Oliver Vince, Co-Founder at Basecamp Research: BREAKING: Basecamp Research raised $140m Series C

0:00Oliver Vince:Hello and welcome back to Scaling Europe show. I'm Seb Johnson. Today on the show, we've got one of the co-founders of Basecamp Research, one of the most exciting companies coming out of Europe doing really cool stuff. They've just announced a huge round of funding. But before we get into it, Oliver, thank you for joining me. How are you doing today? Very well. Thank you for having me. it's absolutely my pleasure this is phenomenal news but like before we get into the round what unlocked it and some of the the amazing stuff that's going on can you give like a one or two minute quick pitch what is it that you've been working on and yeah tell us what is what was it all about yeah i mean base camp at a very high level is looking at using ai to understand biology at a massive scale.

0:41And so what we do is we basically build huge genomic data sets, we build huge language models on those genomic data sets, and we use them to interesting and important applications. Obviously, the better and the bigger these models get, the closer and more relevant they become in medicine. And so we're now sort of starting to see them be very, very useful in the application of like, ultimately, AI for curing disease, which has been a very, very, very exciting journey to go on.

1:08Oliver Vince:Which is amazing, right? You know, there's a lot of talk about AI, there's a lot of backlash against AI at the moment, but like really using AI to cure some of the most dangerous and horrible diseases. And it's a real use case for AI for good. You've just raised how much and who by? Raised 140 million, which is a great new round contributions from, I mean, led by S32 over in the US, but contributions from, you know, investors, existing investors, new investors all over the world across tech across pharma one of the you know our longest supporters has been andre hoffman who's you know the vice chairman of rosh to give you the complete other side of the sort of tech pharma split if you like and yeah just super proud of all the different people that come around the table and you know excited to build with them basically and i guess a lot has changed in the company since your series b you raised i think it was million dollars led by singular i think towards it was announced at the end of 2024 since then there's been some really exciting announcements can you talk about some of those maybe like uh some of the models eden models the gene atlas that you've launched can you maybe talk about those two releases and why i guess i assume they were so important in unlocking this big series c yeah so i mean as i said at the start right it's a relatively simple idea we're just going to try and teach ai to understand biology and then the question is kind of like what do you do with that why?

2:31Why do you care? And effectively, we're seeing the same thing happen in these models as happened in the language space, which is the biology of ornamental language. You know, you can learn to speak biology if you like. And that basically the bigger these models get, the more data they have, the smarter they get, and the more capable they get for useful applications. So to run you through some of those, you know, we have been building this database for a long time. We have these partnerships in sort of 30 countries, more than 200 different locations around the world. and we have been scaling this data set accordingly on that and then one of the big releases that made our first public partnership with nvidia was the eden models um so these are gbd4 scale the largest biological models ever trained and then we've done a bunch of interesting things with them so one of the things is you know we put it into claude science with anthropic uh where you know you could do things like upload a patient microbiology report so someone who's got an infection, you can put that file into Claude.

3:24Claude will work with Eden to design a new antibiotic for that infection and then give you back the antibiotic. And, you know, there's some barriers to that to actually implement it in hospitals, of course. But, you know, it works very, very successfully in the lab. It works very, very successfully in the first sort of preclinical studies. And so it's sort of demonstrating the point of where these models could go. And then sort of underneath all of that, we've worked, you know, we've partnered with NVIDIA Anthropic, PacBio and Ultima Genomics to then build what we call the Trillium Gene Atlas, which is taking our data set and expanding another 100x.

3:53So it's kind of saying, look, we've seen the start of these scaling laws in biology. We believe it's worth continuing. And then there's a sort of very, very clear sort of technical case for that. But also, you know, at the end of that curve, we believe that there's some sort of profound implications of what these models will be able to do for the benefit of humanity.

