Latest / Making Medicine / Can AI Reinvent How America Makes Medicine? | How America Makes Medicine Ep. 6

Transcript

Can AI Reinvent How America Makes Medicine? | How America Makes Medicine Ep. 6 · Making Medicine

0:00
Speaker 1: At Making Medicine, we're reporting on the headlines, not making
0:03
Speaker 1: medical recommendations. For personal health questions, always ask your doctor.
0:08
Speaker 1: Nothing in this episode constitutes investment, financial, or legal advice.
0:13
Speaker 1: Consult your own advisors before making decisions.
0:16
Speaker 2: The thesis that we are most excited about is the
0:19
Speaker 2: possibility that in the future, there will be a single
0:22
Speaker 2: person or a couple people founding a biotech company that
0:25
Speaker 2: will grow to be a billion-dollar-plus biotech company.
0:33
Speaker 1: Welcome back to the Making Medicine Podcast, where science, policy,
0:37
Speaker 1: and innovation meet. I'm your host, John Stanford. This is
0:40
Speaker 1: part six of our series, How America Makes Medicine. And
0:44
Speaker 1: today we're sitting down with Patrick Malone to zoom in
0:47
Speaker 1: on the piece that's reshaping every part of this industry
0:49
Speaker 1: right now, and frankly, every industry, artificial intelligence. Patrick trained
0:55
Speaker 1: as a physician scientist, earned his MD and PhD at
0:58
Speaker 1: Georgetown with research in computational neuroscience and machine learning. before
1:03
Speaker 1: deciding the best way to move medicine forward was from
1:05
Speaker 1: the investor's seat, rather than the bench or clinic. He's
1:08
Speaker 1: now a partner at KDT Ventures, an early-stage venture firm
1:12
Speaker 1: investing in computational science and life science firms. Their active
1:16
Speaker 1: portfolio covers more than 60 investments, with more than 80%
1:19
Speaker 1: in the healthcare industry. But beyond investing, Patrick's become one
1:24
Speaker 1: of the more outspoken voices on where the industry is headed,
1:27
Speaker 1: from how AI agents are reshaping what a biotech company
1:31
Speaker 1: looks like, to a recent piece arguing biotech badly needs
1:35
Speaker 1: its own version of a political power broker, the way
1:38
Speaker 1: AI and crypto found in David Sachs. It's gonna be
1:41
Speaker 1: a great conversation, and it's one that is gonna talk
1:44
Speaker 1: about the future, the next 250 years, and the role
1:48
Speaker 1: AI will play.
1:49
Speaker 3: So let's get into the show.
1:52
Speaker 1: We'll get to Patrick in just a minute for a
1:54
Speaker 1: great conversation about AI, but it hasn't been a sleepy
1:57
Speaker 1: August here in Washington, and we'd be remiss not to
2:00
Speaker 1: hit on a couple headlines that are making news as
2:02
Speaker 1: we record. First up, a landmark approval for pancreatic cancer,
2:07
Speaker 1: a disease that frankly has needed good news for decades.
2:11
Speaker 1: Just minutes ago, the FDA gave expedited approval to a
2:14
Speaker 1: daily pill from Revolution Medicines that blocks a mutated protein,
2:18
Speaker 1: fueling tumor growth in more than 90% of pancreatic cancer cases.
2:23
Speaker 1: It's the first drug to successfully target this pathway something
2:26
Speaker 1: researchers had chased for decades.
2:29
Speaker 3: The results are striking.
2:31
Speaker 1: In a randomized trial of 500 patients with previously treated
2:34
Speaker 1: metastatic pancreatic cancer, median overall survival nearly doubled from 6.7
2:40
Speaker 1: months to more than a year, especially when compared to
2:43
Speaker 1: standard chemotherapy, and.
2:45
Speaker 3: Had fewer side effects.
2:46
Speaker 1: The drug picked up public attention earlier this year when
2:49
Speaker 1: former Senator Ben Sasse described his experience on 60 Minutes.
