New, More Precise Cancer Therapies Could Soon Be Here - EP 51 Richard Fuisz

Hello friends, this is Ashlee Vance. I am coming to you with a little bit of a different version of the Core Memory Podcast. This week we have Aaron Morogi, who writes for us sometimes. He's a scientist, he's becoming a doctor, and, um, he's going to be the host of the show. He's talking to Richard Fuse, who is also a scientist, and he, he just started a company called Nonfiction Labs. They have a new type of potential cancer therapy they're working on. Erné's gonna explain it all to you. Just wanted to give you a teeny little bit of context as to why I'm not hosting and, and to let you know this is a special extra, extra, extra sciency version of Core Memory. And if you're listening to this, keep listening and then, and then go to Core Memory and check out Erné's story on, on Richard. He's just, he's Broken the Story on Nonfiction Labs. So we're, we're, it's doing double duty for us. Um, as ever, we hope you enjoy. Please like, subscribe, leave a review, all that kind of stuff. And, uh, thank you for listening.

All right, well, welcome to Core Memory. Um, I'm Aaron A. Uh, today's guest is Richard Fuze. Um, thank you very much for coming here, Richard.

Yeah, thanks for being here.

Um, I'm going to do a little bit of an intro about you for folks who don't know, um, and then you just correct me anywhere you feel like is, you know, correctable. Um, Richard is a little bit of a legend, I feel like, in biology in, in the Silicon Valley area. Uh, you have done quite a number of things. So you graduated a while ago from Stanford, right? And you're kind of in the bioengineering space, but you've done a few, I think, like pretty big ticket projects. So you're involved in Arcadia, uh, you're also involved in something that's pretty popular nowadays, kind of hot, uh, is, uh, Future House, which is like an AI for science project. And then most recently, and I think what I'm most excited to talk to you about, is your, your most recent venture into to, uh, kind of like a, a bioengineering, uh, startup called Nonfiction Labs. Um, yeah, maybe we can go from there. Yeah. I don't know. Does that— you feel like I kind of just listed 3 things, but that, you know, I summarized you poorly in 30 seconds?

I think that's decent start.

Okay.

That's pretty good.

Okay. Um, yeah.

So I mean, maybe it would be interesting to talk about, uh, Well, I certainly can say a little bit about what we're working on right now. But I think it's a couple of directions.

Yeah. You know, I will have plenty of time to talk about what you're working on now. I definitely have plenty of things I want to ask you. I'm actually, I'm mostly just curious, you know, you went to Stanford, you were there during like the peak of like the, you know, everyone wants to have like a startup, not in biology, obviously people were doing startups in everything except for biology. How does someone go from, go to Stanford and then not end up making like B2B SaaS.

Yeah. Uh, I think that was interesting.

Okay.

That's a place we can work.

I think that's a good start. Yeah.

I, so, well, for first off, I had a bunch of friends who did go into B2B SaaS stuff. Uh, crypto was, I think, increasingly becoming a big thing.

Okay.

This is like, yeah, mid-2010s.

Okay.

Uh, I had, at the time I had a good friend, Noor. Who went on to start Orchid. And, but at the time, this is in undergrad, she was already sort of like lining that up and she'd done her Thiel Fellowship before. And so I think I sort of was watching some of, some of that build out. I think it was also, I guess really what I would say is I think it was a, like that mid-2010s was a part in time when you could start weird biotech stuff. And I had been working with some friends of mine. We found this way to get some money at Stanford to start this lab where we basically, we had access to like, we'd go into the lab like after hours, after 5 o'clock, we'd like take over the space and we'd work on various projects. And it was treated as like a club. That's how we got it funded. And then there I started working on doing DNA synthesis stuff because the idea was we would like, well, the grand idea was sort of like Church lab, like panspermia. So the idea was like, let's seed the galaxy with life and, you know, however interesting what is on these planets is currently, surely it would be more interesting if they had something living on them. And if we could make a, if we had a DNA synthesizer, and if we had the ability to transfect cells to get that DNA into cells, we could do remote genetic engineering. And in the 100 years it would take to get the, like, hardware on the planet, uh, you know, we'll be better at engineering then. And so that's what we'll do. So this was sort of like the idea we were working on, like, out of undergrad. Maybe this is the answer to your question of like, what I was doing wasn't software.

Um, it's just, it was like adjacent-ish to tech. It was like tech adjacent, but it was really bio-focused, which is cool. I mean, that's, you know, ultimately I think that's, did you feel like back, back then there was a lot of interest in biology at Stanford? Because I know this is like 2015-ish or so.

Yeah. So, I mean, I spent, I spent some time around like Gerandi's lab around then. And so Gerandi was sort of like, uh, the West Coast synbio guy. He'd come out of the MIT bioengineering department and like that whole ecosystem. Um, and so it was also probably a lot through that that I got into DNA synthesis was, was Drew Endy's lab.

Okay.

Um, and so, yeah, so then that got us into doing enzymatic DNA synthesis. The idea was like, if you could make a DNA printer, it would be like safe. Uh, then you could, you could sort of use that as part of that device and then kind of built out.

All right. What do we do at Core Memory? We cover innovative, fast-moving, forward-thinking companies, which is why Core Memory is sponsored by Brex because Brex is the intelligent finance platform for many of these companies. 30,000 companies from startups to the world's largest corporations rely on Brex's technology for their finances. They've got smart corporate cards, high-yield business banking, and expense automation tools that are fantastic. I hate doing my expenses and Brex's AIs and software run right through those expenses, figure out where we're spending money, and take care of so much stuff for you so you don't have to waste your time on it yourself. Go to brex.com/corememory to learn more and just, you know, get with the program. Let's get going. Let's get out of this archaic finance software and move toward the future. CoreMemory and Brex.

So you did that. It seems like very much like an academic project. My guess— maybe I'm assuming— my guess is you understood. You're a smart guy. You knew that was not going to be a viable business. My assumption. Hopefully you're going to say you're going to agree with me. Otherwise, I have to adjust my perception of you. But then, you know, you kind of— you do that very academic and it's in its focus. You're still like at Stanford. Biotech is very much a degree-driven or sort of like scientifically driven enterprise. You decide not to go get a PhD. Kind of what What was that decision point for you? You do something else actually entirely, right?

Yeah. The first bit was feeling like I wanted to do something that like actually mattered.

Yeah.

Um, feeling like, uh, I think in retrospect, so I did end up starting this DNA synthesis company off of that. Uh, DNA synthesis as a market is like one of the worst markets like you could pick. Uh, it's a very small, like the actual number of the amount of DNA sold every year It's probably generously in the hundreds of millions, you know, ambitiously in a billion a year. Uh, it's a commodity product that's sold at like negative margins most of the time. Okay. Uh, and so, so as a business, I think it was, it was, it was a tough place to go into. It was also one of these places where a lot of other like smart people had tried and failed.

Okay.

And people said, don't do it. Um, we had kind of this argument at the time that like, that everyone was thinking about it wrong and like you could have higher margins. And the pitch for that was basically, if you could get the DNA in a day, then you could do a very different type of experimentation. And so, you know, iteration times being super important in biology, like that was kind of our frame of like, wouldn't you want to do an experiment every couple of days instead of waiting 2 weeks and batching your synthesis?

And was this company mostly working on like pools or was your pitch like, oh, we wanted to do like IDT So like IDT is like the big DNA synthesis company. At least most people in biology are probably very familiar with IDT's products. At least for me, I use IDT a lot for like individual G-blocks or primers or whatever. Were you guys trying to go for that? Or are you like anticipating the pool market that like Twist and Agilent now kind of occupy?

We were more, we were more lucky at the IDT, like G-block kind of thing. Cause the idea was like. Make your starting product and then vary it from there.

Okay.

A lot of the time when people are making pools nowadays, that's like, well, people do it for different reasons, but.

Right.

Um, our philosophy was if you wanted to have a, um, yeah, just getting that first primer would be what mattered.

Okay.

And then you could either mutagenize or you could fuse it to different stuff.

Okay.

Um, maybe this is getting too in the weeds. I don't know.

Well, yeah, you know, I, I, I was mostly just curious.

Yeah.

Yeah. Okay.

Yeah. So, so we basically started Uh, uh, yeah, we started basically doing, doing that. Uh, we were building like out of this weird warehouse.

Okay.

Uh, in the Mission. Uh, it was just like burned out. It was an old punk rock venue. Okay.

Yeah. Okay.

Uh, good.

Those are being used for something profitable.

Yeah. Yeah. Similar.

Yeah.

Yeah.

Yeah.

So we were, we were working out of that for a while and then I think we, Uh, I got a team of folks to work with me, which was great. Cause to your point, like credentialism, being 22 was like right on the margins of like where you could raise money, where you could build a team, but it was like, all of that was hard and in a bad market. Um, so we, we got a great team together and we got some money and we built it and we're doing that for a couple of years until COVID. And then we were kind of like, yeah, like, this is, this is, this is pretty tough.

Okay.

Um, so, uh, it became hard to hire nucleic acid chemists.

Okay.

Like, in the mRNA world. So, uh, so yeah, so then off of that, I, uh, uh, that stuff was sort of like coming to a close. Uh, and then in the middle of that, you had this like new science era in like 2020 or so. Uh, and then I kind of got, got swept into that.

Yeah.

Yeah.

Yeah. I guess where you're going with this is your Arcadia is kind of what you did next.

Yeah.

Yeah. How did, how did that, how did that start? Maybe if you can talk a little bit about what the goal was there and then a little bit about how you got started into that.

Yeah. So the, so where I got involved in that project was when, um, so Jed McCaleb, previous Core Memory, um, podcast guest. He, you know, had made some money in crypto. He wanted to sort of, you know, take some of his winnings from that and like put it towards something which he thought was sort of valuable for the world. And so he said, okay, he knew he wanted to do some sciencey stuff. My friend Noor was also a mutual friend. And so basically said like, hey, you should go talk to Richard. He has some crazy science-y ideas and like you should talk to Jed. And so we started talking around that, um, when sort of Arcadia was coming up and, um, and obviously Sime and Prachi and a couple other folks there who were involved early on. And so kind of what emerged from that was, uh, yeah, uh, was, was, was Arcadia. So, um, there was an initial big focus on sort of open publishing.

