Dr. Jason Kelly, thank you so much for joining me. I've been looking forward to this conversation a lot.
Yeah, I'm looking forward to it too, Tim. Thanks for having me.
Very, very cool. Now, so you are the CEO of Ginkgo Bioworks. This is a really, really cool biotech firm which is the leader in autonomous labs, which I think to some people sounds unattainable. To me, at first, when I first, you know, heard of this, like, autonomous lab, that's crazy. Having worked in a lab for over a decade, you know, day in and day out being a wet lab rat, it's hard to imagine an autonomous lab being something anybody would pay for. But when you stop and think about it, it makes total sense. And I think the best place to start would be to sort of paint a picture for people as to what it's like working in a bio lab. You know, one of the biggest constraints with biology is it's slow because you have to play within the rules, which is the rules of life. Things grow. You have to wait for cells to grow. You have to wait for mice and rats to live and die. Takes a long time. And one of the most fascinating things that you learn when you work in a lab is the concept of good hands. It's very strange that some things work in some people's hands and some things don't work in other people's hands. And I think that's an addressable market for your product, right?
I don't usually get to do interviews with people that are like practicing lab folks, you know, and so I think it was going to be really fun to talk to you about this, Tim, because you have the visceral experience of what it's like being... Yeah. Tell me about your experience working in the lab. You did undergrad, you did grad work at MIT. What was that like?
Yeah, I mean, so I got to MIT, I was sort of a product of Jurassic Park. So like I came into MIT really wanting to be a genetic engineer. This was back, I started college in 1999, dating myself here. And Jurassic Park I think came out when I was 13 years old.
I was just committed, you know, right? And there was no bioengineering major. And so I signed up for, I did biology and chemical engineering because I was like, well, chem, you know, cells are made out of chemistry. Turns out chemical engineering is mostly about oil and gas plants, but it's the closest I could get. And so I spent the summers working in labs at MIT. And I distinctly remember my freshman year, I was in, um, Herman Eisen, who's passed away now, who was in the cancer center at MIT. And I spent the summer trying to clone a single gene into E. coli under the tutelage of a very nice graduate student or postdoc in the lab at the time, and I failed.
Right. Like I was unable to move one gene into a bacterium.
Yeah. Let me just define that term, cloning. It simply means taking a piece of DNA out of one source and putting it into another source.
Yeah, 100%. And this is something, this was the birth of biotechnology was Genentech and Cohen and Boyer in 1978 to the early 80s out at UCSF and then at the company basically doing this, and with the first product being human insulin. So what they did was they took the gene for human insulin, which they take a human cell, they amplified that gene using PCR, which was also a recent invention, polymerase chain reaction, where you can make many copies of a particular gene. So that gives you enough of that material to work with it in the lab. Then you add restriction enzymes, which are little scissors basically, for cutting DNA, add those to the tube. They cut at a certain spot. You also cut the genome of bacteria at a particular spot and you paste in that human insulin gene just right behind a bacterial promoter. So the part of the bacterial genome that, one of the elements that says turn on the gene that you see next. And so you stick the human insulin gene behind that promoter. You grow that E. coli bacteria up and suddenly it's magically making human insulin and diabetic patients can get actual human insulin instead of back then you used to have to get insulin from the pancreases of pigs.
That's right. So my dad was actually a type 1 diabetic, like, you know, childhood diabetes. And so he at the time was taking pig insulin and got to get the Genentech insulin for the first time. Like, I remember it when I was like really young. And so that process, which by the time I went to school was 20 years old, I still couldn't, you know, like a novice could spend a summer, a smart kid really dedicated, and not be able to do the lab work. And that was demoralizing, I would say, for my interest in genetic engineering. And, you know, I stuck with it because I really wanted to do this. And by my junior year, I'd done two more summers in the lab, one more academic one at Genzyme. And I was just kind of like, this is hard and I don't know if this is going to work. And I was fortunate, I met a professor at MIT, this fellow Drew Endy, who's now at Stanford, and he was sort of pitching a vision that, look, DNA is code, and if we approach it with more of an engineering mindset, we can ultimately be able to do the sort of Jurassic Park dreams. And I was like, "All right, well, at least somebody wants it. There's at least somebody out here who believes it's possible, right?" So I joined Drew, and then kind of the rest is history. But I can talk more about it, but I had a really tough, I would say that experience I had means that most people don't stick it out.