2:53
Speaker 1: It drove a surge in demand and led the FDA
2:55
Speaker 1: to open expanded access before formal approval. Regulators ultimately cleared
3:01
Speaker 1: the drug six and a half months ahead of schedule,
3:03
Speaker 1: a signal officials said the agency's push to move faster
3:07
Speaker 1: on serious and life-threatening conditions. This is good news for
3:11
Speaker 1: everyone in the pancreatic cancer community. Our other headline from
3:15
Speaker 1: a few moments ago was scooped by our friend Naya
3:17
Speaker 1: Pankcity over at Bloomberg, which is reporting that the Trump
3:20
Speaker 1: administration plans to announce new most favored nation pricing agreements
3:25
Speaker 1: with several mid-sized biotech companies as early as Monday. This
3:29
Speaker 1: builds on a string of similar deals the administration struck
3:32
Speaker 1: with major pharmaceutical companies starting late last year. The next
3:35
Speaker 1: round is notable because most of those companies were large
3:39
Speaker 1: cap pharma names. And this will be the first time
3:42
Speaker 1: we see some deals coming out of mid-sized biotechs. This
3:45
Speaker 1: is certainly something we'll have more details on soon, and
3:48
Speaker 1: we'll be talking about next week. As we have with
3:51
Speaker 1: the rest of our How America Makes Medicine episodes, a
3:54
Speaker 1: little history before we bring our guest Patrick in. Let's
3:57
Speaker 1: take a step back for a second, because the idea
3:59
Speaker 1: that AI could transform medicine isn't actually brand new. People
4:04
Speaker 1: have been chasing this for decades, and honestly, the story
4:07
Speaker 1: of why it kept failing tells you a lot about
4:10
Speaker 1: why this moment feels a little different. Let's take a
4:13
Speaker 1: quick trip back to 1997. That's the year Deep Blue,
4:17
Speaker 1: IBM's chess computer, beat world champion Garry Kasparov. Everybody freaked out.
4:22
Speaker 1: Machines have surpassed us. But if you actually think about
4:25
Speaker 1: what happened, it's kind of a narrow win. Chess has
4:28
Speaker 1: fixed rules, a fixed board, and the computer can see everything.
4:32
Speaker 1: But biology is nothing like that.
4:34
Speaker 3: It's messy. It's incomplete. Nothing's fixed.
4:37
Speaker 1: And it turns out people had already tried to throw
4:40
Speaker 1: AI at biology and hit that exact wall. far earlier
4:44
Speaker 1: than you'd think. Back in 1965, researchers at Stanford built
4:48
Speaker 1: something called Dendril. It's widely considered the first, quote, expert
4:52
Speaker 1: system ever built. Basically a computer program trying to figure
4:56
Speaker 1: out the structure of a molecule just from lab data,
4:59
Speaker 1: like a chemist would. In the 70s, a similar program
5:03
Speaker 1: called Mycin came along that could actually recommend antibiotics better
5:07
Speaker 1: than most doctors could. Except neither of these things ever
5:11
Speaker 1: actually got used in the real world.
5:13
Speaker 3: Why is that?
5:14
Speaker 1: It's because you can't write down every rule of biology
5:17
Speaker 1: in advance. It's too complicated. It took the field about
5:21
Speaker 1: 30 years to really absorb that lesson. So what actually changed? Well, honestly,
5:26
Speaker 1: it wasn't the algorithms. It was more the quality of
5:29
Speaker 1: the inputs, the data. The Human Genome Project wrapped up
5:33
Speaker 1: in 2003. And after that, the cost of sequencing DNA
5:37
Speaker 1: just fell and fell and fell faster than computers themselves,
5:41
Speaker 1: which were also getting cheaper. So suddenly there was a
5:44
Speaker 1: mountain of biological data for machine learning to chew on
5:47
Speaker 1: at an ever cheaper level. Then around 2012, you get
5:51
Speaker 1: this moment where a deep learning approach built by people
5:54
Speaker 1: who weren't even chemists beat everyone else at predicting how
5:58
Speaker 1: chemical compounds might behave. From there, it moved fast. And
6:02
Speaker 1: I think we've all felt that over the last couple
6:04
Speaker 1: of years. But within less than 10 years, AI basically
6:07
Speaker 1: cracked one of biology's oldest and hardest puzzles. Google's DeepMind's
6:12
Speaker 1: AlphaFold2 solved a problem that had stumped biologists for 50 years,
6:17
Speaker 1: predicting the 3D shape a protein folds into from its
6:21
Speaker 1: amino acid sequence. AlphaFold2 would go on to the CASP
6:25
Speaker 1: competition where it performed close to experimental methods like X-ray crystallography.
6:31
Speaker 1: DeepMind later used it to predict the structures of over
6:34
Speaker 1: 200 million proteins and released the database free to researchers.
6:39
Speaker 1: And then AI stopped just analyzing and started actually designing,
6:43
Speaker 1: proposing brand new molecules, not just predicting how existing ones behave.