4:10Oliver Vince:It's crazy stuff. You've been working with NVIDIA and Anthropic. Both of them have now come in as investors through their sort of investment arms. What do you think that says about the partnership so far? and why was it great or important to have them on board? So I think that the work that we do has a lot of different technical elements to it, right? There's, you know, to split it into three categories, you've got the data collection arm, which requires, you know, running expeditions around the world, building partnerships in, you know, we've got partnerships in Africa, Australia, like all seven continents, but, you know, some of like the most exciting, you know, environmental places all around the world.

4:50and so that's sort of one arm then there's the tech arm which is basically right how do you scale data sets how do you scale models trained on these data sets to sort of you know currently we're a gb4 scale going beyond that now uh and then there's a whole sort of like pharma set which is like how do you actually take that and make medicine yeah you know how do you how do you get that to the point where it could actually be in a in a patient in a hospital down the line and i guess all of those different pieces have very very different expertise and effectively who worked with NVIDIA and Anthropic on that sort of middle piece and not to, you know, not to do ourself a disservice, but to like help us to leverage what they have learned in that space and apply it to our field rather than trying to relearn the same thing, you know, ourself, if you like.

5:31Oliver Vince:And it seems like you've been very, it's been very thoughtful around who you're getting involved, right? Because you mentioned, you know, Andre Hoffman, you've got NVIDIA, Anthropic, uh like there's a stack of people from across biotech pharma ai it feels like you're trying to build in like this like avengers scale almost like cap taping for this company have you been very intentional about that you're trying to get people who are both a combination of what i call realistic and optimistic right like you need to be you're trying to do something that is very serious, very dangerous, you know, in terms of like, you're trying to treat people with very serious diseases who are, you know, in a very tough situation.

6:17This is not you're not training chatbots that, you know, if they make a spelling mistake, it kind of is okay, you're training things to have, you know, the potential to have a really meaningful impact on people's lives. And so you need to be realistic, sensible about what that is, you can't just be as flippant as some of the other companies have been in sort of the AI space. But at the same time, you'd be optimistic that the way we treat cancer, the way we treat autoimmune diseases, isn't going to be the same for the future. And, you know, the way we look at it, just as a little bit about that, right, take something like cancer, it's a very, very difficult disease to treat.

6:48But the reasons are relatively simple, right? It's individual. So each person generally has a new type of cancer. So even if you have the same like broad cancer, it's normally individual to you. It's very, very complex. It's never normally one thing that's causing it. And it's also a moving target, right? So you've not got very long to treat it. And so the way that we approach the problem from an optimistic sense is that the only way to treat that will ever be for lots and lots of people, of course, is with an algorithm that can come up with a solution to that sort of very, very complex set of challenges and develop a solution.

7:20but the realistic part of you also needs to be there which is that this is very very difficult very very you know uh hard to do and we need people around us who have spent their lives and their careers actually doing it to make sure we have the trust and the credibility to allow ourselves to actually get to that point um because i think in you know in the in the deep domains we've been working on this for a long time we will be working on this for a long time it's a very different you know to just scaling software if you like oh yeah absolutely this is

7:49Oliver Vince:yeah an order of magnitude more complicated and more regulated and a whole other host of issues i imagine but when you talk about specific diseases you spoke about cancer autoimmune problems do you think there are certain health issues whether it's bacterial infections or cancers or degenerative brain diseases where this type of technology that you're developing can have the impact first i so what we are at a very simple level trying to do is learn to read and write dna and so it will be applications where DNA has a primary driving cause and that the solution can be found by designing DNA so examples of that include genetic medicines so this is something that you know anthropics talked about going after which is their sort of rare disease or neglected disease piece but what that basically means is a disease where a piece of the DNA has gone wrong or maybe not gone wrong but it's sort of changed in a way that is adversarial or causes you know in some form of illness in the patient.

8:47And effectively, you know how to fix it. But you have to do a very specific fix for that person. And that's sort of normally at this point expensive to do. And if you can get an algorithm to figure out how to fix that piece of broken DNA, then you have a solution. In cancer, it's sort of similar, slightly different, slightly more challenging. But in cancer, it's still a disease of the DNA, very fundamentally. You know, your DNA has gone wrong, and your body has not recognized that mistake. And so your body has not recognized to shut that down. And what you're trying to do is teach the body to recognize that.

