September 16, 2025 | With the rise of AI and automation, what is the core human skill in drug discovery that is and will be valuable in the future? Derek Lowe, director of Chemical Biology Therapeutics at Novartis and author of the In the Pipeline blog, discusses the impact of AI—and non-AI technologies—on the business of drug making. With host Christopher Bahl, cofounder and CEO of AI Proteins, their conversation covers the value of a PhD in the tech world today, the long-term impacts of knowledge creation in US leadership positions, how the industry can navigate a biotech winter, and more.

GUEST BIO

Derek Lowe, Director of Chemical Biology Therapeutics, Novartis; Author of In the Pipeline Blog
Industry experience includes 10 years to Vertex, 9 years at Bayer and 8 years at Schering-Plough as a medicinal chemistry lab head, project leader, and specialist in new drug discovery technologies.

Freelance writer on science and pharmaceutical topics, with columns and articles appearing in Chemistry World, Contract Pharma, Pharmaceutical Executive, Genome Technology, and other publications. Author of the widely-read industry science and industry web site In the Pipeline, quoted in Forbes, Fortune, the Wall Street Journal, etc. Specialties: Drug discovery research experience in oncology, metabolic disease, Alzheimer's, infectious disease and other therapeutic areas. Target experience in kinases, proteases, GPCRs, ion channels, nuclear receptors, and many additional categories.

MODERATOR BIO

Christopher Bahl, Cofounder and CEO, AI Proteins    
Christopher D. Bahl, PhD co-founded AI Proteins in November 2021 and serves as president, chief scientific officer, and director.  Chris is a protein designer, biochemist, structural biologist, and entrepreneur.

Prior to founding AI Proteins, Chris joined the Institute for Protein Innovation in Boston as one of the founding faculty members, with co-appointments at Boston Children’s Hospital and Harvard Medical School. He pioneered the ability to computationally design miniproteins de novo as a postdoctoral fellow with Dr. David Baker at the University of Washington’s Institute for Protein Design in Seattle. In recognition of his innovative approach to creating therapeutics with the potential to tackle long-standing challenges in medicine and human health, Chris was selected as a 2019 TED Fellow.

In addition to his role at AI Proteins, he is the Founder and Co-Organizer of the Boston Protein Design and Modeling Club, a community of computational protein engineers and modelers from both academia and industry and sponsored project of the Open Molecular Software Foundation. He also sits on the Executive Council of The Protein Society, a non-profit scientific and educational organization that provides forums for scientific collaboration and communication and supports the professional growth for early-career investigators. Chris also serves as a Scientific Advisor for Applied Photophysics, Ltd. and BioLoomics, Inc.

Chris holds a Ph.D. in Biochemistry from Geisel School of Medicine at Dartmouth and received his MS in Biochemistry and BS in Biochemistry, Molecular & Cellular Biology from the University of Maine.

TRANSCRIPT

Welcome to the Chain, the podcast exploring the lives, careers, research and discoveries of protein engineers, scientists and biotech professionals. We look at the impact their work is having on the field and where the industry is headed. Tune in to stay up to date on the newest advancements and to hear the stories that are impacting the world of biologics newest advancements and to hear the stories that are impacting the world of biologics.

Hello everybody, my name is. Chris Bahl and welcome to the Chain Podcast. I'm here today with Derek Lowe.

Hi, thanks for having me.

Yeah, it's great to chat with you today. For some brief introductions, Derek, you're the Director of Chemical Biology Therapeutics at Novartis Institutes for Biomedical Research.

One of the directors. One of the directors and it's something where I should chime in that whenever I give interviews or podcasts like this, that I'm speaking on my own behalf and definitely not as a spokesman for my employer. They don't want that, and neither do I.

Fair enough, I see you perhaps as a little bit of a superhero, so you know scientist by day, but also one of the leading voices in biotech. You know in your spare time as the author of In the Pipeline that's been published by Science, plus many other articles and magazines like Chemistry World or Contract Pharma world or contract pharma, and so, yeah, really excited to be here chatting with you today. My background also, just for folks, is in protein design and drug discovery. I'm the founder and CEO of AI Proteins and, really, to kick things off, we'd like to talk not about AI, but about the things that people might be missing because of all the AI hype. So, Derek, we'd love to know a little bit about you know what are some technologies that you see that are not AI that you're excited about, and how you think they will impact the business of making drugs.