So, um, now actually, sorry, maybe, maybe you can sort of just more broadly, like, what is, what is Arcadia like about?

Oh yeah. Well, I mean, so it's sort of changed over time. Yeah.

Maybe back then, what were they about?

And back, so back then the, uh, so, so at the beginning with Arcadia, it was basically focused on non-model organism development. So, uh, usually we work with mice, we work with E. coli, yeast, like a couple species get almost all the research funding. But if you look at like the actual gains from science, it frequently was coming from organisms that were not those. So, you know, GFP, like, uh, the standard protein people use as like proof of concept for most stuff comes from jellyfish. Um, we had seen some interesting results, um, with sort of an early project around ticks, um, and like tick saliva because ticks like when they bite you, they inject you with a bunch of like anticoagulants and, you know, other novel stuff. And so, uh, the idea was if we could get good at working with weird organisms, that might be sort of an interesting and sustainable, like, edge and a new place where you could have something, something new.

Um, and this was always non-profit, it's just sort of like academic, just not at a university. It was kind of like even back then that was the pitch. Yeah.

So it was structured as a for-profit.

Okay. Uh, actually I didn't know that, but okay.

It was, uh, it was structured as a for-profit, but like a for-profit that sort of had some flexibility for exactly.

Okay.

Um, one of the challenges with nonprofits is that you get into some restrictions around like how you can use funds for, uh, like for example, the IRS will like be concerned if you as a nonprofit are paying way above market for somebody.

Okay.

It's like, okay.

Yeah. It's like, okay. I see.

It's like, is that because that person is really good or is it because that person is like your buddy? And, and so like, as a nonprofit, you have these sort of restrictions. Uh, they make sense if you have to do them. Okay. Um, okay. Uh, but if you don't have to have a nonprofit, it helps to not.

Okay.

Um, and, and, and that also meant we could issue equity and stuff like that to the, to the scientists, which, yeah. Um, which is a nice thing. So, uh, yeah. So that was, that was the Arcadia thing. So initially around like non-model organisms. And then you had a second component, which was basically around open publishing. So sort of free access to content. And then you had a final bit, which is really about like trying to elevate the, like the people at the bench, which is something that I thought is like super important. Um, a lot of nonprofit, like, or larger institutes that were getting started would like use this PI model, which ends up being really easy to hire for, but, is not as good if you're trying to build out a team. Right, right. Um, I don't know if people know what PI— Yeah, yeah, but yeah, go ahead. Yeah. I mean, basically the idea would be, so it's like a pyramid scheme form of hiring.

So you basically say, are you calling academia a pyramid scheme, Richard?

It's, it's, it's pyramid adjacent. It's like a shopping mall. It's like, it's like the modern—

Yeah. It's like an MLM.

Yeah, exactly. I'm gonna get this big shopping mall and I'm gonna like hire a bunch of sole proprietor, kind of like the, the professors, and then all the money will be in their name, and then all the work will be done, uh, by these other people in their name, but then like we'll strap the PI's name into the end of the paper.

That's the only one that gets referenced going forward, basically.

Yeah. Well, yeah.

Nobel Prize comes out of it. The PI.

Yeah. So you end up with sort of this, you end up with this like very favored relationship for the, for the PI. Um, and I think that works. Works in the academic system. There's reasons why it is that way. But the idea was like, why are we taking, and this was not just Arcadia. This was like, I think a commonality between a lot of the orgs at the time were asking really the same question, which was like, why are we taking the smartest people, the best PhD students and the best postdocs, the people who like did go to graduate school and like have like demonstrated—

I think you're a smart guy. I think you're a smart guy.

Well, I, people who've really demonstrated amazing work at the bench, like are actually doing really cool stuff. And we're saying, congratulations, like, thank you for being really good at that. Uh, we're now going to have you work on something which is unrelated, right? Like congrats. Now you get to be a recruiter and grant writer and a manager and dealing with the internal politics of universities, which is a non-zero portion of your time. Um, and, uh, our attitude was very much like, what if you had a place like where inventors could be like inventors and people could be like talented and get to keep doing the thing they're doing without the administrative overhead.

Yeah.

Um, and that's the same thing that we did over at Future House. I mean, that basically is the same. That was sort of the originating goal of that. Yeah.

Yeah. Yeah. I mean, in some ways it was very prescient, sort of, I think like 2025 has been the big year of like, you know, there's a search for alternative models for doing science, I guess, you know, it's really been focused on science, R&D really.

Yeah.

Um, it seems like that's kind of what you were doing for, you know, at least until, until recently, it was like a fairly big part of your life. And I do think you kind of anticipated, or at least, you know, folks who were back then interested in the space anticipated where we are today, which is, you know, people have sort of recognized, as you were, as you were maybe alluding to a little bit, the academic model has a lot of flaws in it. Obviously, there are things that work well, but then a lot of things don't work very well. And so you went from Arcadia working on some of those things, eventually becomes more of like a, uh, similar to like we're doing nonprofit, nonprofit science or like we're just kind of trying to make it easier to, to work on kind of, you know, weird problems that are not in, you know, not at a university, but then you go into Future House and actually that's how you and I met some time ago. Yeah.

Yes. Yes. Uh, so with Future House, we ended up doing It's a different situation. So, uh, uh, different philanthropist. So one thing which is true, I think, across all of these is these orgs are, um, sort of co-created like with the funder and the way that the funder likes to think about the problem. Um, which is different than like a normal startup where it's often much more like the founder's vision is the whole vision. Like in most nonprofit orgs like this, like the funder is one of the founders in a, like a very real way. Um, so Future House was funded by Eric Schmidt. Uh, Eric Schmidt, uh, has a lot of, like, he's a believer in, like, more process and system planning. Uh, and so we spent about a year with Future House basically, like, um, building out a more complete version of, like, what we wanted the plan to be. So we had this big, like, 150-page proposal. Um, and we also talked to a ton of scientists, uh, which is, I think, how we met.

Yeah.

Yeah. Um, Yeah. So I think we were introduced through Alexi Guzzi. Is that right?

Yeah, I think, I think that's, I think that sounds familiar. Yeah. Another person who is in fact working on a sort of new science model of, of doing things. Yeah.

Yeah.

Yeah.

Um, so yeah, so that kind of got, got us into that question. Future House was not focused on the non-model organism bit, but was very much focused initially on sort of doubling down on that, like, like Inventors Institute, like postdocs run the place perspective. So we just talked to like a ton of scientists and we're like, who are the people who have like a really nice, interesting set of projects and mostly focused on like tool development, which, uh, has a particular meaning in biology. Usually when people talk about tool development in bio, it's like, um, things that other people will, can repurpose and use in like 20 different ways. You're not making a drug. You're making like a way to make more drugs, like you're focused on the factory part. And so that was the focus. The other early focus of Futurast we also dabbled a little bit with was like human enhancement or like, you know, drugs for healthy people sort of perspective, which I think that bit has definitely become like a big deal in the last year.

Yeah, definitely. I mean, that's what all the, you know, I think the, you know, the embryo editing companies I think are targeting that genre. And then also obviously GLP-1s. Not, you know, as we're seeing, you know, it's not all diabetics who are taking these drugs. It's a lot of people who have a BMI of 28 who want to be at a BMI of, you know, 22 instead. So yeah, I think, yeah, again, another thing I think you and others had anticipated sort of like where things will go based off of trends.

Yeah, I think it was an interesting time to be thinking about that, that kind of stuff. Mm-hmm. Yeah. Yeah.

Obviously, as we know, you know, Future House today is working on a different kind of thing, working on AI for science or tools for AI and science. So that's, but that's not totally where you went. You kind of went on a different path and now you're working on something exciting to me anyway.

Yeah.

Yeah.

Yeah. Yeah. So my, so my, uh, my co-founder for the current thing, I met her at Arcadia and then wanted, or sort of in the early sort of formation of Arcadia, and she was very involved. Um, wanted to be more involved with the Future House stuff. So she's almost, you know, sort of, we tried to work together for like 3 other previous times. Um, uh, it's these two scientists, Andy York and Maria Ingramo. Um, and, uh, they were over at Calico at the time. Um, and they basically were working on a bunch of, a bunch of different interesting stuff sort of there. They were like officially part of the core lab at Calico. So Calico is like a large, like it's Google's life sciences arm. And so, um, they've been going for maybe a decade. They were, I think Calico, I think that's one of the first places to do longevity research. They were definitely early on that. Um, but Andy and Maria didn't usually do much.

Yeah. Well, it's definitely not based off of, you know, where I think you're going to go, but yeah.

Yeah. Yeah.

Yeah.

So, so, so, um, yeah, I mean, I can tell basically the story with that.

Um, all right, let's hear it from, from your mouth.

Yeah. Yeah. So, uh, so Maria, uh, if you, if you're very active on like Twitter science, you might've seen some of this stuff maybe a year ago or a year and change. Um, maybe 2, 2 years ago now.

I think it was like summer of '24. At least that's right. Yeah, I was at Caltech and I think I remember seeing it then.

Yeah. Yeah, exactly.

Yeah.

Um, so, uh, so she started taking fluorescent proteins. Um, well, actually, so the original, the original person to really credit with this, Adam Cohen over at Harvard, um, really brilliant professor there. Um, he had done some work in magnetogenetics, had some interest in it.

Yeah.

There had been this big flare up in the field and like 2015 and like people saying, hey, well, birds can see magnetic fields and we have these like, uh, cells that if you put them near a magnet, they'll sort of like, there's a big hunk of iron in the cells. And so they will like align to the magnetic field and then they can swim in that direction. And so, uh, that also means you can pull them with magnets. So there's all this excitement about it. Uh, there was this paper that sort of came that ended sort of slowing the field down where someone pointed out, actually, maybe some of this behavioral stuff is a little bit suspicious. Uh, some of the biophysics of this is like, you're like 100x off from what is physically possible. Like you can't actually torque tiny little bits of iron and get them to pull.