6:47
Speaker 1: There's even a few drug candidates out there now where
6:50
Speaker 1: AI picked both the biological target and the actual molecule
6:54
Speaker 1: that's shown real positive results in human trials. But are
6:58
Speaker 1: we there yet? This is something that there's been a
7:01
Speaker 1: lot of hype about and something that we absolutely get
7:04
Speaker 1: into with Patrick in the episode. But the blunt truth
7:06
Speaker 1: is that AI made the discovery part faster, years down
7:11
Speaker 1: to months potentially, but did basically nothing to fix the
7:15
Speaker 1: fact that roughly 90% of drugs still fail once they
7:18
Speaker 1: hit human trials. Let's go back to chess and to
7:22
Speaker 1: Patrick's recent substack, which we'll get into, because the more
7:25
Speaker 1: useful part of the story isn't Deep Blue winning. It's
7:28
Speaker 1: what happened after. A style called centaur or freestyle chess
7:32
Speaker 1: took off where humans team up with computers instead of
7:35
Speaker 1: competing against them. And at a tournament in 2005, the
7:38
Speaker 1: winners weren't grandmasters with supercomputers. They were two amateur guys
7:42
Speaker 1: who had just gotten really good at knowing when to
7:45
Speaker 1: trust the computer and when to trust their own gut.
7:48
Speaker 1: Kasparov himself later summed it up, a mediocre human plus
7:52
Speaker 1: a computer plus a good process will beat a great
7:56
Speaker 1: human and a computer with a bad process. Medicine's still
8:00
Speaker 1: catching up to that idea. Multiple studies on doctors using
8:04
Speaker 1: AI tools found something a little unsettling. When the AI
8:07
Speaker 1: gives the wrong answers, doctors don't just fail to catch it,
8:11
Speaker 1: their own accuracy gets worse, sometimes worse than a coin flip.
8:15
Speaker 1: So the problem was never really, is the model smart enough?
8:19
Speaker 1: It's how does a human actually work with this? Which
8:23
Speaker 1: is probably why when US and European drug regulators put
8:26
Speaker 1: out their very first joint rule book for AI and
8:29
Speaker 1: drug development last January, the very first thing they emphasized
8:32
Speaker 1: We're human oversight and good design, not accuracy scores. And
8:37
Speaker 1: this isn't the FDA's first swing at the problem either.
8:40
Speaker 1: Back in January of 2025, the agency put out its
8:42
Speaker 1: first ever draft guidance specifically on using AI in drug
8:46
Speaker 1: and biologic decisions. It laid out a step-by-step framework for
8:50
Speaker 1: companies to prove an AI model's outputs can actually be
8:54
Speaker 1: trusted before regulators rely on them. That draft came after
8:58
Speaker 1: the agency had been quietly reviewing AI-assisted submissions for years.
9:02
Speaker 1: Those numbers had crept up fast, hundreds of drug applications
9:06
Speaker 1: touching AI in some way since the mid-2010s. But notably,
9:10
Speaker 1: that early framework didn't even cover AI used in the
9:14
Speaker 1: way Patrick might consider doing it, the target and molecule picking.
9:18
Speaker 1: Regulators instead were focused on AI touching safety and efficacy decisions.
9:22
Speaker 1: The January 2026 joint principles built directly on top of
9:26
Speaker 1: that and a fuller finalized version of FDA's guidance is
9:30
Speaker 1: expected sometime this year. We can't stress enough, by the way,
9:34
Speaker 1: that guidance out of the regulator sets the tone for
9:38
Speaker 1: how the private sector and entrepreneurs adopt a new technology.
9:42
Speaker 1: And that's why we'll be working closely with the FDA
9:44
Speaker 1: to hopefully get these AI guidances right.
9:47
Speaker 3: And that's basically 60 years.
9:49
Speaker 4: In a nutshell.
9:50
Speaker 1: Rigid rule-based systems that never left the lab, and then
9:53
Speaker 1: a data explosion that finally gave AI something to work with,
9:57
Speaker 1: and a hard lesson that the smartest model alone still
10:01
Speaker 1: isn't the answer. For investors, someone like Patrick, that's the
10:04
Speaker 1: real takeaway. AI clearly has a lot of promise, but
10:08
Speaker 1: there's work to be done to make it truly revolutionary
10:10
Speaker 1: and for patients to actually benefit, which is conveniently exactly
10:15
Speaker 1: where he spent his time thinking. So with that, Let's
10:19
Speaker 1: bring him in and talk about how AI is changing
10:22
Speaker 1: how America makes medicine.