9:16And so when we say curing cancers, it's less about we're going to develop another chemotherapy and it's going to hit it even harder. It's saying everybody gets cancer all the time because every day, you know, your cells mutate and your cells divide in ways that are not great and your immune system shuts it down most of the time. And so in a scenario where you get the disease like that, what you need to do is teach the immune system to say, hey, look, look, this is bad and go after it. and then you really can just like restore your body effectively back to its healthy state right which is what you're trying to do so that's all not all very simplistic way you can think of it as well just recoding the dna um and then you know you've got things like the antibiotic model we've already produced which is where the pathogen is in dna and then the the antibiotics also in dna so you can sort of figure out how to get it to work on that on those sorts of problems there's a whole host of things you can do it doesn't mean that dna is the only thing you ever need to do or it's the only thing you'll ever want to do.

10:09But if you can figure that out, there's a long way to run.

10:12Oliver Vince:And can you talk about where you're at in this process? So I know that a large part of this announcement and this money is going to take you, I guess, towards more clinical development, but you've been gathering this data, you've been training the models. The models I understand have also been used to some extent within Claude Science. So where are you? Where are you at today? Where have you seen people taking and running with some of these models or designing things? and what do you need to do to get to clinical development? Yeah, so we kind of have three parts of the business. So we have the data expansion piece, that space, or the data collection piece.

10:46We're still doing that. We're still working, building partnerships around the world. All these different companies recently signed some in some very exciting locations that we're going to talk about soon. We have also, you know, that's what the Trillion Gene Atlas is based on and that's grown and that's on track. That's growing very, very quickly. It's on track to be completed by the end of 2027. then the sort of we're still scaling the models we're still building newer models getting better at those models and evaluating them on tasks that people previously haven't done and that includes tasks that we're what we call long horizon so rather than saying to the model can you design me one molecule does this it's saying can you think about the patient can you then you know analyze you know what needs to be done can you then design a molecule can you then assess safety and then can you do all these things and what you're working towards is not a model just as one task at a time.

11:34But in the similar way to like, you know, these, these frontier models today, you, you input a patient file on one side and outcome something that's on the validation. Now, to be clear, we've got versions of that, such as the antibiotics and clawed science, but we're not, you know, we're not at the point where it can do the whole thing yet. And, you know, there's a lot of work to do and understand that, but then the sort of validation that we are actually at the point where we can start treating, or we think we start getting close to treating patients. We have this pipeline of medicines that we are working on.

12:01We are developing, and that we will work with pharma companies on to develop further. And there's all this sort of stuff to be built. So back to sort of realistic and optimistic, like realistic is kind of like, we've got a very, very strong validation of all the technologies, but we've not yet really lent into it at the scale that will be required to deliver the impact of this technology to those who need it.

12:26Oliver Vince:And is that a point in time that you look forward to as like the very first patient? You know, is that something where you have that milestone? What does that look like working backwards? Like how far off is that? It's hard to say exactly. And it's also difficult. You don't want to promise things because, you know, it's people's lives, right? So you don't want to. But it's, you know, in the next couple of years, I expect we'll be there. And, you know, my hope is that when we are doing this, It's medicines that are curative. So they're curative for diseases that previously or currently cannot be treated.

13:05And they start to demonstrate the potential of AI in this space, which is that, you know, in my view, even if your first medicines sort of equal or match what can be done today by humans, the fact that an AI model has got there means that if you continue to work on it, it will go through that. And again, realistic and optimistic, right? Realistic, we don't know exactly how that's going to scale, how that's going to work, but optimistic that there is a better way of treating disease than we have today. And that if we can get that and we can invest in this technology, we can get to a point where certain diseases, which are very commonplace today and are absolutely disastrous, are kind of not a thing anymore.