Okay.

And so, and so the field ended up sort of going into this dark age for a while. Um, but then Adam Cohen over at Harvard comes in, uh, and he basically says, uh, that there are some small molecule dyes that respond to magnets. And like, that's like an objective, obvious, like, you know, you can bring a magnet over to it, you can see it change like color and you're like, okay, like this is real. And so then we asked the question, there are these, you know, proteins, like naturally occurring proteins in biology that are fluorescent. And, you know, Maria asked, can I get that to modulate with a magnet as well?

Right.

So can I change the color? Um, and that kicked off 2 years of directed evolution, which she started doing, where she basically would mutate the protein, pick the one that responded the most to the magnet, and then, you know, remutate and do it again. And so, sort of this like accelerated evolutionary process. Um, and so she ended up getting these, these, these proteins, which like you can wave a magnet over them and you can like see like visibly without needing to do any statistics on it that like they're changing. Color.

So this was like, this was a GFP, right? Yeah.

Some of the early stuff was with GFP and then it ended up going to these, uh, couple different, turns out basically every domain she tried with it, she could get some effect, but it ended up with this, with a protein called Aslov2, which comes from oats.

Okay. Uh, okay. And, uh, the magnetic, uh, the magnetic organism, the oat.

Yeah.

Well, so that was the interesting thing. So normally all these proteins don't respond to magnets. That's right. And that's actually weird because usually like when people do bioengineering, they, you know, you're sort of stealing a little bit, right? So you're taking, you're taking something that already exists in biology, but in some different species and you're pulling it over and you're like, well, okay, this bacteria can sense light. It's using a channel rhodopsin to do that.

So I guess you're referring to like optogenetics there.

Exactly.

Yeah.

Yeah. Yeah. And I can pull that channelrhodopsin into my human neurons and I can hit them with light.

Yeah.

And now the same way the bacteria was sensing by changing its ions, the neuron sensing by changing its ion gradient.

Yeah.

And so I can get the neuron to fire by sort of stealing this thing.

Yeah.

With magnetogenetics, like there are not, um, these aren't, these proteins don't normally respond to magnets, right? Like at all.

Yeah. I guess you're trying to invent a mechanism or you're sort of trying to like, there's some, some more physics-based thing you're observing and you're trying to engineer around that as opposed to, as you were referring to, like take channelrhodopsins, you're like, okay, this exists in some organism, this is how they operate. Let's bring that into a human. I mean, a lot of biology, I guess, is done that way. Like that's how we got CRISPR-Cas9. It's like these bacteria have some immune system thing. Let's like use that for our purposes. Or, you know, reverse transcription from viruses. Let's co-opt that to human biology. So this is different. Like what you guys are doing is you're Observing some phenomenon and saying like, this could be used in some broader engineered system, really engineered, right? Like this is not just co-opting.

Yeah. So you're looking at like the physics of it and you're saying, okay, we look at these proteins, you have some general sense of like where the electrons are moving in these proteins and what are happening when we hit it with light. And you have a theory that if you, you know, hit it with a magnet as well, you can modulate that response to light.

Okay.

Um, I, the truth of how a lot of these things really work though is like, I mean, you know, like you don't really start from like what the theory is and then you work back.

That's what you're supposed to say. You're supposed to say that when you talk to the media.

Exactly. Yeah. Yeah. Usually, usually the, like the, the, the story is always supposed to be like, oh, well, of course.

Yeah.

You know, we did this from quantum physics principles.

Yes.

Yes. Of course.

You've thought it all through.

Yeah.

Yeah. But, but really the answer is like you have this emergent property that you find. When you do evolution and like, um, whatever. Now it's very helpful. So like now, uh, like Adam Cohen and this group went back and they did a paper where they said like, we think this is what the mechanism is explaining these things. And now once you have that stuff, it helps sort of understand like how to make it better.

Right.

Um, but yeah, these are just like weird things that biology can do if you try to make it. Yeah.

Yeah. So, okay. So you basically, so I have the timeline right in my head.

Yeah.

Basically you went, from Future House doing one type of sort of inventor model, then that's, you know, you, you got acquainted with Maria, your co-founder.

Yeah.

Um, and she was working on something which was like these magnetically activated proteins, GFP fluorescent stuff really, where you started out, she started out with some really minuscule observation of like, oh, this gets a little bit intensified or dimmed with the magnet. And then Let's poke on that for, oh, 2 years, it sounds like. Eventually getting something that was like actually visible to the naked eye where you like pass a magnet over this protein and all of a sudden, like it goes from no light to visibly lit up. And then from like, you know, you knew Marie at that point and then you, you know, what was your thought where you're like, okay, this is a weird thing in biology. How do we, how do I take this like bizarre thing that nobody cares about seemingly and make it into something that I think is actually pretty exciting.

Since day one, the Core Memory Podcast has been supported by the fine people at E1 Ventures. They are a young and ambitious VC firm in Silicon Valley investing in young and ambitious companies and people. Thank you so much to E1 Ventures for all your support.

So we were working with— so, okay, so I knew this guy Max Novenstern, and I've been working with him on some other stuff. At the time. Um, uh, he was, he previously was involved with Worldcoin, uh, one of the founders over there. Uh, and I knew Maria from the Future House and Arcadia days and like we wanted to do something and, uh, you know, but, um, Maria and Andy and I basically had the conversation of like, well, we think there should be something like here. Um, and, uh, kind of the question that, that the company we wanted to ask with Nonfiction was like, okay, uh, you can do this stuff with these fluorescent proteins for context. Like, I think that is very helpful for some imaging applications. Um, gets you— there's ways to exploit that effect to get more sensitive measurements. But we wanted to ask this bigger question, which was basically, uh, like, can you control biology, like useful biology with the magnets?, and it seemed like there was maybe, uh, sort of a narrow path if we were, you know, if the physics worked out where we could, where we could do that. Um, and so we started the company to basically take this same, uh, take this sort of idea of, of, of magnetic control and apply it. Yeah.

Um, so you guys are, my understanding of things, you're working on like magnetically inducible therapeutics, but then also like enzyme, like, yeah, basically trying to take the, you know, the field of optogenetics. So the idea that you can like make these proteins that are, um, light activated and sort of people try to adapt those systems to do things in brain, like the neurons in the brain. And then, but you're trying to do that with magnets basically. But, and more specifically, you're trying to make therapeutics. You can actually, uh, I don't know, maybe you could talk a little bit about that.

Yeah. Yeah. So there's a, there's a field called optogenetics, which, which I'm sure you know about. Um, but it's weird.

A lot of— I don't assume I know anything.

But a lot of, a lot of people don't know about it who like weren't, aren't biologists.

Sure. Yeah.

Uh, where basically you can have the genetics part is not that it's usually controlling genes, but it's genetically encoded. So it's proteins. Ways of engineering like living things to respond and respond to light specifically.

Yeah.

Um, uh, people have been doing optogenetics for maybe what, 15 years now?

Yeah. Like 20, like 2005, I think was like when the first paper on it was out. Yeah.

Yeah. Yeah. But popularized by Ed Boyden.

Yeah.

Um, if you're West Coast, you have to say Karl Deisseroth as well. And you certainly should. Yeah.

I mean, yeah, we go Ed Boyden all the way here.

As a Stanford guy, I have to, I have to pick my side.

Yeah.

Okay. Um, Uh, so, so actually in Ed, Ed Boyden's lab, also filled with lovely people. And so is Ed. Um, but, uh, but yeah, so the optogenetics was this like amazing research tool. People spent basically 20 years taking any sort of biological or therapeutically useful thing you could do and controlling with light. Um, but the problem is that like, if I have cells in a dish, I can hit it with light because the dish is pretty shallow.

Yeah.

Um, but if I want to get light into the human body, I can go like a centimeter.

Uh-huh.

Uh, yeah. Which, uh, most things, if you're going to go a centimeter, you can like deal with surgically or you can do other stuff. And so, um, being able to penetrate through the body is really important. Being able to penetrate like without like, you know, surgery, obviously a knife can also penetrate, but you're going to damage a bunch of stuff when you go. So being able to harmlessly get deep into the tissue is the is the new thing.

Right.

Um, so yeah, so that's what we basically set out to, set out to do. So the, yeah, so the idea basically is, uh, mostly we're focused on like antibodies, but also thinking about a bunch of other stuff, which we can talk about all the crazy stuff you can do with this.

Yeah. You know, there's a lot of places we can go with this, I guess.

Yeah.

Uh, it's most useful for me to think about sort of the problem area. So in my, in my head, based on our conversations, um, You know, what you guys are working on is antibodies and the idea is like antibodies for cancer, really for anything, but mostly for cancer. If we're talking about the applications people use it for, um, you know, you have things on the cells, the antigens that healthy cells exhibit, also cancer cells exhibit. And the idea then in the immunotherapy world is we, we want the world to exist such that like cancer cells all have these like unique identifying antigens and all healthy cells don't have them and we can just exclusively kill the cancer cells. And as we know, this has worked out and we have, we've cured cancer.

Um, yeah.

Okay. But we know that's not how it works. Actually, what we, what we have realized is that what happens is cancer cells, they evolve from healthy cells or whatever, something breaks and then they become, they grow too much. Uh, and the issue is it's really actually rather hard to target only the cancer cells and, and spare the healthy cells. And so people kind of work on different classes of antibodies where, you know, you target broadly, but those have toxic effects, or maybe target really narrowly, but then you, you know, you're missing some portions of the tumor that have an antigen that isn't expressed. But you guys are kind of trying to make the broad distributing drug more specific with a magnet? I don't know, maybe you can—

Yeah. Yeah. So the idea is if you can turn, if I can turn a drug on and off anywhere in the body, whenever I want to, I could have the drug be active at the tumor, but not active somewhere else.

