How AI Will Accelerate Breakthroughs in Biotechnology with Benchling CEO Sajith Wickramasekara

23:45So I think while it's early on sort of the agentic how we work side, I think you're going to see very unique models come out of pharma where their computational scientists are building interesting predictive models that, you know, similar to what's happening in the open source world. What do you like? Can you help characterize what useful models we already have on the discovery side? And then like where you think we are in the cycle of having, you know, enough to make a real change in the overall cycle time or rate of success, whatever you think is more important. here. Yeah. It's been really cool to see like the whole ecosystem of these tools grow a ton.

24:25Like open source wasn't really a thing in biology before or in science before. And like, I think in the last two years, it feels like a thing that's here to stay. Interesting. All these interesting models come out like the bolts, for example, on the structure prediction side. And I think like that's like a really interesting thing that's going to change, change biology. I think you're seeing new approaches to like federated learning as well. Eli Lilly put out this announcement about a project they have called TuneLab where they're taking their internal models and making them available to the broader scientific ecosystem.

25:01So a farmer company is saying, hey, you can use our models, but it's give to get. So we get to train and like it's federated and whatnot. So they don't get to see the actual scientific data. But I think like those approaches are like this is the beginning, basically. And so we've got some really cool stuff. There are problems that I think are fairly tractable in terms of structure prediction or antibody developability and so forth. So like that, that's really good. But there's, I would say, a lot of work in front of us in terms of models that are sort of more predictive of what happens when you get into patients, for example.

25:32and don't get me wrong, like discovery is really important, but there's so many other steps that have to happen after you have a concept molecule. Even if you're much faster, for example, at making molecules and you have a higher success rate, you still have to come up with a process to manufacture that molecule. And now you have even less time to do so if it's like a byproduct of success. And so how do you optimize manufacturing processes to get, you know, a lot of yield, speed cost out of out of a molecule is that like where you would say the um highest value missing predictive model opportunity is because i think a bunch of naysayers would be like okay yes i've heard about alpha fold and like people working on antibody prediction yeah creation platforms but we are still you know many years into this premise no drugs out the other end of the pipeline that are ai discovered yeah i feel like the naysayers kind of have this sort of worldview in mind where it's like, oh, I just like type a disease and then I like get a molecule out and like amazing.

26:32Yeah, I discovered drugs. And then this is where I go back to like my mental model is like there are so many steps. Those steps are all cumbersome and difficult. And this is a game of like making each single thing better. And like some of the steps matter more than others, like having the right target or having like a great molecule generated fine. But like there's still many, many years after that that we can compress and shave off. And so right now, I would sort of almost argue that we should be thinking about like, what's the share of experiments that have been touched by some kind of predictive capability, some kind of simulation or some kind of AI.

27:02And I bet that share is like getting higher every day. Part of what I think has been really interesting and there's like good and bad about the investor enthusiasm of, you know, both, let's say, AI's potential impact on biotech and then the potential for platform companies is this theory that we're going to have like very different business models in biotech. Do you think that's going to happen? I would like it to happen. I think right now, for companies, I mean, so one, as a toolmaker, I think there should be more tools. Tools are good. I do think with some of these model companies in the bio world, there's gonna be an interesting question of do they morph into in the fullness of time morph into their own therapeutics companies with their own pipelines?

27:43I think it's, it's unlikely, it's possible, but it's unlikely that sort of, hey, they're just going to remain pure model companies who just do deals with pharma, where pharma, you know, pays them$100 million up front or something like that, and they have five customers and whatnot. Like, I feel like the sort of model building is probably commoditizing too fast for that to be a trackable business model. But to take that expertise and to be fundamentally better at doing research and early development to make molecules and sort of morphing into a biopharma company, like that seems like one logical path.

28:14I think there's a world where like, and we're experimenting in this space where sort of models can be more effectively distributed to the larger biopharma communities. So rather than going and doing BD deals with five companies, it's actually a little bit more like kind of a traditional software sale where, you know, we've actually got a bunch of models in BenchLink. They're mostly open source, but also we've got, you know, we've got Chive and AlphaFold, things like that. Is there a model where like some of these are like pay-per-use or like fee-for-service, almost like SaaS and the entire biotech companies benefiting that?