10:24
Speaker 3: Patrick, welcome to the show.
10:25
Speaker 4: Excited to be here.
10:28
Speaker 1: Well, the number one question we get for guests on
10:31
Speaker 1: the show is, how did they get to where they're at?
10:34
Speaker 3: And how did he find KDT? And how is he
10:36
Speaker 3: so successful?
10:37
Speaker 1: So the first question we always ask is, how did
10:41
Speaker 1: you get here? Walk us a little bit through the
10:43
Speaker 1: resume that led.
10:44
Speaker 4: You to KDT. Yeah, it's a winding road of serendipities,
10:47
Speaker 4: I feel like most stories are.
10:48
Speaker 2: But let's see, my first academic interest was computer science
10:53
Speaker 2: and machine learning. So I studied actually neuroscience and CS
10:56
Speaker 2: in undergrad. And that was actually my path to medicine.
10:59
Speaker 2: So I was interested in ML, graduated university in 2012,
11:01
Speaker 2: and kind of deep learning was taking off. And observed
11:05
Speaker 2: then there were very few physicians that kind of operated
11:07
Speaker 2: at this intersection that knew how to code and knew
11:09
Speaker 2: the math behind machine learning models, et cetera, all of
11:12
Speaker 2: which are obviously very much relevant for patient care now
11:15
Speaker 2: these days. So I ended up pursuing an MD-PhD, did
11:19
Speaker 2: my PhD focused on machine learning applications in healthcare, and
11:24
Speaker 2: was on the path more or less to be a physician.
11:26
Speaker 2: Data scientist, didn't really have much of a business or
11:28
Speaker 2: finance interest, but just, this is a serendipitous speech, which
11:32
Speaker 2: I'll kind of skip the story of, but basically stumbled
11:35
Speaker 2: into someone that had basically founded a venture capital firm,
11:38
Speaker 2: North Bond Ventures, my previous firm, Kind of sold me
11:42
Speaker 2: on venture and startups as a good career path for
11:45
Speaker 2: someone with a technical skill set and was building at
11:48
Speaker 2: that time, building out a team focused on technology applications
11:50
Speaker 2: and the life sciences. So actually in medical school, kind
11:54
Speaker 2: of in my town time or as I was born
11:57
Speaker 2: in medical school rotations, kind of worked part-time as a
12:01
Speaker 2: fellow at this firm and just got totally hooked by
12:03
Speaker 2: startups and investing in and building in biotech companies. So
12:07
Speaker 2: after graduating, my MD PhD moved full-time into venture and
12:10
Speaker 2: So fast forward to now, I've been for a handful
12:12
Speaker 2: of years now at a firm called KDT Ventures, a
12:15
Speaker 2: partner here. We are basically all former physician scientists that
12:19
Speaker 2: got disillusioned with medicine at some point along the way
12:22
Speaker 2: and ended up having more fun investing in and helping
12:24
Speaker 2: build healthcare and biotechnology companies. So we have been around
12:28
Speaker 2: since 2017. We invest early stage. So usually first institutional
12:32
Speaker 2: capital into a company. invest across the biotech and healthcare continuum,
12:36
Speaker 2: everything from therapeutics, where we spend a bunch of time,
12:38
Speaker 2: through biotechnology infrastructure, through healthcare technology, et cetera. Yeah, that's
12:45
Speaker 2: kind of the short bio.
12:47
Speaker 1: Well, it's been an impressive one, and KDT is certainly
12:52
Speaker 1: making its name for itself. And one of the things,
12:56
Speaker 1: you know, it's interesting that you have that computer science
12:59
Speaker 1: background because not a ton of people do. There's a
13:02
Speaker 1: lot of that MD, PhD, and if you had come
13:05
Speaker 1: along 15 years earlier, you might not be putting that
13:09
Speaker 1: skill set to work.
13:12
Speaker 3: Do you see those.
13:13
Speaker 1: Intersections relative to maybe when you started at KDT Do
13:17
Speaker 1: you silo these things still that we've got a therapeutic
13:21
Speaker 1: over here and we've got a deep tech company over here?
13:24
Speaker 1: Are you seeing things start to converge?
13:27
Speaker 4: Definitely converging.