13:43as in people will still get them, but in the same way that most infections in the West, at least, like they're not a thing, as in if you get an infection, you go to the doctor, you get some antibiotics, whereas PVC would have killed you. We hope to move more diseases into that space, basically.

13:59Oliver Vince:Yeah, that'd be amazing, right? It's just a game-changing way of approaching disease. When you talk about it, do you see yourself becoming a drug company or will it always be partnerships with pharmaceuticals? I think in the long time, it will be both. I think that we will always focus on concentrating where we believe we can contribute, where people have already built the infrastructures, or they've built manufacturing plants, or they've built hospitals. We should not try and rebuild them, and we should try and work with them. But where we believe, you know, for these AI models to have the most impact, they need new infrastructure, or they need new ways of interfacing with the patients or they need new data and that sort of thing, then I think it's our progress and our responsibility to build it.

14:44And, you know, it's a bit of a, we've built quite a lot of different things already. And I'm more focused on sort of the end outcome, right, which is demonstrate AI can cure disease. And it might need other things to help it get there, but it's the premise of that actually working and, you know, all these pieces, you need to build AI, you need it to focus on curing, not just treating, and you need to actually get to the point where you're treating disease it doesn't like it almost doesn't get more complex than that so you need to do all of it and i don't know what you identify as when you do all of that but it's definitely exciting

15:18Oliver Vince:and when i look at the landscape of other companies in this space whether it's isomorphic in london and it's also some of the u.s maybe like evolutionary scale it feels like and maybe i've got this wrong you've taken a different approach to data and training the models and And you seem to be betting really big on building this almost unique proprietary data set, whereas I think a lot of those other ones are kind of using public data. Firstly, is that right? And then why have you taken this different approach to some of the other companies in this space? That's correct. I mean, we are, for want of a better word, trying to build general biological intelligence.

15:57We're trying to build models that get generally smarter. And I think other people have approached the problem a bit more like linearly. So a bit more saying like, look, here's how the drug discovery process works today. We're going to build a model for this and a model for that and a model for this. And we're going to sort of treat each one like a bit of a game. And we're going to play each one. And we're going to get really, really good at each first. And we're going to stick them together. And what that means is in each of the individual things, they get very, very good at that. But they sort of, in our view, you know, I think that's very promising.

16:30and I don't want to be negative on any of it, but I think that has certain limitations in the fact that you're replicating the existing process. Now, we've taken a bit more of a leap and said, look, what if you just trained this thing to understand biology and you learned the ability to talk to it, what would it be able to do for you, right? And so an example of that would be, normally when you treat antibiotics, you would take the pathogen, you would do a bunch of modeling on the pathogen, you'd find the target, you'd find the mechanism of action, you'd find the thing, you'd do the docking, and you'd do all these steps and then you design a peptide or an antibiotic that worked against that.

17:04And then you test the sort of five steps in that process and you build a model for each one. What we did is we just built this model and we showed it lots of pairs of pathogens and antibiotics and just said, here's a new pathogen, give me a new antibiotic. And it just jumps from that. And 97 % of the time it works in the lab and it works equivalent to the last line in mice and that sort of thing. And we've no idea in principle how it did that jump. but we know that it's able to jump from A to B, right? So kind of in the way that you're not quite sure exactly how ChatGPT took your prompt and jumped to something at the output, but you know that it did and you can check that it did.

17:41Rather than saying like the equivalent with ChatGPT would be saying, we're going to try and understand what your sentence structure is. And then we're going to try and understand the meaning of each word. And then we're going to reason our way through all those steps. We've just said, look, train it to be big enough, scale it, and then ask it a question and see what it can do. And quite funnily and quite hilariously, when it's smaller, it can't do a lot of these things. So you ask it the question, it spits out nonsense. And so it's kind of very simplistic in some senses. Obviously, it's a bit more complicated.

18:13The team wouldn't like me saying that.