Absolutely. So I think there's a lot more people who would love, like, they're enamored with biology. They see that DNA is code. They read about biotechnology and that we can read and write DNA code, and they're like, "Let me do it." And then they spend, you know, they take a class or they join a lab for a month or two in college, and they're like, "Wow, this is not for me."
Absolutely. I can totally sympathize. Your path is, we have a very similar path. I got into this for the same reason. I wanted to get into this lab because I had learned about molecular cloning just as you described, and I myself, I'm also type 1 diabetic. And so the story of insulin being cloned and, you know, saving, who knows, untold numbers of lives or making their lives better at least, that was very motivating. And I too, the first, it's funny that you mentioned your story because I too, when I got into the lab, my project was clone these genes that we were focusing on. And for me, I had poor success in the beginning, but I was fortunate, I had a pair of those good hands when it came to molecular cloning. I ended up cloning 150, 200 different genes, different versions of mutants and so on and so forth. And I just wanted to contextualize something. The insulin story that you just mentioned, why put it into bacteria? It's because we're hijacking bacteria to basically use them as a protein factory. You know, we can't use humans. We can't grow enough pigs. We can't grow enough dogs to make human insulin. So we can grow up huge thousand-liter vats of bacteria to make this insulin for us. So I totally sympathize with you.
It's very much, it's beer brewing.
It's exactly what it is, right?
Like it is a process extremely similar to brewing. So if you ever visit a brewery and you smell, you see the yeast and it's a big tank. I mean, you're feeding sugar to a microorganism or fungi that grows up in a big vat and it makes beautiful product.
Yeah. In fact, the same companies that build the conical fermenters for breweries build them for labs too. And funny that you mentioned that, that's how I got my love for beer. My side hobby was to start brewing, you know, beer science that you can drink rather than stuff that you can't. Right.
Yeah. That's very common among biologists. And it's also, people would find it interesting that a lot of the ideas that you come up with in the lab are usually over beer or drinks after the lab. Right. It's also... Yeah, it goes full circle. Yeah.
And so there's something really cool about cells is that you mentioned that they run on code, which is absolutely true. And something that I heard you mention in another interview is, and it's a great analogy, is that there are really only two technologies that run on code and that are cells, that's cells and computers. So tell us a little bit about your perspective on that because it's fascinating.
Yeah. I mean, listen, I'll give you the big dream on this stuff and then we could talk about what makes it so hard. So there's something magical about a substrate that can be changed by changing code inside of it. And we have all experienced this viscerally with computers where, you know, you've got your phone in your pocket and you go to the app store and you download a new application and suddenly this thing can order a car for you. And it couldn't the day before, right? Or suddenly this thing can take a human language query and just answer it, right? And it couldn't the day before. And that is because your ability with a coded object that runs on code is if you install new code, it can adopt a new capability. That's not true for non-coded things. Right? Like if your car is sitting there, you can't just like smash new code into it and change the color of the paint. Right? Because the paint is not coded. You can obviously make some changes now. You have computers inside cars with your dash system or things like that, but the physical thing is not changeable by a code. All right. Cells run on DNA code, right? It's A's, T's, C's, and G's, not zeros and ones. But there's just this unbelievable miracle of evolution that cells figured out the same thing we figured out when we invented digital code in computing and just invented digital code generally and went away from like analog, like record players, okay, or analog. We figured out that if you want to copy information, a great way to do it is encoding it in zeros and ones in something that is binary, right? Like that. And that digital means that I can transfer it to you and it's easy to like get the transfer right.
Yes. It's virtually lossless.