13:28
Speaker 2: I mean, one thing I called us a biotech firm,
13:31
Speaker 2: but we're actually at KDC much more of a, as
13:33
Speaker 2: we call ourselves, a science driven venture capital firm. So
13:36
Speaker 2: we invest a lot in biotech, but we do some
13:39
Speaker 2: stuff in non-healthcare industries as well. Things like industrial biotechnology,
13:43
Speaker 2: novel materials, novel chemicals. The idea is that there is...
13:47
Speaker 2: we kind of operate in technical and scientific ecosystem and
13:50
Speaker 2: founder networks. And there we've observed like one of our
13:53
Speaker 2: fund one companies, it's a multi-billion dollar company at this
13:56
Speaker 2: point called Solugen. This is an industrial chemicals company founded
13:59
Speaker 2: by a physician scientist that has PhD in tumor biology.
14:02
Speaker 2: So the point is like many of these scientific disciplines
14:04
Speaker 2: are all intersecting and you can be trained in a
14:07
Speaker 2: biomedical scientific discipline, but actually build a company industrial chemicals.
14:12
Speaker 2: One example that we'd love to say is that the
14:14
Speaker 2: technology for Dippin' Dots, the ice cream that you get
14:16
Speaker 2: at baseball games, that technology has actually been outlicensed to
14:19
Speaker 2: a pharma company for biopharmaceutical manufacturing. So like these technologies
14:23
Speaker 2: go all over the place. Uh, it's all, you know,
14:25
Speaker 2: a complicated web of interdisciplinary, uh, convergence. And that's the
14:29
Speaker 2: stuff that we get really excited about.
14:31
Speaker 4: Okay.
14:31
Speaker 1: I had no idea about that. And now I'm never
14:34
Speaker 1: going to look at Dippin' Dots the same way. And
14:36
Speaker 1: I'm going to be seeing if I recognize proteins when
14:39
Speaker 1: I get my next, uh, batch of Dippin' Dots. So,
14:42
Speaker 1: and I think all the listeners are going to be,
14:45
Speaker 1: you're going to have illuminated them to the link between
14:48
Speaker 1: Dippin' Dots, which also I think had a space tie
14:51
Speaker 1: back in the day, like NASA was.
14:54
Speaker 3: Part of it.
14:55
Speaker 1: So this intersection is happening, this episode, why you're such
14:58
Speaker 1: a perfect guest. And this is part of our How
15:00
Speaker 1: America Makes Medicine series. And we have taken a look
15:05
Speaker 1: at the last 250 years. I don't know that we'll
15:08
Speaker 1: get to the end of the next 250 years in
15:10
Speaker 1: this series, but without a doubt, the hype in our
15:13
Speaker 1: space is about the role of AI. And you are
15:17
Speaker 1: such a guru in this space. And Anthropix announced that
15:21
Speaker 1: they're going to have a company doing this. And even
15:25
Speaker 1: with policymakers, we hear, well, AI is going to start
15:28
Speaker 1: discovering drugs. Give us the 101 on where you see
15:33
Speaker 1: AI and drug development intersecting. What's real, what's hype, and
15:39
Speaker 1: what's still 20 years off?
15:40
Speaker 4: Yeah, such a good and complicated question.
15:43
Speaker 2: I think I believe two things, which maybe is controversial.
15:47
Speaker 2: I simultaneously believe that this is the most important general
15:50
Speaker 2: purpose technology in drug discovery ever. I also think it's
15:54
Speaker 2: completely overhyped in drug discovery, meaning I think this is
15:57
Speaker 2: going to be, it is a critical part of how
15:59
Speaker 2: all of our companies develop drugs. We were very early
16:01
Speaker 2: in investing in computational drug discovery, companies like Dino Therapeutics,
16:06
Speaker 2: Ture Therapeutics, and many others.
16:08
Speaker 4: And we continue to invest heavily in the space.
16:09
Speaker 2: We also invest in what are considered to be non-AI,
16:13
Speaker 2: more sleepy biotech companies, more asset focused, but we think
16:17
Speaker 2: there's a tremendous amount of value there as well. I mean,
16:20
Speaker 2: the way I see it is that it is a,
16:22
Speaker 2: Within drug discovery, there is a ton of potential for
16:25
Speaker 2: machine learning. It's a tool that basically touches every step
16:28
Speaker 2: of the therapeutic development workflow, and I do believe is
16:32
Speaker 2: actually influencing probability success in a significant way. But that said,
16:36
Speaker 2: the kind of absolute or relative odds of success in
16:41
Speaker 2: drug discovery are so low. Even if you take probability
16:44
Speaker 2: success on a preclinical asset up by threefold, it's still
16:49
Speaker 2: overwhelmingly likely to fail. So what I'm getting at when
16:51
Speaker 2: I say it's overhyped, I don't believe that the folks,
16:54
Speaker 2: and I think you see less of this now than
16:56
Speaker 2: you did a few years ago, but there's many folks
16:58
Speaker 2: that say, you know, in five or 10 years time,
17:00
Speaker 2: we're gonna be able to just print, you know, zero
17:03
Speaker 2: shot de novo drug development, take an asset right into clinic,
17:05
Speaker 2: and it's gonna work a majority of the time. It's
17:07
Speaker 2: just not gonna happen. So for us, we think it's
17:09
Speaker 2: an incredibly important technology. We invest across basically every step
17:13
Speaker 2: of the value chain that's using AI for drug discovery.