18:14Oliver Vince:But it's scary. This black box of we're not quite sure how it works, but it's almost magical. What was that like for you and the team, I guess, when you had those great 97 % results come back? it's it's it's a pretty like it's hard to un when you exist in a world that requires all that scientific rationale and has previously built on like you know very very rigorous experimentation to just take a new approach and just say look we're going to do this to see what happens like you know it's a high risk kind of but you know it's a rewarding way to be because it's not even rewarding in the result that you get today so it's not the 97 is good but it's not the reward right the reward is that you have a direction of something that has got there in a different way so you know you can get the same result by putting lego blocks together but the fact that you get to it using something that is completely unsupervised means that you you have a technically feasible path to giving it more and more complex questions and then back to my earlier point on cancer being a personalized complex and a moving target it gives you the potential for something that one day could understand all those different parameters, think about them in a way that we don't understand that's far beyond our current comprehension, generate a solution and hey presto it works right and we're not there yet to be clear but like that's the sort of like landscape that we open up by having a sort of unsupervised scaling.

19:40Oliver Vince:Yeah I mean it demonstrates a real level of intelligence versus just a sort of like training it to respond to certain questions I guess or information retrieval. Exactly yeah. And so So why wouldn't this work? Let's just say three years time, it doesn't work. Where will be, do you think, the areas where this would have failed? Is it the data itself, maybe not quite having the scale or atlas that you really wanted it? Is it the intelligence that maybe the intelligence can only be scaled so far? Could it be in the application itself? Where do you see at this point the biggest chances of failure being?

20:15I think they are realistically tactical, right? They're, they're, do you overspend in some, you know, do you, I think it will work. And I think if you had infinite money and infinite time, I would be a hundred percent sure that it's going to work. The question is whether you, from an accompany building standpoint, whether you can do that. And I think like, like more specifically, right? Like it's kind of, you, it's again, back to being very, very simple. All you're doing is learning to read DNA and write DNA. and basically the bigger the model gets the longer piece it can read with higher accuracy and the longer it can write with higher accuracy and so basically the question is like how much how big does the model need to get to write you know 20 letters 30 letters 40 letters and then what can you do with 20 letters 30 letters 40 letters and how valuable is that right so like isn't if you need to operate in a world where you need to monetize 20 letters for like five years and then monetize 30 letters for five years, you can do it, but it's a very different approach to saying, look, the fact that we've gone from 20 to 30 to 40 shows that we will get there to 100 or 150 or 1000.

21:29But that requires a bit of a leap of faith, which ultimately is where the tech companies are interesting because they've taken that leap of faith and they've seen it before, which is very, very different in some senses to many of the established ways of developing medicine.

21:42Oliver Vince:Have you faced any cynicism or skepticism from that more traditional industry versus the tech industry?

21:51Of course we have, right? And like, I mean, to be clear, like I, you know, many of us have, I have background in my PhD was in cancer research. I have like many of our team have, are like trying to be realistic and optimistic at the same time. And so we're aware of our own flaws and where this doesn't work better than anybody else. but it really is just like look there is a promise there and therefore it's worth doing and you kind of if you can get people to accept that it's a very very very exciting and very fulfilling thing to do because you're not guaranteeing anyone anything but you're saying look here's the potential it's unequivocally huge if it works now let's go on the ride and see if it

22:31Oliver Vince:works right let's take that swing it's amazing i love it this is like it's um it's amazing to hear you know, there are some companies today that are building truly frontier technologies, especially here in London, that are really game changing for the world, right? And, you know, there's a lot of stuff going on, a lot of doomerism around AI at the moment. It makes me so happy to see that so many of these great companies that are truly tackling some of the biggest problems in the world are like here in London doing amazing stuff. Well, anyway, thank you so much for this phenomenal news. Huge congratulations.

23:01Oliver Vince:And I look forward to seeing your progress and your journey and maybe be treated one day. You're very kind. Thank you, my friend.