Yeah, that's a good question because, you're right, I do get asked an awful lot about AI and I seem to be slowly turning into the designated AI skeptic on a lot of conference panels and things like that. But I always tell them look, I'm not a skeptic, I like the stuff. It's just that compared to the outer fringes of the hype, I sound like I'm really down on it. But as far as the non-AI stuff, there is a lot going on and I have been telling people that I consider myself really fortunate to have lasted long enough in the industry to see all these new modes of treatment and new ideas coming on. And these were things that were rising up before we got into all the AI and machine learning cycle.

And by that I mean things like the sorts of strange large bifunctional molecules that are being made and a lot of people know these from protein degraders.

And I have to say, if you had proposed molecules like this, you know, back when I was in the industry back in, say, 1989, 90, you probably would have been fired Because these things would have looked ridiculous back then and no one would believe that the mechanism could work.

But it's a way of targeting specific proteins in the cell and causing them to be destroyed by the cell's housekeeping systems, the proteasome and others. That has effects well beyond the usual small molecule inhibitors, because any given protein is doing a lot more than just that one thing that has to do with this active site. It has all sorts of other partners and localizations and signaling networks. Now you're standing up there like Zeus and zapping them with thunderbolts and just making them vanish, and it really causes some significant effects. So I'm thrilled to be able to see this sort of thing happening. The idea is now being extended to all sorts of other stuff. You know you have all these post-translational modifications like phosphorylation. You can bring the enzymes involved with those. All of them, haul them over with bifunctional molecules to proteins that otherwise they wouldn't even be touching, and that, of course, is going to do things that we've never seen before.

And that's very exciting.

Oh yeah you can mislocalize proteins entirely. You can take them to parts of the cell they shouldn't be in and see what the cell does with that.

As an unabashed proponent of AI in drug discovery, I do have to comment that those molecules sound like they're pretty challenging to engineer and AI might be helpful for doing that.

Eventually. People are trying to do that, but at the moment we don't have enough solid data to feed into any kind of AI model.

It's being done we would love to but it's being done pretty empirically as it stands because there are all sorts of effects that kick in. You have to form what we call the productive ternary complex, your protein one which is the target, your business protein that's going to be doing the degrading or the phosphorylating or the deubiquitin or whatever, and you also generally need a third species which is at least the bifunctional, but sometimes there's another cofactor that needs to come in there too. And modeling that is a major challenge. You can get big effects if you just change the size and shape and geometry of the linkers between these two parts the ligand for the target protein and the ligand for the working protein and it varies in ways that are extremely hard to predict. I'm sure that we'll get a handle on it eventually and I think machine learning techniques will really be helpful for that, but right now there's just not enough solid data to dump into the hopper.

Totally agree, so maybe let's pull that thread a little bit. What sort of initiatives do you see unlocking that data, and how much of that do you think we really need in order to be able to start to make inroads here?

Yeah, people are piling this stuff up as quickly as they can because there are a lot of big organizations that are doing a lot of this internally and they are hoping to get an edge by machine learning their way to happiness. So we don't really have an idea of where the horizon is on these things. I'm hoping that will start to become clear, but it's going to be a few more years. Honestly, no-transcript you want to make and you say how should I make this? And that's traditionally been our job. We sit down and go.

Hmm, you know, I bet you could make this by combining these two things and forming this bond. And you could make that other one by doing this and this one could come from a so-and-so and one time I saw someone, and et cetera.

It's been very human-centered, but in theory you could mine the entire corpus of synthetic organic chemistry, all the reactions, all the substrates, all the examples, all the conditions, and use machine learning on that, where you put in the molecule structure and it would come back and say, ah, okay, this can be made by a metal-catalyzed coupling here, and piece A can be made from this, which can be made from this, which can be bought from this catalog.

Piece B was actually already made in 1982 by this group, in this reference and et cetera, et cetera. In theory you still can do that. In practice it has turned out that the organic chemistry literature is a little bit too much of a steaming, disorganized pile to get that to work easily. And it's to the point where I think we may and this pains me to say it, but we may have to actually go back and recapitulate a lot of synthetic organic chemistry under more controlled conditions to get data of high enough quality to do this kind of machine learning on. Right now it's shot with all kinds of biases and BS and worse, fair enough.

So there's still a long ways to go. Ai is not going to take all our jobs yet.