17:16
Speaker 2: But I think it's going to be something that is
17:18
Speaker 2: like over time an important benefit, but it's not going
17:20
Speaker 2: to be a panacea.
17:22
Speaker 1: One of the things just reading your stuff and I
17:25
Speaker 1: will say you're prolific in putting some thoughts out there
17:29
Speaker 1: that I do think bring the conversation forward. And so
17:32
Speaker 1: if you're not following Patrick on the various channels, you
17:34
Speaker 1: should be because he puts interesting stuff. And one of
17:37
Speaker 1: the things that stood out to me was the biotech
17:40
Speaker 1: of the future from a human infrastructure standpoint could be
17:45
Speaker 1: theoretically much smaller.
17:46
Speaker 3: Will you paint a little bit of a picture there?
17:49
Speaker 2: Yeah, this is the piece that we are. So we're
17:51
Speaker 2: excited about investing in many different applications of machine learning
17:55
Speaker 2: and drug discovery. But I think the thesis that we
17:57
Speaker 2: are most excited about is the possibility that in the future,
18:01
Speaker 2: there will be a single person or a couple people
18:04
Speaker 2: founding a biotech company that will grow to be a
18:06
Speaker 2: billion dollar plus biotech company. And the reason why I
18:09
Speaker 2: think that is now possible is is one, just the
18:11
Speaker 2: independent of AI, the degree to which drug discovery now
18:14
Speaker 2: can be virtualized and externalized using CROs and kind of
18:18
Speaker 2: outsourced expertise. So you no longer need to, you know,
18:21
Speaker 2: the kind of the non-biotech industries, there's this view, which
18:25
Speaker 2: is true, that you want to own capabilities in-house. You
18:28
Speaker 2: want to hire full-time employees in-house because they just care
18:30
Speaker 2: more about the mission. That is true. But in drug discovery,
18:33
Speaker 2: I think that is actually the exception oftentimes. And you
18:36
Speaker 2: can do a lot with kind of part-time or consulting companies.
18:39
Speaker 2: basis employees. But beyond that, I think the tools now,
18:42
Speaker 2: there's so many tools, both in terms of molecular design,
18:46
Speaker 2: whether it's de novo antibody design or molecular optimization for
18:49
Speaker 2: small molecules, both open source models, but also closed source
18:52
Speaker 2: models in many startups that are awesome companies developing tools
18:55
Speaker 2: here that will allow a drug developer to reach for
18:58
Speaker 2: whatever tool is most necessary for their particular task. Whereas
19:01
Speaker 2: historically that may have required some medicinal chemist that has
19:05
Speaker 2: 30 years of experience and is a very expensive hire
19:07
Speaker 2: to you can now do supplemented by many of these
19:11
Speaker 2: tools and software packages. But beyond that, kind of the more,
19:15
Speaker 2: maybe less sexy stuff, but still important stuff of just
19:17
Speaker 2: being able to help you manage a bunch of CRO
19:20
Speaker 2: relationships as a single person. You have like a bunch
19:23
Speaker 2: of program managers that are now essentially agents that can
19:25
Speaker 2: help comb your email when CRO data come back from
19:29
Speaker 2: the latest experiments and help comb those data propose the
19:32
Speaker 2: next experiment, have like an email draft ready for you
19:34
Speaker 2: the next morning without you ever having to do it.
19:36
Speaker 2: So I think there's lots of different ways that AI
19:38
Speaker 2: can enhance the operations of a biotech company, both in
19:41
Speaker 2: terms of the actual R & D, but also the
19:43
Speaker 2: actual operations of the company. It won't happen, this single
19:47
Speaker 2: person biotech company won't happen overnight. but I think increasingly
19:50
Speaker 2: over time, the team required in order to get to
19:53
Speaker 2: a development candidate and even beyond is gonna be much,
19:56
Speaker 2: smaller and therefore the speed and capital required to get
20:00
Speaker 2: to that preclinical and then clinical proof of concept is
20:02
Speaker 2: much lower as well, which I think is a good thing.