Right.

So the reason why you can't get, usually, again, there are some patients who are very lucky and they happen to express something really crazy on the tumor. Usually the reason that doesn't work, as far as I understand it, is because your immune system is actually pretty good and if you have some tumor that looks totally different than the rest of you, your body's probably figuring that out and killing that tumor when it's like too small to show up on a scan.

Right.

So at the point, it's like that, like airplane, the survivorship effect.

Exactly.

It's like, it's the airplane meme. Yeah, exactly. So like, you know, the subset of tumors that are a problem for you are mostly ones that like look a lot like the rest of your body.

Right.

And it might look, it might might not necessarily look like all the stuff that's nearby, but it might be turning on stuff that maybe is active, you know, in your gut or active in your lungs or like these, it's turning on these things to help them survive and grow. And it's pulling these, you know, enhancing genes from your genome. And so these proteins are expressed somewhere else too.

Right.

And so the idea is, you know, usually you cannot look at those targets for oncology because you don't want to hit the tumor and also hit the lungs. Right. Because you've, you've traded one problem for like a worse one. But if you can locally activate the drug here and only here, you can turn it off if it's starting to cause a problem. You can start to dose stuff that would usually be pretty dangerous. But you can do that safely because you're only hitting the tumor.

Right. Right.

Or you're turning the drug off everywhere except for the lungs where, where the target usually is. So, so this opens up, um, like a big problem in biotech exactly is like we have this tiny set of tumor, like tumor targets that like we think are good. And there's all this crowding and like, you know, most companies are developing towards like the same set of targets and finding new ones has been hard. If you can, if you can open up these targets that like we've known about for a long time, maybe we've even tried them in clinic.

So like really like broadly expressed surface antigens that, you know, like every cell has or like maybe a lot of cells have them.

Exactly.

Yeah. And so you're kind of— is the pitch something like this? Like, you know, PD-1 maybe is like a good example of like a surface antigen that is pretty broadly expressed.

Sure.

You could dose somebody with a good amount of antibody against PD-1 and then, you know, it goes all over the body. But you only activate around the tumor using a really powerful magnet such that you're kind of controlling the response. And so you're like sparing the tissues that also have the antibody in them. Like the drug is there, but it's not working. Is that kind of— yeah. Okay.

Yeah, that's the idea. So you have sort of, I think about this as like two different categories. One is like existing blockbuster drugs. That like already work. They're just like, they have some safety issues and we can come in and say, hey, we can take that existing drug and we can make it better. Like a fresh version, better version of that drug. Or you say we could do some like crazy drug. Right. So PD-1 would be the former. So that would be like these drugs already, people use them.

Yeah.

You know, for a lot of stuff. They have some talks. I mean, I think if you ask a patient who's on PD-1, they don't love the drug, but they usually survive. And from, you know, in oncology, that's usually considered Thumbs up. A win. Yeah.

Patient lives.

Patient lives, I guess. Yeah. I mean, you know, and maybe, and maybe there's like a discussion about like what the philosophy of oncology care like should be and how we should think about that. But like, you know, um, as it stands, that ends up being a pretty useful thing. Yeah. So that's sort of the two, the two ways we think about it.

Yeah.

Um, so yeah.

So it's like you could, you know, maybe, maybe there actually is some other target. Every pharma company in America knows about it, but they're all like, well, we can't make the dosing work because people will just die. All of a sudden, like that drug actually might become viable or like that target might become viable because you could more specifically, you know, activate it somewhere.

Yeah. Those are the sort of conversations we've had with like pharma, pharma, like partners and stuff. There's often some like VP in the R&D department who's like, I wish I could hit this target.

Right.

We've got great data on it, but like, we just can't, you know, whatever the toxicity is just not going to not going to work out. You can't kill the patient and also kill the tumor. Nobody, nobody gives you points for that. So yeah, I mean, that's, that's basically the, that's the gist.

Yeah.

Um, and then you get lucky with like, you know, with magnets because like Maxwell's equations are like easier to simulate than biology. Um, I think you have these fundamental problems a lot in medicine of like, you get into the clinical trial and like only then do you find out how are humans different from mice in whatever way it turns out to be important.

Yeah.

Um, actually, like, there's a lot of things which, I mean, hopefully we get better. I mean, this is like the virtual cell, like, which we could talk about.

Yeah. Uh, there's some, actually there's some discussion to be had about AI and biology. I definitely want to have that, but yeah.

Yeah. But like, but like the advantage of like electromagnetism is like, you can predict it. Like my computer can run a pretty good simulation on a regular laptop. Map of like where a field is. And so the hope is that that is like a more predictable thing that will make more sense. And to the extent we can move this targeting question from this sort of obscure, like, which genes are expressed here and how does that translate to protein expression? And like, you know, all this sort of like complex biology that's like irreducible in the trial. Like, can we move that into like something that's easier to simulate, right? Like making the problem simpler is good.

And yeah, so I guess what you're saying is, you know, there are two things to worry about. One of them is like the mag— does the magnetic response work at all? So like that's the thing you're talking about being like easy to simulate.

Yeah.

And then if you have a target, I think, I guess all of the AI for biology tools, like the protein language models, I mean, there was an episode, I think, Ashlee just had Josh, you know, the Chai guys, right? And their whole thing is like, we're going to make these models that are really good at, you know, making, making it easy to create drugs for a certain target, like protein language models. So you're saying like, we have two, we've solved two different problems that we can make this space super easy to operate in because we can simulate the response to a magnet and we can simulate the creation of a drug. And so it's like, The pipeline is easy. Is that—

I think it may be. I mean, am I extrapolating?

Am I, am I—

there's so, so look, I certainly think that like, this is a, this is a technological thing. What we're doing here becomes a lot easier if the protein models like work.

Yeah.

Um, uh, yeah, I, I, I won't comment about the protein models working, but— well, I, I mean, so look, I, I definitely, I think the, the specific, like the protein antibody designs that I've seen, they work.

Yeah.

The AI enzyme designs, those also work. Like we use some of those now and they're great. Okay. I think there's a secondary question, which they're just starting to get into now is like, are these proteins developable? If it's not developable, then it solves the llama problem, like the alpaca immunization problem. And that's like a 3 months and like not that much money sort of thing. It's certainly helpful to solve it. I think it's, you know, if I was an alpaca immunization farmer, I would be like, you know, bummed out. Yeah. Yeah. I've got this friend who We did it recently down in South Africa with some like alpaca farm. And like, I'm sure that guy is like sweating.

They're eating my lunch right now.

Yeah. Yeah. But like, you know, if you're, if you're, uh, yeah, like, you know, the developability piece is the piece that then takes years after that.

Right.

And so is it immunogenic?

Right.

Does it have good stability? Like, will it actually last in the infusion bag for long enough to give it to the patient?

Yeah.

Um, uh, those things actually take a lot of engineering and most of the time and most of the expense. And, uh, that's the interesting, that's the thing people are starting to claim now. And like, that's what the Chai folks are, right? You're like, and that I think is really cool if that works.

Yeah. Yeah. No, definitely. I guess.

Um, yeah.

So I guess, you know, in, but you guys aren't only working on antibodies, right? So your vision is maybe like antibodies are very important. Obviously therapeutics is like ultimately, as much as we don't want it to be the case, it's very much what drives the biotech industry. You know, obviously it's good, like it's important to have drugs that work, but you also like, you have a vision and a step sort of beyond that with magnetically inducible proteins. Like, you know, what is, what does the next step look like?

Yeah. So, so, so it's important for a company to be focused. Yes.

Yeah.

So, so we're most of the company's aims are basically towards making really good magnetically controllable protein domains and finding out all the best ways, at least, to put those onto an antibody and to control those. But we also now increasingly work with a lot of like academic labs. Because from their perspective, like, this is pretty cool. Like, these are like new things to be done.

Yeah.

Um, uh, there's a nice, I think, spread of they get to do the, like, the cool science and get the cool paper., and, and we get to find out how these things work and, and mice or, or that sort of stuff. Um, so I think there's interesting, interesting stuff for, uh, like as a research setting, I don't know, could you have a mouse that like a little remote control mouse you control with a magnet or, you know, that's the, that's the long-term, uh, fish there. Yeah. Yeah. Yeah.

Yeah.

As a, as a company, I think we need to stay focused on the antibody stuff, which is important. Which is important.

It's very important.

It is.

I think, yes, it's very exciting and it's both important and exciting. Yeah.

Yeah. Yeah. But this idea of saying like, like for the first time, this means that there is like a way to control biology that doesn't affect healthy stuff, but it does affect the stuff you've engineered at arbitrary depths, you know, in opaque creatures and, you know, most of what you're interested in is opaque. So, so yeah, I think as a research tool, it's amazing. And we, you know, So like love talking to academics who are, um, yeah. Yeah. Um, working at, yeah.

Okay.

Yeah.

All right. So you guys, you're working on antibodies. You're also working on like enzymes as well as like magnetically inducible enzymes. So not just therapeutics, but maybe also like other tools for biologists to use, the industry to use.

Yeah. So the first one that we engineered off of that, so we, we, uh, first thing we got was this magnetically engineered luciferase.

Uh, which is another workhorse of biology.

Yeah. Yeah. Another workhorse. Also one that's like very, that shows up in a lot of like conspiracy videos as well.

Luciferase.

Yeah.

I'm not familiar. What's the, involved in the moon landing, JFK shooting?

I think it's some, it's like Lucifer, like Lucifer component. Yeah.

Of course.

The name of, they got us. Yeah.

You got me.

Yeah.

Okay.

Um, yeah.

So, so luciferase is this, uh, enzyme that's in the bud of a firefly. Right. Or there's a bunch of other species that also make it.

Yeah.

Um, some deep sea shrimp, some weird various critters.

Yep.