That's the, yeah. And evolution had a lot of time to think, I guess. But the way we pass information across generations, okay, the way you give your kids your eye color, your hair color, the way a bird passes the beak shape that worked out so well for it to its kids is through a digital medium of DNA. It's the same problem, like biology solved the problem the same way that we ultimately realized it was wise to solve the problem for transferring information, which is to transfer it digitally. Okay. And so that's just kind of a cool miracle of evolution. Sorry. Do you want to say something there?
Yeah, I was just going to say it's got one thing that computers don't have, generally speaking. It also has the ability to error correct, which is...
Well, I mean, yeah, so it is interesting, right? Like, so I think the way we've built our code is very frail to change. So we don't have the idea in software that the software code will change on us and thus the code needs to be tolerant to errors like that. There's an interesting discussion about where things are headed with neural nets, which are a very different type of software, like what is really at the heart of these AI models. It looks a lot more biological. Okay, it's produced through what they call reinforcement learning or, you know, these training exercises, which is really very much like evolution. It's like feed it things and select for the, you know, select the weights based on what does a good job, then kill the weights that do a bad job. You know, they're not programming. It's very fascinating. So we are seeing sort of software becoming more biological, but to date, what you said is true. It's not meant to be robust in that. So I think there is a lot to actually learn, probably embedded in how biology has been so robust to change that we could think about adopting in other disciplines. That's a super interesting scientific area. But in any case, it does run on code. All right. And so, but what's very special about cells, unlike computers, is that computers are coded. It's magic. You can give them new things to do, but everything a computer does is in the virtual world. So what computers move around with all that code is bits. They move information around. What cells move around is atoms. Okay? Cells build things. Cells manufacture things, right? And so when we can program a cell, when we can put new code in, we don't get a new application in the virtual world, we get a new application in the physical world.
That's very exciting, right? And this is kind of like viscerally obvious if you look at wild biology out in nature, right? Like imagine we didn't have, you've never seen a plant, and, you know, I don't know, who's a new CEO, let's say Tim Cook, but there's a new guy at Apple gets on stage at the Apple developer conference and he holds this, a seed, and he says, "Look what we have invented. We call this the seed. And you plant it, you put it in the dirt, okay, and you add air, water, and sunlight, and this thing begins to self-assemble a device out of the air, right? It starts taking carbon out of the air. It starts assembling that carbon into solar panels, in other words, we call them leaves, to collect sunlight. It uses that energy to bind even more carbon, and then it starts to build advanced nanotechnology, which we call fruit. But at a molecular level, they have more complexity than our current microprocessors that run our phones. Oh, by the way, it does all of that for free, right? Like, you know, and so like we, like that would be the most advanced alien manufacturing technology you could imagine. And it's happening outside right now in your garden. And the issue with it is we at the moment aren't able to design it and program it to do the things we want to do with the same level of capabilities we can program computers. And the core reason for that is humans invented computers. So we understand them all the way down to the last transistor. And humans did not invent cells. And so we are having to reverse engineer how the heck that seed works in the first place before we can design it to do more and more things. You know, like grow buildings, as you, right? So does that make sense? Like, and so that's the potential, right? Like the reason I've committed my whole life to biotechnology is eventually we are going to get to do all that stuff because it's latent in the substrate, right? Like we already know wild biology can do absolutely magical things. It can self-assemble. It can self-repair. We have healing. We have molecular structures that are absurdly complex on a three-dimensional basis. And we have very low cost to do all that. Yeah. Yeah. It's like... it's just, let's have it do things that are more in line with what humans want, you know, and that's it. Like that's the game, that's the game of biotechnology, and it'll go on for the next hundred years. And the use cases, I mean, are enormous. The, I mean, from health to materials to buildings in the background, whatever it might be. We, I mean, there's so many that we don't even know yet.
And I've heard you use another great analogy, like you just mentioned, computers, we invented, we know them down to the last resistor on a sub-nano level. Biology is like an alien technology that was dropped down in front of us and now we have to reverse engineer. And we're in that process right now. And if you had to put a number on it, and I know this is difficult, how far along do you think we are on the path of mastering that alien technology, understanding it?