No, no. In fact, it's going to keep us all even busier for some time to come. So I guess that's the good news on it if you're worried about your employment. But it also means that you have to be very mindful of the effects that it can have and the areas in which it's most likely to start performing and be ready for those. You can't just you know, you can't assume that AI is just going to do everything perfectly next week and you can't assume that's all a bunch of garbage and you don't have to worry about it. Neither of those are true.

So, with the rise of AI and automation, what do you think is the core human skill in drug discovery that is valuable today, that will continue to be valuable in the future?

Yeah, yeah, this gets at. Something that I like to talk about when I'm talking with groups about AI is that it gets back to the definition of grunt work in research, and if you go back to when I was in graduate school, you know writing the mastodon over the lab through the ice flows. But no, this was back in the 80s. And synthetic organic chemistry, for example. We did a lot of things every day, several times a day, that no one does anymore because we have automation that does that for us, at least in the better equipped labs now, and it's working its way down through all parts of academia. By this time we wouldn't do these things anymore because we'd consider them beneath us in a way.

They're not a good use of our time. They're grunt work that machines should be doing, and the bar for grunt work keeps on moving up. I think that's a good thing. I think that if a job especially is repetitive enough or automatable enough, then it should be automated, and that frees up the humans to work at a higher level. Now, one thing I have to add at that point is that when I first gave a talk like that a few years ago, someone you know buttonholed me after the talk and said what about the people who are not able to work at that higher level? And I told him I deliberately did not talk about that because it does mean that a lot of jobs that used to be around that were just like okay, you know, keep your head down, make this, make that, make this, make that you don't have to worry about any bigger issues, just make this, make that those jobs are disappearing.

They were already disappearing over the years due to outsourcing to lower wage countries and now they're disappearing due to automation and machine learning. So it means you're going to be forced to work on the bigger, tougher issues and if you like that, great, if you've always wanted to have more time to do that, this is your time. If you don't like that, if you've been trying to avoid those things and just kind of plow along, it's not going to go so well for you. So, in general, those are the things that are not going to disappear, because there are a lot of things that AI can't do now and are going. Ai is going to have trouble doing for some time to come.

I can, actually I've got one for you. Here's sort of a hierarchy of what medicinal chemist problems are Right in front of their nose. It's what molecule do I make next and how do I make it? And that, how do I make it? Gets back to that literature searching I was talking about Past that you have to ask yourself. Hmm, how potent does my new compound need to be? How potent is potent enough? How selective against the other possible targets? How selective is selective enough?

These are much harder questions to answer, but there are even tougher ones. After that you say, all right, is this the right target for this disease? And got to tell you 80 something percent of the time it turns out not to be the right target. We have a clinical failure rate of about 85 percent in this business. Some of that's due to toxin. A lot of it is due to picking the wrong damn target in the first place because we don't know enough about the biology of disease. So that's a huge question Should we be working on this target?

But there's a bigger one. Should we be working on this disease, this therapeutic area at all? For example, should you be working on Alzheimer's? There's a provocative one Alzheimer's costs vast amounts of money to research. The timelines are slow and agonizing, incredibly expensive, and we still don't know what we're doing, because we don't know what causes Alzheimer's. Should you put your money into that or should you put it into something else? Very big question. Ai is not going to help you out with that or with those other big questions. That's up to the humans.

So maybe just to pressure test that assertion a bit, I think what a lot of people have anxiety about today actually is AI coming for the intellectual work and the large language models. Read all literature and all blog posts, all articles, and, in many ways, can start to chip away at the role that would typically be performed by a scientific advisor. Right, I mean, it's not very good at it compared to a scientific advisor, but people always talk about the. You know the rate of exponential improvement in AI. So while today it's, you know, maybe a precocious high schooler at best and its ability to provide advice, you can imagine a world and you know five to 10 years where it's actually providing world expert level. You know comprehension and feedback.

Yeah, I can imagine that, but unfortunately, so far what I'm seeing is that for and this is all for large language models not talking about others, but for large language models. What I'm seeing is the improvements are in producing faster, even more facile and convincing bullshit, and that's not what we need. We, as human beings, have been conditioned by a great long history of speaking to each other. We have been conditioned to think that anything that can manipulate language must be intelligent.

So people are, as you see, are assigning agency and saliency and sapience to large language models, which they do not deserve in the slightest. You see people going on and arguing with chat, gpt or grok or whatever, over and over, and, as I saw someone say the other day, he said my man, you are arguing with a vending machine. What is wrong with you? So, as it stands now, I am least enamored with the LLM approach for human language. Now for proteins. Sure, that's a much more limited space. You have very far fewer words, the grammar is more constrained and you have, you know, a huge corpus to train things up on. Well, you have it in human language too, but the problem is so much larger If you're familiar with the weird short story by Jorge Luis Borges called the Library of Babel.