Um, but, uh, the idea with that was we've made this magnetically controllable luciferase and that's like a shortcut to engineer optogenetic tools because you say, hey, uh, luciferase makes light. If we bring a magnet near it, the luciferase dims, we can use that as just a to just fuse it on the opto system and it's like a hacky, but a very fast way to make all this opto stuff for 20 years, make that quickly magnetically available. Right. Um, so then, yeah, there's a lot of interesting questions, just like speed run, uh, redeveloping opto using, using this, this new enzyme. Yeah.

I mean, it's, it's, it's really ultimately, I think when you told me about when you were working on this a while ago, um, It's, it's, it's kind of like immediately to me, like all the problems that Optogenetics runs into all of a sudden, like become, if things work out, like it becomes resolvable. So like the main one you talked about is like, yeah, obviously it's really exciting if we could have these like channelrhodopsin or whatever, these light-activated proteins. People use them for like trying to map neural pathways and stuff like that. But again, as you said, limit, right? We can't, this light that's above me right here, it's like, it's not penetrating into my basal ganglia, like the deep nuclei of my brain.

Yeah.

And so people try to like do these very invasive things, like stick wires deep into the brain that have like an electrode light or whatever. It's complicated. But you're saying like kind of the pitch, at least when you, when you gave it to me a while ago, was all of a sudden like a magnet, not a fridge magnet, but like a, you know, a powerful one can activate in the same way. And then, you know, you can kind of skip a ton of like engineering challenges that like light-induced proteins have. Instead, it's magnetically induced. I don't know. I think it's very compelling.

I would also say, speaking of the fridge magnet question, so the magnetic field strength you need to get the effect saturates at about 10 or 20 millitesla. So that is actually the field that like is at the surface of a fridge magnet.

Oh, is that right?

Yeah.

Okay. But I assume like you need to be like, you know, very close, right? So if you're trying to like penetrate a skull, the fridge magnet is not going to work for you.

Yeah. Yeah. Yeah. Okay.

So 10 or 20 milli Tesla, that's about the level that you have to have a pacemaker warning sign. Conveniently, it's pretty much that same zone. Yeah.

I should know that. I don't know.

I mean, there'll be signs.

It's important. Okay.

So, so like actually technically, like the FDA says that like if you have an iPhone with a MagSafe magnets on the back, you need to put that in a big case if you have a pacemaker because, um, that the surface of it is strong enough. But once you go like a millimeter off, which you'd get from bulky case, right? Um, yeah, you're not, you're not causing the effect anymore. So, so yeah, we actually got really lucky with that because like, if you do the math with like a neodymium magnet, it's, it works out to be like maybe a couple pounds magnet to get the full depth in. But if we were 10x off of that, then the magnet would be too big. Uh, and if it was like a tenth of that, then, you know, then everything would be setting it off. So, Yeah, we got really lucky that the—

yeah, well, yeah, it was good luck, you know? Yeah, we could say luck. I don't know. I feel like engineerable solutions maybe, and some probably engineer— probably could have gotten there regardless, but you know, okay, we'll say luck. You got lucky, Richard. Everything you've ever gotten is because of luck. There's no talent. It's all luck.

But Maria got lucky with that at the beginning.

Yeah.

Yeah. And then, and then, and then from that we've been trying to say, you know, the minimum set of, set of things need to get lucky.

Yeah.

To win. Yeah.

Okay. Um, um, All right. Well, I, I have more questions for you. We could talk about nonfiction. I have a lot of stuff about biology in general if you want to, I don't know. Is there—

Yeah, let's talk about something that's not nonfiction stuff.

Okay.

Okay.

Other interesting things to talk about. Yeah.

I mean, there's some threads that we can pull on here. I think, you know, interesting one, because we're talking about AI for, for science, like we're talking about protein language models. We're talking about the idea that you can simulate how proteins fold and therefore make it easy to design drugs.

Yeah.

Um, you know, in my mind, part of this conversation that exists, like, right, you know, right now there's a lot of momentum around trying to like redesign the machines of biology to run on AI. So like, for instance, there's like, uh, Mission Genesis or whatever, like the government, um, program that's trying to like use all the, the, the data from all like the national labs and stuff for, for scientific, you know, AI for science, like all the AGI labs and stuff will get access to that data.

Yeah.

You know, but so like I see that in my mind, I think like it definitely makes it easy in theory to like maybe design drugs better, right? Like maybe if you have all these like AI tools that can look at different receptors, they can come up with new drugs, right? But the thing that I guess like, you know, what you worked on a nonfiction sort of what people, CRISPR-Cas9 for genetic engineering. Like, I guess the big question is, do you think that like AI actually allows you, at least as it exists today, does it allow you to like actually interrogate these like really out there mechanisms? Can you actually like learn that from an LLM? Just say, okay, this all exists in the same, you know, do a latent search of your LLM mind and come up with some way of inducing a protein that doesn't involve light.

Um, so I think the simple answer is like, it's not— I think we're a couple years off from it being like—

I need you to have a hot take. Hot take. I need you to say the words, uh, AI for science is never going to work in biology. I need you to say that.

No, I— okay, I— here, okay. I'm not going to take the bait. Um, uh, here's how I think AI has like actually been helpful, like for this sort of stuff. I think AI literature review is like objectively like very helpful. Right. Um, uh, it reminds me there's a professor I'm going to butcher. You'll just cut this out if I say the guy's name wrong.

No, we're not. Uh, don't cut this out.

There's a guy, there's a guy, I think his last name is Brandywine. Um, there's, do you know this lab?

No, I don't.

This guy's papers are really funny because like you read them and then like he's almost always doing something that's very prescient, but it's like 20 years too early. So you can go back and read this kind of stuff.

Okay. And they're amazing.

Beautiful papers, but like using the tools, like you can just barely get it to work with tools then. And then 20 years go by and then people are doing it all the time.

Yeah.

And so like, I think being able to read, so one answer is okay, if you wanted to have an edge in like the pre-AI era, you'd go back and like read all this guy's papers.

Right.

Um, but like now you can just have, uh, Semantic Scholar or ChatGPT Pro or whatever, like read 50 papers for you.

Right.

Um, you can rig up your own system if you want. You know, certainly don't use, you know, Sci-Hub or anything like that.

Definitely do, definitely do not do that. Yeah.

Yeah. But, but, you know, you can, you can very quickly read a bunch of, you can, you can comb through papers for old insights. And I think that's actually super helpful because often like the good info that you need, if you're doing something new, is often buried in like some weird part of the results section. It's not like, like they weren't thinking and framing the paper like you use it the way you're going to use it.

—right, I guess it requires like reading some basic bench biology and then being like, oh, how is this useful for this other unrelated, completely unrelated problem space? Yeah, I guess like the thing— so I hear you on that. I agree. I think, you know, today if I were like, I want to do some, you know, cool biology using some emergent thing from 2007 that just never took off, like, I think an LLM is useful for that. Yeah, I guess the bigger question that I have is like, You know, right now a lot of the momentum is towards making AI-based biological research. Like, Ark Institute is definitely oriented around that. That's a large new research institution and AI is like embedded in their stuff. You know, there's Future House, obviously what they're doing, and they got a decent amount of money to do that. Now the AGI labs, and now again, as I mentioned, the government program is oriented around that. But I think the thing that concerns me is like, you know, you're correct that you could dig up stuff from 15 years ago using an LLM, right? But what if now all of a sudden we're just saying we're abandoning the idea that people should work on like random biology and especially with, I think like academic, the academic system where like, you know, it's smaller. Yeah. It's getting smaller. We've been lucky that there's been at least, you know, you can say there's been money that's been wasted, but ultimately there was money that existed for people who wanted to work on weird shit. Right. So it was like, oh man, you know, it'd be cool to look into like, uh, uh, shrimp, the deep sea shrimp. And, uh, oh, actually there's this alkyl, you know, phosphatase and, oh, actually that's pretty useful. Um, or, you know, jellyfish. Oh yeah. And I'll just win a Nobel Prize for that. I guess like, do you, maybe my question, maybe I'm baiting you. I don't know. I'm baiting you.

Yeah. Yeah. I know you're baiting me. I think the, uh, Well, okay. Well, I have a comment about like the virtual cell stuff, which I think, uh, okay, let's hear the, like, yeah. And then I'll get back to, I think there's, there's other, like to directly, more directly answer your question. Okay. Um, I think the, the virtual cell stuff reminds me a little bit of, um, were you like kind of around for the systems biology flux balance analysis? Oh yeah. Yeah.

That, yes, that joyful thing.

Yes. So like people had this idea, like maybe 10, 15 years ago. Um, I think that's kind of roughly the timeline. Um, closer to 15. Yeah. Maybe 15. Yeah. Yeah. Yeah. It was kind of tapering off 10 years ago. Um, that like, you know, the cell is just merely a system of stocks and flows and you have these metabolites and then you have these enzymes that interconvert the metabolites and I can calculate the rate of enzymatic, I can get the enzymatic kinetics. Kinetics, so I know how fast they convert. And if I measure enough of the stocks, I can model like through a series of like, you know, matrix math, basically how the system is running. And my understanding, like to condense many years, perhaps lifetimes of work into like a pithy statement. That's what we do here. Yeah. Yeah. I would basically say like the, The problem you ran into was you just, it's hard to get the data. So like cells are heterogeneous, the concentrations aren't the same everywhere. Different cells also will vary in their enzymatic kinetics and there's all these like weird variations. And so the data just wasn't there to answer all the questions people wanted to answer. Right. In like toy models, it was useful though. And like there was like, there's a lot of metabolic engineering companies that got going and like, yeah, maybe some of them went into bad markets, but like a few of them went into good markets. Yeah. A lot of money into bad markets.

Uh, yeah, mostly bad markets.

It was a glorious time. Yeah. I, I, but like, you know, the, the end of this story was basically like the, the data, the models are only as good as the data used to train them. Right. And like the models usually don't tell you something that wasn't in the training data. And like, this gets back to like the classic, like, like the classic good bearish take on all this stuff is like, You know, any training set sufficiently large to train a good model usually has a perfectly adequate working example, like, stored within the training set. Right, right. And so, um, you know, that's going to be the question for these virtual cell models of just, like, can they answer arbitrary questions, or are they best equipped to answer the questions which, like, we're specifically training them to build, right? And, like, LLMs did have all these, like, emergent properties, right? Um, but it's, you know, maybe you say that we weren't emergent. Maybe like we—

maybe they were actually just in the training set.