28:43And you can build models and like have a scalable business model on the other end? Like, I think that'd be really interesting. And then there's going to be like more data transactions, I think, as well. like data wasn't is not is interesting for for a field that really depends on data as its currency like everything is about data on the molecule you see very very few transactions of data that's because no one trusts anyone else's data you wait till there's a clinical trial and the data is positive and you buy the molecule but you'd think that you'd see a lot more selling of of data before that but you don't because the data like it's very hard you don't know what format it's in do you trust the way it was created like it's just not there's tooling and normalization about it You might be able to transact on it.

29:20Yeah. Okay. And like, will people be selling like their negative data at some point into some pool that other people can learn from that? There's all kinds of crazy stuff I can think of. I'm sure you've heard in the 13 years you've been building Benchling, the conventional wisdom is that the only way to create value in pharma is assets, not tools, right? Like where were they wrong? Or maybe the tools just like weren't that important before and they weren't as embedded as they needed to be. Yeah. Yeah, I don't know if this is a... Thermo Fisher and Danaher are like sneaky big companies. And I think people don't always realize that.

29:53And they've done it largely by like systematization of tools in the physical realm. So like instruments, reagents, services around them and so forth. So I think there's some kind of at least thing that rhymes with building great tools on the digital side. I think just frankly, like looking back, the technology probably hasn't been there. Like when we started BenchLink in 2012, cloud was like the norm everywhere. but most of life science was like paper on premise spreadsheets like okay so like we spent the first couple years like it's wild for such an advanced field in other areas yeah i mean and that is because like you you could argue like hey we're we're it's such a like high stakes game of poker for them that the only thing that matters is like does this drug get to get to get to patients and is successful and they can you know pharma has pretty healthy margins and so like the operational efficiency isn't always going to like improve the odds of success.

30:47So we spent the first couple years basically just evangelizing, like bring science online, like it's going to be better. And then spent the next like 10 years after that kind of convincing people that structured data mattered. And because that's that's sort of like the core premise of Benchley. It's a system of record will help you have like a data model. And every time you do experiments, that data model is getting populated with information. You can ask questions. And there's a set of people who they got it and they believed. It seems obvious to tech people. It seems obvious, but there's like, it's not for free.

31:16Like a piece of paper is much easier and an Excel spreadsheet is much easier. But there's a set of people who believe and a set of people who maybe weren't convinced. But now with like AI, one, I think the benefits are a much more immediately obvious to everyone. And so that's going to be this amazing tailwind to try to like do better here. And I think it will convince a lot of people who might've been skeptics in the past. Yes. I don't come at that from a holier than that view because one might actually claim that in venture investing, the only thing that matters is the quality of the next decision and whether or not you found the winner.

31:50And so it's a lot of tech people with a lot of pen and paper, actually. Yeah. Yeah. And so, but I think that's likely to change. Two things. One is like all the foundation model companies, DeepMind, Anthropic, OpenAI, they love to talk about AI for drug discovery. Like and, you know, I think that there is fundamentally like a mission orientation there. I also think I'm a bit of a cynic because it's hard to be like, that's a bad idea. Like that seems like just just roundly good for humanity if we have more medicines, as you said. It's like 10 years ago when the crypto people were like, it's all international remittances.

32:29Right. Like, why do you think it is like both so popular with the labs and then even more popular the last few months? And then like tell us about your partnership with Anthropic. I go back to that sort of returns to intelligence piece where I think science is a is a problem that has some shape to it that really benefits from the LLM architecture. Like you just think about the corpus of scientific literature as like this vast pool of unstructured text. And like these are kind of these are like roles where these are pursuits where like that's a ton of domain knowledge to hold in your head. And there's so much specialization.

33:04And so the idea that like you could be like truly standing on the shoulders of giants, I think is very appealing. I've got a scientist and I'm in an early stage biotech and now I can have access to the world's best like clinical design expert or the world's best toxicologist or the best research assistant who can even like read papers better than me to figure things out. Like there's a lot about science that, again, is so artisanal and inefficient that it seems like a problem that AI is going to be much better at. I think that's one thing. I think the other is like, I mean, it is like I think bio is I don't know.