It's a thought experiment about a gigantic universe of books which contain everything that can be put between 150 pages with the letters of the alphabet and a couple of punctuation marks Everything. And that means everything is in there Every insight, every bit of knowledge, everything that can be stated in words, but also every wrong thing, every stupid thing, everything that looks beautifully convincing but is completely misleading and, of course, oceans of gibberish beyond belief. And that's what we're doing now. We are opening up the library of Babel, we're filtering out the gibberish, but we are not able to filter out the convincing, wrong stuff from the right stuff.

Plus, I think there's a conservation law at work, a conservation of knowledge in the same way you have matter and energy. I do not think that rearranging existing texts will provide you with any new knowledge whatsoever. It may call your attention to things that were already there, that you didn't know about, and that is when it's functioning at its highest level, which is rare. It may call your attention to stuff. It may help you think about things from a different perspective, but it is not creating any knowledge. Knowledge creation is a separate scolium, it's a separate task and it's mostly up to us humans.

So maybe along those lines and folks that may be listening and thinking about career advice, particularly early career folks think about the value of a PhD in the world today. If you look over in the tech world, I would say it's not that common to find folks working in the tech industry that have a PhD. That's been the norm in biotech, but times may be changing. I would just love to know a little bit about advice that you would give to people should they pursue a PhD, and is the value of a PhD in the biotech industry decreasing?

Oh yeah, well, that is a provocative question, and you're absolutely right that, in tech, PhDs are uncommon, and in chemistry and molecular biology in these fields, they are very common. There's a reason for that, though, because a fair amount of the stuff that goes on in the tech world not all, but a fair amount of it can be done without a lot of equipment on hand. You can teach yourself to code, you can get really deep into a lot of very technical subjects on the software side without you know, on your own hook more or less, and that's great, but it doesn't happen in biology and chemistry so much. You cannot really have a well-equipped laboratory in your basement and you can't really have all the equipment that is surrounding you, like nuclear magnetic resonance machines and LC mass specs and cold rooms and 37-degree rooms where you're growing things cultures. There's just too much infrastructure, and that sounds funny because, of course, tech infrastructure can be very complex indeed, but there is just too much of it in biology and chemistry, and that is why spending time studying in the laboratory and doing research under academic conditions, that is to say, getting a PhD is still valued, because there's no other place to do it and you can't just join a small biotech and learn it all on the job. You're learning enough on the job as it is.

I didn't know any medicinal chemistry when I came into the industry, and that's pretty typical. Chemists don't come in with med chem degrees. For the most part they come in with degrees in synthetic organic chemistry, because that's what the company wants them to do and you pick up the rest on the job. But you can't pick it up from scratch. There's just no way. No one's going to pay for that.

So personally, I think the value of a PhD and a PhD in postdoc for that matter is still going to be high in the biotech industry. A postdoc, by the way, demonstrates that you can go off into a new lab working with a new group under new conditions on a new project. That you can go off into a new lab working with a new group under new conditions on a new project that you haven't worked on before, and pick it up and make progress. That's a very valuable thing to demonstrate and that's why a postdoc has traditionally been something that can look good for you. So I think that we're actually going to not see an erosion of the value of the PhD in biopharma.

Yeah, that's a very interesting viewpoint, I would say. As a founder of a biotech company, I actually see the opposite, where we I think that many people experience that the PhD is somewhat of a mixed bag for a lot of folks and that the first half of the PhD is often where you receive training and the back half of the PhD is where you demonstrate to the world that you have been trained and the salaries are sufficiently low. And I think the mental health of graduate students is also sufficiently challenging that we don't necessarily place a lot as much value on whether someone spent time in academia. We care more about what they know and I think the ability to learn on the job is also becoming as you know. As you mentioned, people don't learn med chem necessarily in academia. You learn it on the job. We value that they know how to do what we need and we don't care so much about where they learned it, and we provide a lot of opportunities for folks to learn and grow and advance their career.

Okay. Well, do you hire master's level folks instead, or do you hire bachelors?

We hire folks at all career levels and our career ladder doesn't have a glass ceiling that you hit without a PhD, and I think that's becoming increasingly common.