Yeah. Maybe it's sort of in the training set. Something. I mean, it's, you know, amazing stuff is in the training set, but like, um, uh, until you build the models, it's hard to know, um, how good your dataset is. Yeah. So it's good people are building them and then we'll find out.

Yeah. And I mean, I think I definitely, there is, there is utility in a cell model to be clear. Like I'm not, I'm not like, uh, oh, AI is bad for biology, obviously, like I did have a career that was based off of that for some time. I think that, yeah, maybe the question is more like if we build a sufficiently, like what is a sufficiently large biological model look like? You know, AlphaFold was sufficiently large to address like one specific problem within protein folding, right? It solved some type of folding. But there's still a whole realm that exists around like dynamics that isn't really well integrated into protein folding. Like what does it actually take to make like a unified model where you can just like ask open-ended questions of like, you know, maybe you can make a model that's really good at, say modeling like what a glioblastoma, what kind of drug might work in a glioblastoma, but that might not translate whatsoever to some other like endometriosis. I don't know, at least in my head, that's kind of how we view it. But I'm here to generate hot takes and maybe not intelligent ones.

So. Well, I mean, I think like you're talking, you're thinking about like the protein folding problem. Like I think that like, so the sort of stuff we've done with like engineering magnetic proteins, like if I didn't have a way to look at what the shape of the thing was, it would be really annoying to design this stuff. And so like in a pre-AlphaFold world, you know, some of these domains have like a crystal structure and I could have figured out like, I could have drawn out in my head what it looks like, but like, it's really helpful to just be able to kind of open up a simulation. Yeah. See what thing is near something else. Cause you're making stuff that's supposed to fit together physically. Like these are locks and keys you're building. It's nice to know what the lock is shaped like building the key. So like, so that bit I think is like big. Um, but like then there's all this questions about like what's useful clinically. Right. And, you know, I think another key— our mutual friend Rox, Roxana Teslow, you know, has a bunch of interesting hot takes about like, can we get more public clinical data? Right. Obviously large pharma companies, you know, this gets back to the point about like, you know, these cool targets that like are too toxic. If a drug fails and you're not bringing it to market, you often aren't going to fully disclose all those results and publish your paper and get the trial out. Right. And so there's a lot of interesting clinical data that just never sees the light of day. It would be really cool to get more of that out there. Right. But until that's out there, you kind of have to go with what you've, what you've got. So like, that's the, that's like right now, I think everybody's sort of white knuckling it through. Okay. So until we actually see how these things work, we don't know, you know, we don't know where the value is going to, where the value is going to fall. And I think Um, it's definitely interacted weirdly with like the patent landscape, for example, it's like you can't really file, you know, right. If my patents are all on structures of a protein and I can make an arbitrary structure, like, do I have patent? I mean, yeah, right. Like, like how much, how much edge do I have in the sort of, yeah, as a protein designer, if that's all that you do, right?

I mean, it becomes advent— like advantageous to not disclose whatsoever that like like actually having a patent is not useful because you can just design around the patent. I mean, that's like what OpenCRISPR is kind of about is like we can make CRISPR constructs that are like around the original patents. So the OpenCRISPR stuff is interesting.

Have you talked to the CRISPR, some of the CRISPR folks about what they think about OpenCRISPR?

I certainly have spoken to some CRISPR folks about that.

Yeah. Yeah. Yeah. I mean, I think, I think the, the people who hold the original CRISPR patents will claim and will say, I think with some credibility that like really what they've claimed there is like constructions of different domains. And they do have claims which have been, you know, approved. And I don't know how they'll, I don't know if they'll win a case on it or not. But the claim is like, this is not specific to a sequence. This is a construction of the, you know, it's a conceptual thing. It's like this thing like this and this thing like this, I can put them together. And I can get the effect. Yeah. Um, the problem is just you can't do that with antibodies, which is most of drugs. And so like the open CRISPR folks maybe are solving, um, well, it's interesting. Maybe they're solving a real problem if like it's a novel sequence, but if it's the same, if it's just a hallucinated version of the thing that someone else has already claimed, maybe you haven't escaped the original patent. Right. But these antibody things are definitely escaping the original patents because you can't claim antibody, right? Um, yeah, whatever the first, I guess there's Camelid for 20 years. There's like Camelid antibodies, like the alpaca stuff, nanobodies, nanobodies trademark, right? Right. Um, they could claim that, but now you, now you can't. So, right. Um, yeah. And the patent will disclose your target.

And so, you know, you, you've created a shortcut where anybody can basically make your, yeah, I mean, like it's, you, you would just need to take the sequence and then maybe like generate a library around that sequence. But you know, that's, It's maybe weeds. I do think that the alpaca thing is like another example, not, not to belabor, uh, at least my, my talking, not to talk my book here, but I mean, I think that the, the, the, the alpaca thing is actually another interesting example of this where like, you know, if there wasn't some person who was like, I think it's important to study alpaca biology, like maybe you actually never discover that actually, oh, their antibodies are different than human antibodies and that's useful, right? Like you actually We just never, just never learn that. I, at least I personally would be very surprised if based off of everything we know about models that exist today, maybe you could argue that with true superintelligence, like it can just like intuit these things or maybe it wouldn't matter. But I think like it's impossible to make like a leap, like a human antibody looks like this. And actually there has to be some organism out there whose antibody looks different And we can use that to our advantage. Like, I just don't think you'd get there. Yeah. Um, you know, which I don't, it's not to say that people are necessarily saying we should abandon studying weird stuff, but I do think, um, they're sort of implicitly, uh, creating that world through the action of like taking money out of the weird biology system and putting it, funneling it into the AI for biology planet. I think, yeah.

So I guess what I would say is I think the weird biology, the advantage of that is that it's novel dataset construction. Right. And, uh, and the architecture of a lot of these models is not that complicated. Right. Some, some of them are, I mean, like AlphaFold structure is actually like, yeah, there's a lot of weird little parts. That's like a crazy paper. Um, but like novel datasets, I think are generally important. Um, I think letting people, uh, yeah, doing some stuff that's not obviously useful is helpful. Uh, and then occasionally also being willing to like white knuckle it through and be like, I'm just going to try it something. Um, but I would also say like, I think a lot of those, like we should just try it. Some things can often be really cheap. Like people often, uh, like the good stuff, you usually have some idea of what you're doing. Um, uh, yeah, I don't know. Yeah. Yeah.

You know, we, we, we can, we can, we can move on, I guess, like, you know, ultimately, uh, it doesn't really matter. The systems are moving in that direction. I do think that one interesting question, you know, we're talking about biology, you know, you've been in the space for a while. I think China, you know, just on the topic of things that are currently being discussed a lot in biology, I mean, I think China is another, you know, it's a really big open question where I feel like my take when I, when I sat in the chair you were in a few months ago, it was like, well, China is like this, you know, rising power. In biology, but maybe American enterprise is still like, uh, able to win out. I feel like it's actually kind of unclear to me nowadays. Uh, I don't know. I'm curious to hear what your thoughts are on this.

I, so, so like, what are the paths? I mean, I, I watched your episode, of course, but like, what is the argument where, uh, where the US like wins? Uh, you know what, I don't know.

I feel like, wait, hold on. I'm in the, I'm in the host chair. Oh, okay. Yeah.

I'm just curious. Like, I think, cause people have different theories, right? Like how the US beats, beats China. Right. Right. Um, I do think that like, it's a, as an American, it's like a good world if America has, you know, most of the drug development here. Um, there's a lot of soft power sort of stuff and the US chooses not to flex that. Like, we're happy to sell drugs to most parts of the world that are happy to buy them, but like usually steeply reduced prices. But, um, I don't know, like, I think it's good for it to be here. Right. Right. The, uh, some people will argue that the answer is going to be like automation. So we'll just have like lab robots do everything. Um, that has been like a thing for the last, like, again, couple decades of that. Yep. Um, it seems like the challenge is like you can automate a lot of the hours of someone's day. Like you can have an OpenTrons for $5K. Yeah. That will do the pipetting for you. Right. Um, and those are like all over now. They're in a lot of academic labs and they're gathering dust. Yeah. You know? Yeah. And like, I mean, like, like, I think the challenge is basically that like the, a couple of things. One is replacing biologists with software engineers does not necessarily make things cheaper. Right. Uh, uh, sometimes it does, but sometimes it doesn't. And the reason why it doesn't is that like to properly quantify everything you're going to do in every edge case is really hard. And a good biologist on the fly at the bench will be making choices If you're doing research, yeah, we'll be making choices. Yeah. Um, if you're doing like clinical diagnostics, you hopefully are not making choices.

You're, you're up, you're turning a crank. Yeah.

And you really should not be changing, you know, genes. Yeah. Yeah. Yeah. It's like, of course. Yeah. Um, but, but for early stage research, I think, uh, yeah, you want to have kind of like distributed intelligence across the team, not just like a software engineer and a biologist kind of yelling at each other that like, you're being too specific. You're like locking me in. And then, you know, robot Claude is doing all the work. So I think that is challenging. And like these new AI tools maybe are going to make it easier to like, you know, like a Claude code to convert.

Vibe code a cancer drug, you say? Yeah. From my apartment, I can vibe code, you know, Sarkozy's cure or something. Yeah.

So maybe you could have like a vibe coded, like you might be able to write a protocol and convert it and like I've tried doing that actually. You can put like a, an easy protocol in and say, write me OpenTrons code that does this. Right. It'll do an okay job. Yeah. Um, maybe in a year from now it does a great job. Right. Um, but that's still like a part of the puzzle and it's not like the big part. Right. Um, and it slows down the iteration times. And I think that's what kills a lot of like, there's a lot of projects that get killed basically because the iteration times are too slow. Yeah. Um, So yeah. Yeah.