33:35Like I sort of wonder, like, why don't more people work work in bio? Like there are big problems to be solved. There's an incredible like. Because the failure rate's so bad. The failure rate's so bad. And like the impact is huge. I think everyone's now seen what like GLP ones can do. Everyone saw what COVID vaccines can do. Magic. Yeah. Yeah, it's like when it works, it's magic and like people need this stuff. And so like, I don't know if AGI starts, you know, automating away software engineers or whatnot. Like what's left? Like got to make drugs for people. All right. More scientists. And what about the partnership?

34:04Yeah, so we have a partnership with Anthropic. I think we feel very we work with we use sort of all the foundation model labs capabilities. But like we found that there's like a strong commitment to science from Anthropic. I mean, Dario's a scientist, and so it's been really, really good mission alignment with them. And I think sort of they've expressed publicly that science is sort of the next frontier after after code. And I think for our customers, trust is super, super important. And I think their posture plus their technology is one that really appeals to them. And so it's one where just to the start, like Benchling and Claude kind of like natively interoperate very well.

34:38So if you want to, you know, work through Benchling and Claude or you want to, it works pretty effectively. So scientists can, you know, generate reports, ask questions and things like that from a very simple AI interface that they're used to. And I think it's just like the start with them. Can we talk a little bit about company building? Sure. Just, you know, 13 years of wisdom and like two minute takes. Every mistake made at this point. Maybe we'll start with the most recent, like hard decisions, not mistakes. But your co-founder, Ashu, gave up all his direct reports at some point and went all in on it.

35:10I called a bunch of friends around this company and our mutual friends. That's a big decision. Like when that happened, how did you make the decision? Because you guys started way before AI was working at scale. Yeah, it's funny. I think we started early, but at the same time, I feel late still. Okay, all of this, yeah. I yeah you know one of the interesting things is like I feel like the power of being a co-founder is actually just in moral authority and I think like this was a it was a pretty controversial decision in our company and we needed someone who had the right I guess willingness to like a like I don't need any of my Legos anymore anyone else can have them and like b if I look stupid that's okay yeah and so like it's funny he wrote this wrote this post of like I'm quitting my job to like do this other thing.

35:57And I had a bunch of customers call me and they were like, oh, my God, I'm so sorry. Your co-founder quit. Is everything OK? I was like, oh, no, no, no. Metaphorically. Metaphorically. Didn't quit. Just going full time on AI. Yeah. And it was controversial. Like biotech's been obviously like had this kind of dot com level bust. And so a lot of our, you know, our team is feeling like, hey, we've got to focus on like the basics with our customers. The market's tough right now. You have some companies that are laying people off, shutting down. Like, Yeah, how do you invest like that? Yeah, isn't AI like a distraction?

36:26But like we were actually really fortunate. I think we had good sort of outside the building perspective. Asha was very hands-on keyboard himself. And that's how he got convinced. I think he was like playing with one of the models like during Christmas or something like that, building for himself. And I think that really, really inspired us. And also like we realized like if we don't do this for our customers, like who is going to do it? Like, again, it goes back to needing to translate some of these amazing things that come out of the Bay Area and Silicon Valley into like useful vertical applications in very complex regulated domains that we felt like we're the right people for.

37:00Maybe because there are a bunch of entrepreneurs listening to this podcast as well who are looking at industries that are complex and regulated and want to try to bring, I don't know, the cloud and then AI to them. Like, what has been hardest? What are some lessons from that? My I think like the my like most trusted algorithm for this is like go talk to customers and all that I know super obvious, right? But all the times in which like I feel like the company has been like at its lowest or at its worst or like I'm feeling at my lowest. It's because I've gotten too far from customers and like product market fit is this moving target.

37:34And sometimes your market changes like market changed of like a bunch of biotech companies that were going to take over the world and then like then they weren't. And so like I still spend and maybe this is a vertical thing. I probably spend 30, 40, 50 percent of my time talking to customers still. And I think of myself as like the I need to go really deeply understand their problems and how they're changing and I need to go bring it back to the company. And like that's the thing I have to role model so the whole company does it. And that's the number one piece of advice I give people. So that answers part of one question from another friend who used to work for you.