20:05
Speaker 2: Just when you think through what is going to best
20:08
Speaker 2: enhance the treatment of disease, whether it's common or rare diseases,
20:12
Speaker 2: I think having more ideas, more therapeutic assets coming from
20:15
Speaker 2: more people and driving down that cost curve to a
20:18
Speaker 2: development candidate is just net very good for both our
20:22
Speaker 2: business of investment, but also ultimately putting new medicines out
20:25
Speaker 2: there for patients.
20:27
Speaker 4: Yeah.
20:27
Speaker 1: Another way of saying it is this should yield more
20:29
Speaker 1: shots on goal, right? And from the investor perspective, I
20:34
Speaker 1: am assuming also this is going to let you place
20:37
Speaker 1: more bets.
20:37
Speaker 3: Is that exciting to you?
20:38
Speaker 2: 100%. I mean, we'll probably get into this. The bottleneck
20:42
Speaker 2: is still overwhelmingly the clinic. And so I think there's
20:45
Speaker 2: going to be a huge– point of diminishing returns on
20:49
Speaker 2: the number of preclinical assets you can generate. But at
20:51
Speaker 2: the end of the day, I think having more and
20:53
Speaker 2: better preclinical assets generated more cheaply, even though we're never
20:58
Speaker 2: going to solve 100% in the near term, the clinical bottleneck,
21:02
Speaker 2: it's still, I think, great for patients for sure.
21:05
Speaker 1: If I'm thinking about this also, this biotech of the future,
21:08
Speaker 1: one of the complaints, and it's going to be a
21:10
Speaker 1: long question, so stay with me here.
21:13
Speaker 3: We go to Japan a lot.
21:14
Speaker 1: Incubate has developed a deep relationship there, advising the government
21:18
Speaker 1: on how to build an ecosystem that attracts U.S. venture
21:23
Speaker 1: capital and how can the country be more entrepreneurial as
21:28
Speaker 1: it relates to biotech.
21:29
Speaker 3: But they have some.
21:30
Speaker 1: Very arcane, might be the wrong word, their labor laws
21:34
Speaker 1: make it very hard to hire and fire. And something
21:37
Speaker 1: we see in biotechs is You get good data, you
21:40
Speaker 1: immediately hire up a ton of people. You get bad data,
21:43
Speaker 1: 50% of the company's gone the next day because that's
21:47
Speaker 1: just the reality of biotech. If this all plays into fruition,
21:51
Speaker 1: should we see a little more stability that we don't
21:54
Speaker 1: have that dreaded day of, hey, some data came back
21:57
Speaker 1: in and we've got to cut half the workforce?
21:59
Speaker 3: Do you see that maybe we can.
22:01
Speaker 1: Have a little more bit of stability or are those
22:03
Speaker 1: cuts further down the line that AI is not going
22:06
Speaker 1: to be protecting them?
22:08
Speaker 2: I think it's an important role, but I think that
22:10
Speaker 2: also what you're describing is very real. I think it's
22:13
Speaker 2: a problem and it's actually symptomatic of a deeper structural
22:17
Speaker 2: issue in biotech venture capital. I actually don't think it
22:19
Speaker 2: has necessarily the most to do with AI. And that is,
22:21
Speaker 2: if you just look at the top line venture capital
22:25
Speaker 2: funding in biotech over the last many years, you would see, obviously,
22:29
Speaker 2: things go way up in 21. Things come way down
22:31
Speaker 2: in 21, 22, 23. It's gone back up. And I
22:36
Speaker 2: think most people think of that as like a good thing,
22:38
Speaker 2: which it is in general. But the thing that it
22:41
Speaker 2: misses is that when you actually look at the makeup
22:43
Speaker 2: of that capital deployed, it is a ton of capital
22:46
Speaker 2: into far fewer rounds than ever before. And what you're
22:49
Speaker 2: seeing is you're seeing way more mega rounds than we've
22:52
Speaker 2: ever seen, like 50, 80, $ 100 million plus rounds, series A's.