Well, it's true that there are people who can pick a lot of stuff up on the job, and if the entire industry goes in that direction, I think it could work out. Unfortunately, there are still these days, there are still a lot of organizations that will wonder about you if you have not done at least the master's and the PhD. You're right, the first half of it. You're often learning a lot of stuff. The second half is proving that you had the persistence to see this through, and that's not nothing.

I found it extremely painful while I was doing it, especially the persistence part, but I have realized in the years since that it did help and it actually helped me prove to myself that I could get through something like that. Now that gets back to the mental health point you made, and I completely agree. There are a lot of laboratories and a lot of universities out there that are not paying enough attention to that, to the quality of life and especially to the quality of the mental atmosphere in the labs. It used to be worse, but it's still nowhere as good as it should be.

That's fair.

Are you enjoying the conversation? We'd love to hear from you. Please subscribe to the podcast and give us a rating. It helps other people find and join the conversation. If you've got speaker or topic ideas, we'd love to hear those too. You can send them in a podcast review.

Thinking about academia. You know the somewhat the elephant in the room is the cuts to scientific research in the United States and I think the you know. I think it's pretty unanimous in science that folks are very concerned about the long-term impacts that this will have on the enterprise of knowledge creation and the United States' leadership position in this. We'd love to know your thoughts on when will this really start to impact the field of biotech?

Oh God, it's a horrible subject. I expected bad things from this administration but they have really, God help us, exceeded my expectations in that line. You can tell that these folks, the higher echelons of the current administration, are deeply hostile to science and deeply hostile to universities and, honestly, it seems like learning in general. And you can go back and see all these statements from people like Russell Vought and JD Vance and their intellectual forebearers, like that idiot, Curtis Yarvin, going on about how universities are the enemy. We must destroy them. We must, you know, bring them to heel. We must break them down, destroy them. We must, you know, bring them to heal. We must break them down. Well, by God, they're trying.

The damage is going to be severe and long lasting, even if we stopped right now and we're not. But even if we stopped right now, there would be a huge divot that has been taken out of federally funded science in this country. And it's very important to keep in mind that this federally funded science, through the NIH and its various branches and NSF, really has been a huge part of the engine driving scientific innovation in this country. I mean, that's a lot of the basic research, the blue sky stuff, is going on at that level and companies aren't doing that sort of thing because most of that money does not return or it returns in a very long time, in very unpredictable ways. But that, I think, is the perfect use of public funds for things like that, and we're stopping it. So if you were, for example I've done this thought experiment before the Trump administration, when people talk about, oh you know, let's really hammer those damn pharma companies, if you were, for example, to make it illegal to research new drugs as of Monday totally illegal, stop it all, pour it all in the waste. Can everyone go home, go off and go off and do something else? You wouldn't notice for a little while, because there's things that are already in the pipeline and going, there's stuff that's in clinical trials.

If you stopped the early research, no one would notice for a bit. But gradually, slowly. Early research no one would notice for a bit, but gradually, slowly, over a few years, you'd think, hmm, nothing seems to be happening. And after five or 10 years it'd be like you know, when I think about it, there hasn't been much new about. You know, all these diseases that we used to hear about. Some people would never catch on at all. Honestly, some people never would, but others would be like huh yeah, I thought they were going to cure such and such by now, or I thought I heard about some that was going to do so and so, and after say, 20, 25 years, it would be painfully obvious to anyone who cared to look that we had in fact cut our own throats by doing that.

That's what we're doing right now with the basic research. It's a little further back, so it'll take longer to show up, but it means the effects will be even worse. Now, that doesn't mean that basic research stops. It just goes on in other parts of the world. So what we are definitely doing is we are ending the era post-World War II, when the United States was the greatest place in the world for scientific research and technical advancement. We're stopping that. We have decided for some reason that totally escapes me. We have decided that we don't want that anymore. Screw all of it. And that's what we're doing. We're screwing all of it.

So I very much agree, and along those lines, the other thing people are talking a lot about is the rise of Chinese biotech, both in the speed and cost competitiveness of creating molecules. What is the role of Western innovation in biotech over the next, say, five years, as the industry seems to be poised to transform?