38:07He said that Saji is amazing at like understanding the macro of the company and then like being deep in the detail on everything, especially with customers. But given, you know, BenchLink is a complicated company, complicated fields, like how do you decide what to focus on and then like focus your team on? That's a nice compliment, though. It's like telling someone they have a large context window. It is. Good model, man. Thank you. Thank you to whoever said that kind thing. It was Malay. Hi, Malay. Yeah. I, you know, one of the most interesting things about being in a vertical is you talk to all your customers and they all, because no one, they're underserved.

38:45Like they're just, you know, the world of life science software, there aren't, it's not like go-to-market tools where there's like 10 ,000 companies. Much less great software companies. Yeah, there are just not many. And so you have a very underserved demographic where they're not used to someone coming in sort of asking what they specifically need. They're used to very general purpose horizontal software and trying to like do koogee things to make it work. And so when you go talk to them, the first reaction is like, oh, my God, what we do is so unique and we're such a snowflake and there's no way.

39:13But then you like talk to enough customers and it turns out they almost all want the same thing. I find that having five, ten customers is actually like a pretty representative model for what the entire industry needs. You don't want to get like too overweighted in like one category of medicine or one size or something. But everything we've built over time, I thought has been successful. It's because we found a couple of customers and we got very, very deep with them. And we built and built until they were like super happy. And then it took off. And like maybe that's like the YC part that's been like programmed into me that I've never gotten out of it, like do things that don't scale.

39:46I feel like in a vertical that works 10x, 10x better. Even for our new AI tools, we have all of our customers. They're available to all of our customers now. But where our team is on like daily, weekly calls with five or 10 customers to the point of like we had a customer yesterday, you know, text us about like a meeting they had with the FDA and how they like ran a report inside of our deep research tool. Instead of taking them a day, it took them like five minutes and it was like perfect for them. And like that's the level of closeness we get with our customers. One other unique thing that I think would be useful for a lot of people today, including me, is Benchling knows how to get scientists to work in a software company and work around a software company.

40:28Like I work with many more research scientists in different fields than, well, actually, some in bio, but then I anticipated, let's say five years ago when I was just like good old engineering. And it is philosophically different, right? You have to run programs differently. you're like, well, you know, it's not like this is done by the next sprint. It is we do not know. So like what advice do you have on like recognizing that talent, getting them to be productive, managing it? That is a really hard question. I have I have serious battle scars of that. We've had to build a very interdisciplinary company to be successful.

41:01If I was only hiring like software people who knew bio, I would have like exhausted the pool 10 years ago. It just doesn't exist. We have to take the software people, take the science people, make them sit together, learn from each other. I actually, and this is going to sound obvious, but I've actually found that like the most conflict has been around the like actually like how the mission gets solved where actually a lot of like coming from the world of science, especially academia is just a very different incentive structure. And in the world of academia, like your labor is basically free. And so like, it was like very, very, very cheap, right?

41:37Grad student labor. And like the currency is publishing a paper. And that's how you get more funding to do more things and so forth. Whereas in a company, like we have to sell software. And so like, actually the most impactful thing has been like really a lot of repetition that in order for us to achieve our mission and to keep delivering great things to our customers, like we have to make money. And a lot of the tension has come from like that, the need to do that. And so making sure our scientific teams like really understand that the better we do as a business, the more amazing innovation we can bring to our customers.

42:10And by the way, if we don't do this, who else is going to do it? Where are the other like next 10 companies building software for science and R &D? They're going to power the next discoveries of these biotech and pharma companies like where we're like, I think we're the only independent, like scaled player doing this at this point. Do you interview at all for this orientation? I think we try to. I don't know that we've like I have some amazing predictive way to find it. I have a friend who's a founder who asked what I don't find to be a controversial question because it's literally in my title.

42:43I'm a venture capitalist. But they ask people, including research scientists, how do you feel about capitalism? And they think it's a pretty interesting sorting function. I will try that. I'm not recommending it, but I do think it is interesting. the set of answers you get, actually, because I have tried it. One more question for you on culture. You spend a lot of time with biotech and large pharma. You're at your core a software person. What can these two worlds learn from each other? Great, great question. I think the biotech and pharma world can learn something from how tech communicates and tells stories.