I mean, I think like feedback loops in biology, I don't know. I feel like I'm rehashing some of the same things I talk about.

I know. I, I know.

I, but these are things I think are super important, right? Yeah. Yeah. And I think like, you know, Stephen, our mutual friend Stephen Milena has like this article about like fast feedback loops in biology. I think that, um, you know, I think you're correct that maybe towards your question of like, what is, what is a, what's an American win condition look like? Yeah. It's probably something like we can like leverage automation better. I guess the question to me, you know, I think China kind of developed automation for stuff that had nothing to do with biology for a long time to make like, you know, this mug and like various knickknacks that you can find at like a souvenir shop. Yeah. Um, and then it got more complicated and they were like making like, you know, high-end stuff. Um, and then eventually it's like, oh, actually we have this skill set, which is like, we can automate anything. Let's apply it to biology. And so now like CROs in China get to take advantage of the fact there's a huge workforce of people who are very good at creating pipelines to automate stuff. Like we don't have that in America. Yeah. I have to rely on OpenTrons to make me a good tool. Like frankly, yeah, I don't know. I have no idea. I'm like, I would buy an OpenTrons. I mean, I remember meeting, uh, Keoni Gondal, I think like, yeah, a couple blocks from me.

Okay. Yeah.

So, you know, I remember going to his like home lab in, uh, Huntington Beach in LA. And being like, holy shit, like this kind of guy doesn't exist. Like this guy is like a rarity. Rarity here. Right. Yeah. Whereas in China there might be like 10 million people like that who are just like, yeah, I'm ready to make you this pipeline, not in your bedroom, but actually I'm willing to, to do this like at a scale where we can just test any drug, uh, through a completely hands-off thing. I don't know. I feel like win conditions, it's, it's hard to imagine beating that system out.

Out. So, so, okay, so here's a couple. I'll, I'll give you like my best case. Okay, let's hear it.

I need, I, I need to hear this.

Yeah, some hope. Yeah. Uh, uh, no, I mean, I think there are like real reasons to hope. Um, well, okay, a couple things. First off, like the Emerald model of being like, okay, let's automate 90% of it and let's have humans fill in the gaps. Yeah. Like it's probably something that is achievable. Like, right, like the Cloud Lab thing.

Yeah.

I mean, it's happening at Emerald today and like, there will be, my guess is the next 15 years of that stuff in the US looks like that. Unless like, you know, Claude 6 is amazing and, you know, whatever. But like, but basically like that's doable today. Um, there's the other thing which like we've done with nonfiction stuff, which has been this Argentina lab.

Yeah. I actually, I'm glad you brought that up. Yeah.

Another weird tip. Another weird. Yeah. So, so, uh, we sort of had this question of like, uh, we think iteration times are very important. Uh, we also think like cost is very important. Right. And the question was like, how can we have a cheaper cost of experiment? Uh, you know, like, okay. Like one person, I think core memory people obviously knows like a lot of the SpaceX story where they were like going to be cheaper than NASA. They were going to be faster than NASA. And they did that through vertical integration. And like our philosophy was like, well, okay, can we be cheaper than the NIH per experiment? And can we be faster than the NIH per experiment? Yeah. If you're working off of US talent, very hard as a startup to pay less than the NIH pays grad students, which is not a lot. Yeah, right. But there are actually like amazing Western allies where the prices are not very high, certainly in Europe. But in particular, we found in Argentina, we've set set up absolutely like an amazing lab with great people. And they are just like happy to be working on really cool stuff. And we found that we basically can say, okay, we have like a lower cost per experiment by far than like even our like academic collaborators. So sometimes we'll just volunteer to do stuff. Okay. You know, like we can do a lot of crazy stuff there. Um, and so I think that there's a, uh, like there's a lot that the US can do outside of China, uh, with allies that are maybe like have less of an IP concern, right?

They're not going to rip, uh, nonfiction IP and just make like a, yeah, you have me better or whatever.

You have a, you have a, you have a better potential to like, to, to, to have like your own., yeah, like sustainable business and to hire great people. Yeah. I mean, the China, like salaries are not as cheap in China as you would think. Yeah. Uh, they're, you know, it's not that much different than Europe, right. On the, on the scientist prices these days. And so like, um, yeah, like, I don't know. I think, I think lowering costs, I think is very important. I think there are ways to do that by like offshoring some of the scientific team. Um, I think that like, um, tighter iteration loops is something that the US has an advantage on. Um, just in that, like, I do think we have a deep thing in the culture where like, we would rather do a quick and dirty experiment. And sometimes that's better than scaling up a thousand robots to work on the wrong project. Did you find out in two months that it was dumb? Sure. Um, and I also think we have a lasting edge on like cutting-edge science stuff. Yeah. Um, like the magnetogenetic stuff we're doing, like you of like at this point, and China is like changing a lot. Sure. So like, sure, you know, whatever people said 10 years ago about China is, yeah, goofy now.

Yeah. Like, oh man, all they do is like, I don't know, that was, that was my take. That was my take. Yeah. Lotta actually was the one who called me out on that. She was another guest on the podcast. Yeah. She was the first one who was like, no, actually you're incorrect about that. Uh, and I think I've definitely been reformed in my opinion.

But now, I mean, I think what you're seeing now is like China is clearly able to do— so a couple of things. One is like you have the Turing test thing that moves and people just progressively say, oh, well, yeah, sure. Of course anyone can do a VADJAF bispecific. Yes. You know, like the stuff. And those were things that like American companies were saying were hard to do. Yeah. And it turns out they're like, um, like there's a large set of people who can do them if they want to. Yeah. Um, and their clinical trials are obviously running a lot faster than ours and there are structural reasons why that would be hard for us to do it as fast. Yeah. Um, yeah, I don't know. I, I think, I think there's like reason to hope, but I suspect that like probably the future of the US biotech industry, the extent that it like continues to exist.

Yeah.

Uh, which I think it will.

I mean, um, but to some degree, like, you know, yeah, to be clear, yeah, the worst case we're really talking about here is like our industry becomes second to China. I don't, I think I would agree with you. I do think that like the American innovation engine is quite powerful. And it definitely would take a lot of bag fumbling on the American part to allow that, you know, to be completely overtaken by China. Yeah. However, you know, so what someone could argue— one could argue, and in fact has argued, uh, that, you know, we're certainly doing our best right now to drop some of the things that are advantageous to us, which is like taking advantage of— people want to say weird things like, uh, you know, magnetically, uh, activated proteins, But yeah, I think there's another topic I actually wanted to talk to you about. There are many things we could discuss here. I think another area is, you know, biology, biotech is done for many reasons. Mostly it's for, you know, to this point we've arrived at like healing people, making medicines for people is the best thing we can do with biology. Consumer biology is like not really taken off whatsoever. And, you know, people have tried different things. You know, I think there's the, the LA project, the idea that we can engineer organisms and, you know, have a unicorn as a pet or something like that. But, and there's plant engineering stuff that people do, which I think is cool, but largely there's not any like consumer biology products. Maybe you could argue GLP-1s are kind of like that, but it is a therapeutic. I don't know, how do you, do you see a world where there's actual, like a whole class of like consumer bioproducts?

Um, I mean, so there are examples that like exist today of consumer bio stuff. Okay. Um, I think the challenge has just been like getting people to understand them as, so, okay. Like the joke in synthetic biology circa 2015, like the good quip to have was that anything, uh, any successful version of symbio stops getting called symbio. Sure. It starts getting called like CAR-T or something else like that.

So it's getting called kombucha.

Yeah. So, so like, if you think of like two examples, there's, there's Evan Zhao's company, the, uh, I'm blanking on the name. I, I should give them some free, free, uh, marketing. I can't help you. I don't know what it's called. They do, they do, uh, uh, cat, uh, like nanobodies that like catch cat dander. And so the idea is you spray it on the food. Okay. And then, um, the nanobodies like somehow inhibit the allergen that comes from the cat saliva. And so you don't get the aroma. Yeah. You don't get the, like the, the piece that makes you allergic. Um, like that's not branded as a biotech consumer biotech product, but that's, you know, that's the same process that these other companies were doing. Sure. You still had to do the, yeah. You know, the camel at the alpaca farmer, he's calling that guy up, he's doing that stuff. You know, another example would be like Solugen where like their pitch was like, we make hydrogen peroxide. It's like, it's like, would you like cheaper hydrogen peroxide? And so, you know, again, I think, I think that those companies like, it's still outside the mainstream of traditional therapeutics. I don't, I don't think that like,, you know, some of these will probably have large outcomes, hopefully, but like, they won't necessarily be the like billion dollar or $10 billion a year, um, thing. Yeah. Uh, and if they are, they won't be branded as a biotech product. They'll be branded as, you know, a consumer thing. Sure. Um, right. And I think that's like, so that seems to be the future. I think there was like a lot of emphasis in like, again, the 2015s of like a chair that's grown from mushrooms or something like that.

Like, what is this chair business?

Everyone's— people are really into chairs.

People are really into it.

But these chairs are so much better than the ones we're sitting in right now. Yeah, exactly. Like any mushroom chair, right, could be, right?

I guess there's something to be said for like, you know, this is an industrial machine-built thing that biology never needed to make a chair. So there's not really like any, you know, the purpose of biology is not to make chairs. It's not, you know, I guess it's, it's hard to engineer something that's really meant to just like survive into like making, like ultimately it's very cheap to make a lot of industrial processes. It's very cheap to make plastic bags. Yeah. Um, it's actually very expensive to make like a bacterial cell capable of producing a polymer that can then be used to make plastic bags. It's, it's like a lot harder. Yeah. Um, that's always been like my view for why I think like, you know, maybe I didn't say this. Ashlee had asked me about the chair thing and I was like, man, uh, I don't know.