If it was like a baseball game, it is the first batter and it's like, I don't know if there's been six pitches, you know, like it's pathetic, but we're at the absolute infancy really. Like, and that I think is the thing that, like I'm glad I didn't know when I started biotech, you know, as a student.
Yeah, it would be demotivating for sure.
It's just, well, you know, I think now looking back 20 years into it, I'm glad I'm where I am, you know, because there's just like so much cool stuff still to build, right? Like we're so, you know, it's kind of like computers in the 50s or something, you know? Right. I'd agree. It's very like the true believers are here for it and it's going to take some time, but you can kind of see the path and there's a lot of wind in the sails right now thanks to things like robotics and AI that are becoming and being applied increasingly in bioengineering. That got me pretty excited. Like, you know, like I don't think we're stagnant on ideas for what could create the next unlock in our ability to design biology. And I can get my framework for how to think about that for, you know, students that are excited to get into the field or folks that are building companies and startups. I have my sort of like, okay, what have I learned the last 20 years about what works and doesn't, but it's absolutely early days, so no one's missed anything. If you're into this stuff, you could still be the seminal contributor to this field for sure. It's really exciting.
Yeah, I appreciate you saying that because that is true. There's the opportunity. So just to contextualize, if we had to look at the last 20 years from when you and I were in grad school to now, really, I mean, there's a lot of talk about, there's always been talk about personalized medicine and things are moving so fast. If you're on Twitter or X, you see these sort of breaking news, this has been solved in biology, but really not much has changed. What you and I were doing in the lab 20 years ago is still being done almost the same way. Moving genes around, for example. What has advanced is ironically the dry work. You know, when you and I were cloning genes, we'd do it by hand. We'd print out a line of DNA. We'd look for restriction enzyme sites and we'd think about it. It was a big Sudoku puzzle. Now you have AI that just does it for you. So that does definitely increase the rate at which research can work, but really not much has changed. And you use the word unlock. I think it's coming, especially with AI and iterating, which is what Ginkgo Bioworks is going to allow labs to do. So why don't we talk about Ginkgo? How did you start it? Now the motivation for it, we just heard you talk about it. Tell us a little bit about the Y Combinator experience and how you got to where you are now in terms of Ginkgo.
Yeah, I mean, so there's five of us that started the company. Four of us were grad students together at MIT, and then the fifth founder was a MIT professor, Tom Knight, who was like two of the students' advisor. Okay. Austin. And then Barry and myself, we both worked for the same advisor too. It was fellow Drew Endy. And we all were graduating in 2008. And it was just Drew was moving to Stanford and everybody kind of finished at the same time. So we're all basically asking the same question of like, should we apply to be professor roles? We had talked about doing a startup. We had been doing kind of like a few things during grad school, like pitch competitions and things like that. And we decided that, yeah, it felt like the next step for what we were doing at the time and what we were really, we were some of the first grad students in a new field called synthetic biology, which was really what Drew and Tom kind of helped originate at MIT. And there was a number of other schools and an NSF engineering research center around this area, that sort of this idea that you could think of DNA as like reusable genetic elements, like parts, like you might have a standard promoter and a standard chassis strain. And it was like, kind of could you standardize the actual biology? That's actually been much harder to do than we hoped, like parts bins in a factory. Bingo. Yeah. Like could you think of like the way you build the genetic program as taking known reliable parts off a shelf? And I would say the answer to that has been like not really, not that great. And we'll come back to, you know, what's worked and hasn't, but that's my takeaway after 20 years. But that was very much, we were those first students. So one thing I would say is as a grad student, like lean in on areas I think that are new. If you can be some of the, you don't have to be the first professor in that area, but if you're the first students, like all of our peers from that, most of them are, you know, professors or department heads now in sort of like synthetic biology or bioengineering departments around the US where we were. And so I think that's like just a good lane to get in as a kind of like new field. It's like worth the risk. Yeah.