Oh God. Well, this is a fraught subject because, yeah, the Chinese have got a lot of very good scientists over there and there's a lot of really good work being done over there. There is, of course, also a lot of shit and of course, there's a good amount being done over here too. But the Chinese have been wrestling for years now over the system that they put together where people have to publish. They have to publish this many papers in this many journals with this high an impact factor, or they will not get promoted or maybe not even keep their job. It has been very explicit. Now over here we have that problem too, but it's often been a little bit more loosey goosey, like you know. It would really do your career a lot of good, really do your tenure application a lot of good, if you could get that paper accepted at Nature, you know, but with a lot of Chinese institutions, it is just laid out for you on the wall you must do this many, you must do that many, and they must be this good and they must go to these places. So this is why the paper mills are disproportionately located in and used in China. This is why there are so many fake conferences and authorship scams and reviewing scams and special issue editor scams running because these people have set it up, so the incentives to do these things are huge. So if you incentivize something, by golly you're going to get it. This also means that there's a cultural problem here too.

Some years ago I was talking with this is not China, this is Japan, but it's even worse for China. I was talking with this is not China, this is Japan, but it's even worse for China. I was talking with a researcher in Japan who'd found a very interesting possible platform for analytical chemistry that could be related to drug discovery. I asked him I said are you going to start a company around this? He said if I do, I can't start it in Japan. No one in Japan will give me money for something like this. If I want to start a company around this, I've got to come to the US and, honestly, in the US I've got to go either to the Bay Area or to the Boston area, maybe San Diego, and that's it. Nowhere else will I find VC backing from people who are willing to take chances on stuff like this and who understand its importance and who have done it before. There aren't any, and I found that very telling.

And China has that problem even more. China does not have what you'd call a startup culture. They have a tremendous research engine going and it does turn out some very interesting stuff, but it's different from what we have over here. It's different from Europe. Europe is, of course, closer to the US in that regard, but it's still. You know what it's like. Over the decades, people from England, germany, france who had this really wild startup idea. They generally try to come over and get it funded here because they know that that's the most fertile ground to land that on. And we're killing that.

The US biotech ecosystem, you know, has been pretty remarkable and you know all of us are based, are Boston, based here. You know lots of folks, I think, come to this city and try to understand what the secret sauce is so they can try it and replicate it later. And I think there's no one thing that makes Boston what it is. And I totally agree. I think that it will be some time before another culture, another country manages to recreate these ecosystems of innovation that we have in the US and I hope they're durable.

If they do at all, because there's no guarantee. Correct, Unfortunately. I agree with you about Boston. I mean, obviously you have very, very good universities and you have other factors. You have people willing to go out on a limb a little bit and then, when it worked, more people were willing to go out on a limb a little bit and then, when it worked, more people were willing to go out on a limb. But a lot of people are trying to replicate this at other places in the US. You know, over the years I've lost count of how many press releases I've seen about the such and such biopharma cluster. We're developing it right here in and you can just fill in the blank, you know, and some of these sounded plausible but never worked out. Some of them were implausible from the start. We're developing it right here, you know, in Grand Forks, and it's not something that you can just hook us into existence, unfortunately.

The ecosystem has so many components, also critical mass of talent that wants to work and even things that maybe are less glamorous. You know back to one of your points that you made earlier about you know you can't start a biotech lab in your basement. Actually, I think you can. These days there's a lot of stuff that you can use, scientific equipment you can buy at bargain basement prices and you know, for us at a scrappy startup, one of the things that's really helped us thrive and grow is the secondhand lab equipment ecosystem that exists around Boston and we can just drive 45 minutes away and pick up large pieces of capital equipment that we can get, often for 95% off what a retail price would be.

Oh yeah, that's a good point, and that is something that is very much harder to do in El Paso. Absolutely, yeah, that's a good point. And of course there's also the contract and outsourcing aspect of things where you can do more remotely. It's not optimal, but sometimes it's very cost effective. So, yeah, things change because of that. But a lot of people wondered if those factors were going to diminish the importance of the research hubs like the Bay Area and the Boston-Cambridge area.

Whenever we're working with a contract research organization, we always prefer to go local, just for the ability to drive over there and talk to them or deliver samples. It's such an advantage.

It really is, and being eight or 10 or 12 time zones away is the exact opposite of that experience.

Absolutely, and even just the networking events that you have in Boston, just that critical mass, you need it.

Exactly. You really have the feeling that if you're working in the Boston/Cambridge area, that if you just keep your eyes open, almost everybody whose talk you would want to hear is going to come through here and give a talk. Almost every you know event that has the speakers in it that you want to hear and that you want to talk to and that you want to meet other people who are interested in. This stuff is going to happen around here at some point, and it's not that it never happens anywhere else. And there are places also that have great you know speakers and seminars and symposia and stuff, but there's just a flood of it around here. You just can't keep up.