Now I, you know, I think reflecting back sort of what you're getting at here is, you know, well, it's like, I think you need to start from, this is the challenge is like businesses are like, need to start from the thing that's actually going to like work on a timeframe. I think that like one of the challenges in biotech is like, if you're doing therapeutics, people sometimes will cheat on their verbs a little bit. They'll say like, when is the drug going to be in humans? That does not mean it's making money. That means it's like 6 years from making, starting to make money. Um, you know, and so like, uh, biotech ends up being very dependent upon financing and continuing to be in the good graces of getting these, you know, boluses of cash when you need them after the last bit's de-risked or within a timeframe before the next bit. And so I just think like, it's very important as a biotech to be focused on markets that will make money. Yeah. Uh, low risk of making money. And I think one thing we've definitely seen from the Symbio wave, a lesson a lot of people learned like the hard way, um, was like, people do not pay a premium because it's cool. Yeah. Uh, yeah. Yeah. Like, you know, like the things have to solve problems and most of the problems biology have successfully solved so far have been therapeutic problems. Right. Right. And so, you know, cause like dying is like a feeling sick. Yeah. Or like big problems. Right. Yeah.

Yeah. Um, you know, I, we have a few minutes here actually. There's one other thing I wanted to ask you about. You know, it's sort of what you're getting at. You just said something like biology is very, it's very reliant on capital. I think in a way that like software maybe is not, I mean, the capital and software is very much like salaries and maybe some infrastructure, cloud infrastructure stuff, but like it's mostly salaries. Biology is like a lot of not just like salaries actually might not even be half of your cost. A lot of it is not that. And I think like the dynamic I wanted to get your opinion about is like, seems to me a difficult dynamic in a lot of biotech is that it seemingly is like very reliant on like finding your billionaire. It's like you need to find your person who with a spare $500 million to drop on like cool biology and pitch them to get excited about the thing you care about. I don't know. I feel like that is a, it's a difficult place to be for this space right now.

Yeah. I mean, I think this is, I, so some people frame it that way. Yeah. Uh, I think the, the parts that are different about bio is like, um, if most of the people who invest in like Salesforce or like some SaaS I'm picking on Salesforce, but like invest in some SaaS company. It's because they've looked at the CAC, they know the customer acquisition cost, they've looked at your total lifetime revenue, you know, value of the customer. They're doing the napkin math and they're underwriting based on that. It's pretty easy to underwrite, but you're not usually doing it for like some other higher motive of like, you think the world would be better if salespeople remembered the names of their clients or something like that. Sure. Okay. Um, uh, in bio, there is also this aspect of like, you're making the world a better place.. Yeah. And so like often the highest bidder for some projects has been like someone who wants to make money, but also, yeah, like have a legacy of something that you've done. Um, and I think that, that, that is like a, a component of bio. Mm-hmm. Um, it tends to be most true at the beginning when companies and projects are like really kind of, uh, you know, in their like most vulnerable early stages. Yeah. Once they've started to de-risk, it's, it's you can start to do that napkin math and say, okay, your average oncology product has a 10% chance of success and they roll through. Um, uh, I think like this is also a difference between like the SF and the Boston model. Yeah, yeah, yeah, exactly. Yeah. Yeah. Whereas like in SF, it tends to be like a founder-driven thing, you know, the money tends to come in after the idea. Boston, it tends to be like the VCs are in the room and it's kind of like a VC's creation. Yeah. Like they're sitting in the academic lab waiting for the Nature paper to come out.

Right. The data needs to be there first, I think, in a lot of, in a lot of ways, right? Like you need to see, I mean, like Dyno was like that. Yeah. Like there was a science paper that was, the data was anticipated already and that could be used for fundraising, but ultimately you couldn't just like sit here, you and I are just talking about it and I come up with something and then you're like, oh, actually here's $5 million based off of that. That. Yeah.

I mean, I think the, the SF ecosystem is more built on like, it's the SF funding ecosystem is built out of tech. Yeah. Right. And like Genentech in the early days here, um, it was Kleiner Perkins. They did a lot of other like tech funding at the time. It's funny in some ways, I think there's like this digital aspects of traditional biotech that still look the way that like tech funding looked 20, 30 years ago. Um, but it's continued to have this like cultural infusion since. Yeah. And I think that's why the SF ecosystem looks different is just like, I don't know, different cultural legacies of the two places. Yeah. Uh, yeah.

Yeah. I mean, I guess, okay. Maybe you're right. Maybe I'm being, uh, I'm too bearish. That's sort of my, uh, but ultimately I'm here to create takes. Uh, that's why Ashlee Vance has given, you know, given me this microphone.

Uh, your takes are fantastic. Okay. I appreciate it.

Just let me know where to send the money for this. Okay. You know, we just have a second here left. There's one thing I wanted to ask you. I didn't bring it up at the top of the show. I didn't want it to be too distracting. Okay. Richard, you are sort of related to a dynasty of— oh, yeah, of bio, bio-related, biotech-adjacent people. Maybe we can go back to somebody that I think I think people might be familiar with your grandfather, also named Richard Fuse. Maybe you can talk a little bit about sort of his dynamic and maybe if he pushed you towards biology. Yeah, for listeners here who are not familiar.

Yeah. So my grandfather, for listeners who maybe don't know, my grandfather was next door neighbor with Elizabeth Holmes. And then got into a series of disagreements, different levels of unhappiness. Multimillion-dollar disagreements. Yes. Part of the Hulu show is sort of built off of that downfall story.

Yeah. It's also, you know, a good portion of a New York Times bestselling book.

Bad Blood. Yes. A couple few. Yeah. A couple last name-related puns in the chapter title. Uh, uh, yes, I, uh, so when I was like a kid, I knew I wanted to be an in— like an inventor, which basically meant to me doing whatever he did. Yeah. Uh, both my parents were doctors. And so like, and I have two more uncles who are biotech patent lawyers. Uh, we'll get to that in a second. Yeah. Uh, uh, so, so yeah, I don't know. I think I, I think I had a sense that this stuff was possible early on. And I think that was important. Um, but, but yeah, we do different kinds of, well, okay. I would say we do different kinds of stuff, but now we're doing like, you know, if my grandfather did, he did a lot of stuff in drug delivery and my dad has cardiac MRI. So he does, works with magnets and my mom does this, you know, uh, my mom's a general internist, so she runs a startup kind of in a different sense. And now I feel like I've like come back into the family. Yeah.

Actually what your parents had envisioned was like, Richard, we want you to start a mag body, mag therapeutic company. Yeah.

It's actually the entire entirely explainable now based on— yeah, I see. Yeah, yeah. So I don't know, clearly, clearly there's some influence there, although like it's easier to say post hoc. Yeah, maybe.

Yeah, I don't know. Unfortunately, maybe there's some nominative determinism for people who are, uh, uh, more, more, uh, better equipped to do punning right now for doing fusion proteins. Yeah, exactly. Yeah, yeah. Uh, your— one of your uncles is on the patent for Suboxone, right? Uh, I'm not sure.

I think he's on a Suboxone-related patent. Oh, okay. Suboxone-related patent. Okay. Um, okay. Related to the drug delivery around Suboxone. Okay.

You mean like taking a pill and putting it in your mouth?

Uh, I think it was, I think it was trying to solve the question of like, I don't know, apparently this is like a drug that is not easy to deliver to the people who want the drug. Yes. Yeah.

Are not necessarily, uh, uh, uh, it's an important support.

Yeah. Has some user, user issues with it.

Yeah. Yeah. Okay. But yeah.

I mean, he's got his name on a couple hundred patents between the two of them, at least. So, yeah.

Okay. I see. All right. So yeah, maybe you're always destined to—

I don't know all of them.

Yeah. Yeah. Okay. Um, okay. Well, is there anything else you want to share for, for the, for the audience before we close out here?

Uh, let's see. I think the only thing I would, if I can, if I can plug, you can plug. Uh, I would just say, uh, anybody who, if there are any academics listening who want to work on magnetic drive Uh, hello@nonfictionlaboratories.com. Please reach out. Uh, you know, we'd love to, we'd like, we love working with people and just advising and talking about whatever. And it's a, it's a low pressure thing. We just, it'd be great to get this out in the hands of people.

Yeah. Cool. Um, yeah, well, I'm very excited. I'm, I'm really glad that, um, you know, when, when Ashlee Vance and I were talking about, I was interested in doing an episode, bio-related episode for Core Memory. I was like, you know, it has to be Richard. Richard has to be the person that, that, you know, I do this with.

So is this your inaugural podcast host?

Yeah, this— well, uh, real RNA heads will know about one other podcast that I hosted, uh, many years ago that, um, you know, this is for like the 4 people that are familiar. I'm not going to talk anymore about it, but okay, uh, there's some deep cuts, there's some extremely deep cuts. This is very, very much for the 4 people who would know about it. Um, yeah, this is now a girl. So, and I think you, you made it easy on me. Um, so I really, I really appreciate your time. I definitely think for anybody who is in the space of doing weird stuff and they think that magnetics could help them should definitely talk to, to Richard and the folks at Nonfiction. For sure. Um, and thank you very much for, for, for listening. And you know, this is, uh, my first time, so. Open to any feedback from folks who listen to it. I didn't make you say the thing I wanted you to say.

Oh, what do you want me to say? I can—

could you put the headphones on actually?

Of course. Yeah.

What do you want me to say? Yeah. Okay. Uh, if I could just have you say one last thing, Richard, um, can you look at the camera and can you just say biology is done for the love of the game?

Uh, biology is done for the love of the game.

It is not done for the money. Well, okay. Okay. Okay. We're gonna, we're gonna, we're gonna close out there. Thank you very much for tuning in. Alright, cool.

The Core Memory Podcast is hosted by me, Ashlee Vance. It is produced by David Nicholson and me. Our theme song is by James Mercer and Jon Sortland, and the show is edited by the Jon Sortland. Thanks as always to to Brex, NE1 Ventures for making this possible. Please visit our Substack, YouTube, and podcast channels to get more of what Core Memory makes. Thanks, y'all.