And then secondly, what we did was we were like, all right, well, we felt our group that you really needed to apply engineering effort, especially on the lab side, if you were going to make any progress here. And so we wanted to be able to invest in robotics and some of these things, and it felt like that was going to work out a lot better in an industrial setting than an academic setting. And so we decided, okay, we'll try to start a company. And we made, you know, you can actually do that pretty cheaply. You go, there's lawyers that'll work for free for you because they want your future business and you're starting a company. So you get a free lawyer and they'll, you'll say you'll pay them back when you do the funding round and they help you incorporate. You get, you know, so we had a, you set up as a C corp. And then once you have the, once you have the organization, it actually opens a lot of doors for you because you can now think about raising capital. So we went out and talked to a bunch of VCs. That went nowhere. Horrible experience. A, it was 2008. So it's like nobody was excited to give out money. B, we were not making a drug. So the thing you'll learn quickly in biotechnology is that the whole venture ecosystem orbits around drug discovery. Okay? Because drugs are proven to make a lot of money and you can kind of get to a certain point where like, you know, with a team of say 30 or 50 scientists, you can get to a candidate and then you can do a trial and kind of if you're all in for say 50 to 70 million bucks, maybe across that range, you could theoretically have an asset that is worth two billion. And so that's like decent venture math. Not a VC can say, all right, you know, I like this, like what they're going after and the payoff is there. We were not doing that. We were like, oh, we're going to make robots and platform technology. It's much more of like a tech industry business model. And that's not what the biotech VCs are used to seeing. Interesting. Okay. And so that wasn't good. That wasn't fundable. And then we were children, right? Like we're like straight out of school, young people. And again, the tech industry celebrates that, you know, the Zuckerbergs and the Gates and people like that. But in biotech, we haven't had as much of it. And so that's also looked at pretty skeptically. And so what we actually did, and this is advice I would also give like students that be tuning in, is we applied for grants. So we applied for, here in the US, we have like SBIR program, which we applied for a bunch of those. We applied for DARPA grants. We even applied, we had a grant from Scotland. Like we just kind of used our skills as graduate students knowing how to produce a high quality scientific proposal. And we did a lot of work trying to do that. And we took our time. We bought a lot of equipment on eBay. We got the cheapest lab space we could find in Boston and basically bootstrapped for four years before we were able to raise any capital. So one thing I would say is I wouldn't be afraid to do that as a student, as a grad student, especially because A, you've already made a decision to get into some niche technology area. A lot of your peers probably just like went to consulting or finance or tech or somewhere to go make some money and you are already going down a road where you're betting on some technology area. That's a real skill set, real capability you have. And then if you are willing to stand this organization up, you can apply for grants. It doesn't look that different than starting a new lab, right? Right. Like as you're aware, right? A lot of what a PI is doing is applying for grants. So yeah, exactly. And so you kind of like are building that skill set anyway instead of doing a postdoc, right? Like doing a postdoc is sort of like, you know, purgatory, right? And like what's the... low paid, low paid purgatory. Yeah, I think so. Yeah. And so we kind of felt like it was a good way and we just sort of had a deal that like, look, if it felt like things were going okay, then we keep doing it. And if not, like, listen, like we're all friends, like we can kind of dissolve this thing and call it a day. And it kept getting better every year. Basically the technology was clearly getting better. We were also doing commercial deals like service agreements with, you know, companies to kind of get some business. And then in 2014, you mentioned Y Combinator, Sam Altman, now Mr. Famous for OpenAI. He takes over Y Combinator and writes a blog post saying, I think YC should go after hard technology, deep tech. So things like nuclear fusion, materials, biotech. And so I wrote him an email. I was like, hey, listen, we're four years old. We're 15 people in a lab in Boston. We don't really make sense for YC, but this is like oasis in the desert. Right, right, right. Thanks for funding. You know, just saying that like nobody funds this stuff and that YC could fund this stuff and the Silicon Valley model could work for these kinds of companies. And by the way, he was right. Like, so we were the first biotech to do YC. Like I went, Sam was like, oh, you should definitely do it. Like come out and meet me. And so I met him and we decided to do it. And then Helion Energy, for example, was our batch nuclear fusion company that's raised like $5 billion now or something, right? Like that thesis that deep tech can be treated like we fund tech companies is true. And so we ended up going down that lane, basically got treated like a tech company and got raised money in Silicon Valley and New York and really never raised money from traditional biotech VCs. And then we did the company public in 2021.