So, as great as Boston is, I also say part of the reason why we can buy so much great lab equipment at discount prices is because we appear to be in somewhat of a biotech winter at the moment, and we'd love to just kind of get your thoughts on this. I mean, is this going to come to an end, or do you think this is our new normal and we need to adjust expectations?

Yeah, I mean I now. I guess in October I will celebrate my sheesh 36th year of doing, of working in the biopharma industry. I got I started in the in the business in fall of 89 and I've seen several of these cycles, but this has been an especially big one because the boom preceding this winter was especially big and long too, as everyone will recall. Now a lot of folks may not have realized that because they hadn't been around long enough and they maybe came into the industry during this and took it as the normal. It's just great. I love it over here. The money is just raining down from the sky. You can start a company next Tuesday. Oh my God, it's great.

That has not always been the case. So there was a very long boom and a lot of more experienced people were like you know, someday this is going to end, and people tended to call it too early, in the same way that people called the 1929 stock market crash too early. A lot of people lost their shirts, not in the crash but in trying to sell short in anticipation of the crash earlier in 1929. And they got just slaughtered because things kept going up. But eventually these things do come to an end, and this one, by golly, has definitively come to an end for some time now. But I don't think that means it's going to be like this forever. But let me back that up Until we started slaughtering the basic research engines in this country, I didn't think that If you wanted to create a biotech winter that turns into a biotech ice age, we are in fact doing the things that you would do to cause that to happen.

So under normal business cycle conditions, I would say no, it is going to come back. It's going to be painful, it's going to be long. It's going to be longer than you want or can stand, but it will come back Now if we keep on the course we're on. I don't know.

Well, I remain hopeful.

Oh, I try to. I try to too. There's no other way to live, and I'm the optimistic guy around here. When I took on my current job you know this is several years back I was a pretty experienced guy and I wanted to make sure that I wasn't turning into one of those folks who sit in the back of the meeting room and go, ah, that's not going to work. Nah, we tried that, nah, seen that before. What the hell good does that do anyone? I mean, most things don't work. You can sit back there and say nothing's going to work and you'll be right 90% of the time. Or what the hell good does that do I? Prefer to be the person who perks up and says you know, I like that, that's interesting, let's try that, and that is my personality. But they're making it difficult for me to maintain my buoyancy.

That's fair. So thinking about how do we, you know, how do we pull out of this biotech winter? Then, you know, maybe bringing us back around to where we started, the conversation around AI. I think a lot of folks are pointing toward AI to solve the world's problems, regardless of how true. That is one thing I'm a bit curious about. AlphaFold 3 has attracted a tremendous amount of attention, so has AlphaFold 2 and AlphaFold 1. And you know, imitation is the sincerest form of flattery, and we've seen a proliferation of many, I dare you know I don't mean to use this in a pejorative way, but call them alpha fold clones, because, by and large, they're attempting to do something similar. And then there's, I'd say, a rapidly expanding cadre of startup companies that are creating closed source AlphaFold clones, and many of them, I would say, are incremental advancements on what is free, and they seem to be attracting a lot of investor attention. And we'd love to get your thoughts on these companies. What do you think their business model is and do you think big pharma is going to fork over the big bucks that I presume their VC backers are expecting?

Yeah, good question. If they're going to get anyone to fork over those big bucks, they're going to have to show some things that you can't get anywhere else, and that's what I'm skeptical of, because right now, I very much doubt that any of these folks have something to point at to say check this out, you're not going to get this from AlphaFold, you're not going to get this from RosettaFold, you're not going to get this from all these open source attempts at these things. Nope, nope, nope. We've got it and no one else does.

I don't see that that's going to be very hard to realize and it's going to take a great deal of effort that some of these VCs may not realize that they've signed up for. So no, I share your skepticism on these things because, I mean, it's a real market and it has real potential, but it's already being pretty well served and the open source varieties of this are trying to serve it even better. So if you're going to try to do a proprietary version of these things, you're really going to have to bring something startling to the table, and I don't think startling is fast and I don't think startling is cheap.

Very fair. Well, they're certainly raising a lot of money. I guess it tracks the hypothesis that this is perhaps a bubble, an AI bubble, and typically when those pop, the result is cynicism from investors and moving away from the field. Do you see that coming or do you think these businesses are going to figure out some kind of way to monetize these algorithms in time?

