this is that Verdon thinks Extropic will make cheaper, more energy efficient AI processing systems than the likes of Nvidia, OpenAI and Google. His AI computers will not require trillion dollar investments in data centers but rather will be affordable to the masses (possibly).

It’s very early days for Extropic, so much of this decentralized AI fervor is fueled by prognostication and hope.

Obviously, we get into all of this on the pod.

The show starts Extropic heavy and then veers into e/acc and decentralized AI territory. So, if thermodynamic computing is not your thing, go ahead and skip to the more philosophical stuff where you’ll find that Verdon is fun to listen to and something of an engineer philosopher.

This podcast was sponsored by e1 Ventures – the smartest and most noble podcast sponsor in Silicon Valley and all points beyond.

A few weeks ago, we went to visit Impulse Labs, which has produced the world’s most advanced and most ambitious cooktop.

The start-up, based in San Francisco, has a $6,000 induction stove that can boil water in seconds and that lets you alter the temperature on your cooktop degree by degree. It’s an engineering marvel, as you’ll see in the video.

The product is also an undercover agent of sorts.

Impulse has packed its cooktop with a number of lithium-ion batteries. They’re used to make the stove work well but also to store extra energy.

Impulse plans to make a wide range of appliances, network them together and create a type of battery backup system for the home.

Enjoy the show!

First. The news.

The bio-tech player New Limit has raised $130 million from just about the fanciest assembly of smart, rich people imaginable. Nat Friedman and Daniel Gross – via NFDG - Kleiner Perkins, Khosla Ventures and Human Capital are there in their corporate forms and Patrick and John Collison, Josh Kushner, Joe Lonsdale and Fred Ehrsam are there as individuals.

Over the past four years, New Limit has been trying to identify the right combinations of transcription factors – a certain class of proteins – that can rewind cells and take them back to a younger state. Their work piggybacks on the Nobel Prize winning work of Shinya Yamanaka, and it’s among the most exciting technology in the entire bio-tech field – at least for me. As you’ll hear in this interview, they’ve made massive progress over the past 18 months or so.

We’ve talked about transcription factors and related technology with Joe Betts LaCroix from Retro (podcast under the Joe link, and full video episode on Retro here) and with Brian Armstrong, who co-founded New Limit.

In this episode, however, we hung out with Jacob Kimmel, another New Limit co-founder, for a real deep dive on transcription factors and New Limit’s approach to taming them. Kimmel is as clear and eloquent as it gets on explaining this technology.

This pod might feel different than the usual pods. It comes from a sit down interview we did with Kimmel for an upcoming Core Memory video on New Limit. Still, it’s glorious.

The Core Memory podcast is sponsored by the wonderful people at E1 Ventures. Their money and hearts are pure.