Yeah, I think that's an awesome story. And you're right when people should do it. Or grad student, like I think it's not, we still don't do a good job telling people how to do it. There's sort of like a lane that looks like talk to your TLO and take your IP and call a biotech VC. And if you do that, like, no, you're not going to get the time of day. The TLO is going to take a million years. The biotech VC is going to tell you to pound sand. And if you're the luckiest of lucky, you'll get to be the founding scientist at a company run by a biotech VC associate and get...
100%. Yeah. It's nothing like the kind of model your tech peers are doing where they, you know, when we started Ginkgo, there's five of us. We each own 20%. Right? Like, you know, like the attitude was like we own, you know, like you're the founder, you own the right, not you're like an employee of a vendor. You know, you got some pride and ownership. Yeah. So I, yeah. It's just like the, I would just encourage, like I wouldn't, the opportunities are so poor if you're not doing a drug. The venture market is often so bad for you that you might as well just go at it yourself. Right. Like go get the grants, go make it happen. And then once you get a little bit of traction, actually venture is, you know, look for your lane, right? Like nowadays there is actually a lot more investing in deep tech because of places like Anduril, these all the defense tech stuff's done so well and SpaceX, you know, like credit Elon, like all this, like, you know, SpaceX and Tesla have proven you can make a lot of money in atoms. So more, there's now a lot more Silicon Valley venture money that's not just going to software. Now you can kind of get money for atoms businesses, which all those of us who do PhDs are often in the world of atoms. So anyway, I would really encourage people listening to try. Like I don't think it's, I think that path should be taken way more.
I 100% agree. Now, you know, some universities and colleges are doing a better job. A lot of them have offices now that encourage spin-offs. That's great, but you're right. It's when you're a graduate student and you're done or you're still a graduate student, I mean, you're either going to be in low-paid purgatory as a postdoc or low-paid purgatory in a startup. You might as well take the higher risk, higher reward. Yeah. And work for yourself.
Yeah. Why not? What's the difference? Like, by the way, like if you do that for two or three years, like it is such a great learning experience. Oh, if you... Right, like it's in the same way grad school, like everyone, like I think people poo, like I think classically Silicon Valley like poos PhDs and I'm like, guys, like where do you, you know, it's almost like, anyway, what I'll say is for those of us who have done PhDs, you have this awareness that a big part of how you learn is that it's your thing.
Right. Like it's like your thesis.
That's right. Yeah. Yeah. A lot of students ask... You're not working for someone else.
That's right. Like, and some PIs are better than other PIs about this. Like I came from a lab where Drew was great about this. It was, you know, like it was really your thing. It wasn't like you were, you know. But at the end of the day, it's your thesis regardless of whose lab you're in. And so there is a learning exercise that is just the product of doing your own work, right? Like, and that is very much like what founding a startup is like.
It's, there's no one to help you. Like, you know, like it's your...
thing, right? And that is awesome. And so you've already gotten some training to do that as a PhD. If it's your kind of thing, a startup's a great place to keep trying it. Just go.
Yeah. Yeah. PhD is like a furnace that strengthens iron, right? And students ask me all the time, when you're working as a graduate student, how do you decide what you want to do day-to-day? Who tells you what to do? Nobody. You've got a project, a question, a goal. And that's sort of the learning that you just...
people go through their whole life without learning what it's like to be solely responsible for something in that way.
It's really an unbelievable learning exercise. And you do get it at founding a startup too. That's why I kind of think that community has more overlap than people would imagine. And it's not everybody. I'm not saying that there's plenty of people that don't get that, don't really squeeze that out of their experience, out of their PhD, but a lot of people do and it's a unique thing.
So let's talk about autonomous labs now. Why autonomous labs? What do they look like in practice for those who are working in labs right now and who hear the idea of an autonomous lab, how is that even possible? You know, I pipette, you know, a thousand times a day and I think I'm quite accurate. Studies show that usually you're not. So what does an autonomous lab look like and then I'll ask you what can it run, what can it still or what can it not run and who is the ideal customer?