43:19There's the whole go direct wave in tech right now that really resonates with me as a founder. I think biotech pharma companies need to tell their stories. Like most people. This is not what I thought you were going to say. I'm curious what you thought I was going to say. Most most people I don't think could name five scientists or five CEOs of pharma companies, but like we could name every single like tech CEO. Everyone knows who you say Sam and it's like first name basis, right? For a lot of them and or Jensen or something like that. There's a lot of heroes journey. A lot of heroes journey. Yeah.

43:50But like and I think talking about the patients and scientists, though, would like change a lot of the public perception of biotech and science. I don't think people know how hard it is to like make a medicine. I mean, I was, you know, I was just thinking recently, like, you look at COVID even, like, Moderna and Pfizer helped the world, whatever you feel about vaccines, like, they played an important part in like helping the world like reopen. And yet, like, Zoom gets more credit in COVID. Or you have Gilead, who's like, you know, HIV used to be a death sentence in the 1980s. And like, there was tons of panic and fear about it in the 90s and early 2000s and like it's like cured HIV basically at this point.

44:27And like most people have no idea. Like and so I think like because sort of the the way they communicate is much more about almost these like faceless companies rather than the people. I think it's like easy to easy to hate on them and underappreciate. So I think they need to tell their story and go direct. Yeah. That's one thing I think on what tech can learn from from bio is I think once you start I think sort of tech has sort of become everything and sort of gone from, you know, very, so the ambitions of tech have grown a ton. And I do think like some of these other industries have figured something out when it comes to rigor and validity and accuracy and sort of move fast and break things does work for certain domains.

45:06But again, once you get to the point where you want to have credibility with regulators to put things in patients, you know, or make a medicine, like that stuff does come to matter. And like biopharma, again, for all the things that are difficult with this is figure out how to be safe. Like the U.S. is still the gold standard for how to like deliver medicines safely to people. I think that stuff matters more and more. Yeah, it's really interesting. I was talking to a friend who is a top research scientist at one of the top labs. And I asked him a very basic question. probably a year and a half ago about like, well, you know, how can I how can I better use these models in very quality sensitive fields?

45:45And he said, just wait. Right. Like there's like, you know, like, why don't you just focus on the use cases that are not as rigorous, essentially. And this is an amazing research scientist, but I'm like, that seems unambitious. Right. Because there are many benefits to intelligence in these fields that are really important to all of us. And so I do feel like more people should work with that, too. Yeah, yeah, yeah. I think in Silicon Valley, sometimes it's a little bit easy to, like, oversimplify other people's jobs. There's a lot of complexity out there in the world of biotech and pharma. But I'm very optimistic that, like, the sort of tech companies will figure out how to make it work because I think everyone's very excited about AI.

46:20Last question for you. Something you're excited about in AI outside of bio or outside of, like, Benchling's immediate purview, I guess. On a personal level, I haven't written a line of code. I had not written a line of code and probably I'm very embarrassed to say this, maybe eight or nine. It's been. Oh, wow. It's been some time. Like, I remember I stopped coding around the time that React started being a thing. Like, I'm dating. I'm dating. I'm dating. Yeah. All the kids listening are going to turn off. And like, I've tried some of the new agentic coding tools lately. And it's just like on a personal level, fun to just feel like the whimsy of being able to build something very quickly again.

46:57It's like that. That's pretty cool. That's probably the one I'm most excited about. And then I'm excited for my like parents, actually, and for them to have like new technology that's pretty easy for them to access. Like my mom's a chat GPT user and I'm sure it's sort of Google search plus plus at this point. But like that's been like pretty cool, too, of like technology that's so intuitive that like I don't have to go home and be IT support at Christmas. Yeah, I think it's actually undervalued as just because there are audiences that are traditionally not as lucrative as like the fast tech adopting audience of your 15 to 40 year old.

47:33Yeah. But it is incredibly wild how the UX of like natural language and voice are changing who can use it. Absolutely. AI tutoring for my kids and such. Oh my God. Same experience. Learning is such a joy with AI. Help them learn the right things, but like. Thanks so much for doing this, Saji. Thanks, Sarah.

47:55Find us on Twitter at NoPriorsPod. Subscribe to our YouTube channel if you want to see our faces. Follow the show on Apple Podcasts, Spotify, or wherever you listen. That way you get a new episode every week. And sign up for emails or find transcripts for every episode at no-priors.com.