I think the former is more likely. Unfortunately, I think the former is more likely that there's going to be some popping bubbles and a lot of bad feelings after them. But of course, computational chemistry, computational biology and drug design these things have had cycles of enthusiasm and anger following over the last 30 years. When I took on that first job in fall of 89, they were just stuffing a room with Evans and Sutherland workstations there's a blast from the past and they had the sign over the door that said CADD and no one knew what that was, so they had to spell it out. Computer-assisted drug discovery the chemists across the hall didn't like that much. They put a sign over their lab that said bad brain-assisted drug discovery and of course that didn't work out at the time in the early 90s, and everyone got really upset about it. And then it came back and that didn't work out. Then it came back and that didn't work out.

But there are a lot of bubbles set to be popped. I am past drug discovery. I am very skeptical of the ability of a lot of the generative AI stuff. We're seeing vast amounts of money poured into it. I'm talking about OpenAI and Zuckerberg stuff all these companies. I am very skeptical that their business models even can pay off, and if that blows up it's really going to lead to a lot of hurt feelings.

That's fair. So one thing AI and protein design can definitely do today very well is redesign proteins to cheat their sequence substantially while preserving the activity of the molecule. And one of the things that we think about a lot and we worry about is how good our composition of matter patents are for the molecules that we create. How easy will it be for someone to rip off the molecules that we create and, frankly, we are? Our position today is that I don't think that the composition of matter claims that the US Patent Office is currently granting will actually protect any protein molecule.

And do you think that the US Patent Office understands this? Like what, what? What would we need to fix this? And assuming that it's not fixed, because all of the turmoil in our administration, how do you think this would impact the biotech industry?

That's a really good point. That is a really good point because I'm a small molecule guy and composition of matter patterns are still extremely valuable, but of course, being small molecules, small changes can make a big difference. You hear about the magic methyl and I've seen that you throw one methyl group on a compound and suddenly it becomes a hundred times more potent or selected. So yeah, a composition of matter still means a lot. But you're absolutely right, on the scale of a protein you can get away with a lot, and there are a lot of things that are more or less equivalent. Now that's the key word, because I think that in theory, the patent office would respond to this by putting more teeth into the legal doctrine of equivalence, which I think is stated in non-legal terms as a difference that makes no difference is no difference, and you can see that applied to small molecules, where if you just go in and stick a fluorine on the back end of some molecule that they forgot to claim in their specifications in the patent and it makes a molecule that does exactly the same as the other thing, the patent office is not going to allow that. They will say that that is not an inventive step under the doctrine of equivalence.

So I think that that might be where they would head with composition of matter patents on proteins. They'll say great, you replace this other thing, you replace this lice with an orange Whoop-de-doo. It makes the same kind of protein that does the same thing. It is that is not patentable. You did not break this other patent by doing that. So my guess is is that's where you'd head. Now, that's going to be pretty fraught, because there's a lot more issues in expanding the doctrine of equivalence in proteins than there are in small molecules. But that would probably be the way to do it, because otherwise, yeah, you end up with a real Wild West, don't you?

You sure do. And I guess that becomes the argument is is it a change to the molecule? And the same molecule fundamentally is equivalent if you change, say 40% of it. What if you take some loops off and remodel them and change the length, change the spacing between the functional amino acids, how much do you have to change with protein before it's a different protein. ? the protein of that is fairly trivial nowadays.

\All, and this is the sort of thing that's traditionally going to be fought out court case by court case. And I do not envy the judges in these because, yeah, what you would do in a case like that is you would say you know what, by deleting these loops, I have made this protein also have a better half-life in blood serum, or I have made it more resistant to proteolysis if you're trying to come up with an oral formulation, et cetera, et cetera. But if you were to just say I deleted these loops and it does the same damn thing, then I think the patent office would be within its rights to say, okay, you had the same damn thing, we do not patent the same damn thing.

So if it's going to be fought through the courts, that means we're looking at at least a decade before patent law catches up.

I hate to agree with you there, but I do, and it will enrich a lot of lawyers along the way which is always the default ground energy state.

Very fair, all right. Well, on that note, Derek, it's been so much fun chatting with you today. Thanks for taking the time to come have a conversation with me.

I enjoyed it. Definitely. Thank you.

All right. Till next time, cheers.