Okay, so first off, let's talk about why when people who work in labs think about bringing robotics to labs, they're like, oh no, this is super hard. And also why we don't see robots in labs, right? So you mentioned our tools are very similar to where they were 20 years ago. That's very true, right? I took one of my good buddies who is a professor at MIT now in the biology department and I brought my son who was like nine. It was a couple years ago at the time and I was like, oh, let's go visit Joey and I'll show you. And I realized while I was there, I was like, oh, my old lab bench is like the floor upstairs in the MIT biology building. I was like, let me go show you where you... and we go in there, Tim. And it was...
like going into a time tunnel, right? Like it's...
you know, same heat bath, right? Like same kit above the bench, same sad grad student, right? And I was just like, oh my god, it's the exact same tooling I used at this exact same lab 20 years ago.
Okay. And if you walked into like the computer science department at MIT or the mechanical engineering department at MIT, even the tools available to the graduate students would be wildly different.
But not the biology grad.
That is so true. Okay. And so why is that, right? Okay. And so here's my analogy because we call this an autonomous lab. So I'll use the transportation industry as an analogy to explain what has made this problem hard. Okay. So imagine you have a chart and on the y-axis you have the level of automation of a transportation technology.
And on the x-axis you have the variability in the request from a user.
All right. So, low variability, high amount of automation. That's a subway.
Okay. You sit on a subway. I swear to God, Tim, it is totally automated transportation. You don't have to do anything.
Okay. Takes you away. You're doing nothing. Okay.
But you better want to go to one of the stops on the subway line. It is not going to pull up in front of you.
Yep. All right. Now, low amount of automation, high amount of variability. We have that technology, too. It's called a car. You put your hands on the wheel, foot on the pedals, you manually operate this bad boy. And it'll go right to your house or it'll go to the grocery store, wherever you want. Extremely flexible in terms of where it'll go.
All right. And that's the transportation system for a hundred years we've had.
Have you been in a Waymo?
No, I haven't. I've experienced FSD, but yeah, not a Waymo, though.
So, the exciting thing about the Waymo is that you're in the back seat.
Okay. All right. So the Waymo is up here in this corner. It is the automation of a subway. You sit down and do nothing.
But it is the flexibility of a car. You can tell it to go right to my house.
All right. And that's so surprising that we give it a new word. We don't call it an automated car because like automated door, right? Automation just does the same thing over and over again.
This is an autonomous car. Not because, by the way, it's deciding where to go. You're still telling it where to go. It's because, what the hell, how are we getting this amount of variability automated, right? Like we've never...
That's the barrier. Yeah. Well said. That's the barrier. The barrier is the variability. And this is true across lots of areas of physical areas where things were the same. We automated, going back to the loom.
Okay. Right. Like we automated the stuff that was repetitive and the same. Printing press and the...
Bingo. And then as in every industry, once you had a certain variability barrier, the humans came back.
Like making shoes are too complicated. So there's still people involved in that, but t-shirts, right? You can print those out. Okay. Right. Does that make sense? All right. So let's go in the lab. Low amount of variability, high amount of automation. We actually do have these. If you've worked at a pharma company and seen the high throughput screening groups, they have things called work cells. It's a robotic arm and it moves plates like 96-well or 384-well plates or 1536 through a bunch of devices over and over again. And it's amazing. It works overnight. It's totally automated. No pipetting. It's all just doing its thing. All right. However, you better want to do the same experiment you did yesterday because you cannot make this thing do something different tomorrow. It's like a subway. It's doing the same experiments over and over again.
All right. Low amount of automation, high amount of variability. The car, oh, we got that, too. We've been talking about it. It is the manual lab bench.
And all the manual benchtop equipment. Not just your station with the pipettes, but all the other devices around your lab that you walk up to.
Yeah. You know, like, oh, you walk up to the centrifuge, could you share it with your lab mates? You go use the HPLC. All of that paradigm.