22:56
Speaker 2: crowding into many different targets. As soon as a target
23:00
Speaker 2: becomes clinically de-risked, you see 15 different companies spring up
23:04
Speaker 2: all over the world going after this target. When you
23:07
Speaker 2: raise a $ 100 or $ 200 billion venture capital round, the
23:11
Speaker 2: first thing you do is you hire a very experienced,
23:14
Speaker 2: very expensive management team. Now, make-arounds serve their purpose 100%.
23:18
Speaker 2: There are many indications, a late-stage neurodegeneration trial where you
23:23
Speaker 2: need that type of capital for sure. So I'm not
23:24
Speaker 2: saying that there should not be very large rounds, but
23:27
Speaker 2: I think a lot of this is coming at the
23:28
Speaker 2: expense for much more capital efficient lean plans, certainly like
23:32
Speaker 2: novel first-in-class biology, less de-risked stuff, but the high upside
23:38
Speaker 2: in terms of ultimate patient impact. So I think you
23:41
Speaker 2: see a lot of this. If you didn't hire a
23:44
Speaker 2: very large team or overhire a very large team as
23:47
Speaker 2: a result of a mega round, when the data wouldn't
23:48
Speaker 2: work out, you wouldn't have to lay off a So again,
23:52
Speaker 2: I think I'm not trying to argue against a May
23:54
Speaker 2: ground entirely. I think there is a role for it,
23:56
Speaker 2: but I think just needs to be much more of
23:58
Speaker 2: getting back to the way venture capital started, which is
24:00
Speaker 2: that you just kind of a craft bespoke industry, smaller companies,
24:03
Speaker 2: you know, high impact, enterprise.
24:06
Speaker 3: I'll just ask it directly.
24:08
Speaker 1: Is the valley of death growing wider?
24:11
Speaker 3: Is that another way?
24:12
Speaker 1: Is that a corollary to sort of what you put
24:14
Speaker 1: out there about these mega rounds more showing up? Is
24:17
Speaker 1: venture now showing up later and later, which is leading
24:19
Speaker 1: to a wider valley of death?
24:21
Speaker 2: I mean, if the valley of death is reflective of
24:26
Speaker 2: taking new therapeutic targets or new biology for like a
24:31
Speaker 2: new biology that's understood for a disease that could result
24:34
Speaker 2: in a new therapeutic strategy, that is absolutely widening. I mean,
24:37
Speaker 2: you can look at the stats. I forget the specific numbers,
24:39
Speaker 2: but directionally, this is right. Over the last 10 years,
24:42
Speaker 2: the number of clinical trials and funding into drug discovery
24:47
Speaker 2: has stayed consistent or has gone up. But the number
24:50
Speaker 2: of targets with which that we're like going after has
24:52
Speaker 2: gone way down. We're probably down like 10x or more.
24:55
Speaker 2: So that's what I'm getting at. You're hurting into these
24:58
Speaker 2: like single therapeutic targets that are clinically de-risked and people
25:01
Speaker 2: are either trying to compete to become best in class
25:04
Speaker 2: or there is something best in class. They're doing some
25:06
Speaker 2: slight path like extension to be incrementally better, which can
25:10
Speaker 2: be an important therapeutic product for patients. but the whole
25:13
Speaker 2: biotech venture capital industry shouldn't be going after it. I
25:15
Speaker 2: should argue most of the biotech venture capital industry, that
25:18
Speaker 2: is good for pharma. Let pharma, you know, very de-risked,
25:21
Speaker 2: slow bureaucratic organizations go after like, you know, deuterating this
25:24
Speaker 2: molecule to make it like, you know, bimonthly instead of
25:27
Speaker 2: monthly or something. venture capital should be in the business
25:30
Speaker 2: of taking like high risk, high reward bets. And that's
25:32
Speaker 2: where I think the value of death has never been larger.
25:36
Speaker 3: I think that's pretty profound.
25:37
Speaker 1: I think that's something both on your side of the
25:40
Speaker 1: investment table have to be thinking about it. But on
25:42
Speaker 1: our side with policymakers, we've got to be thinking about
25:45
Speaker 1: it too. I mean, we have got to be pushing
25:48
Speaker 1: solutions that close that. I mean, shutting down SBIR didn't help,
25:52
Speaker 1: but it's back on. And so I think that's a
25:54
Speaker 1: place we'll be looking for policies that can either shorten
25:58
Speaker 1: it on both sides if more incentives have to be
26:00
Speaker 1: creative for someone to come earlier stage. But I want
26:04
Speaker 1: to spin back on the AI front, also sort of
26:07
Speaker 1: on the policy front. In my very layman perspective versus
26:12
Speaker 1: your expertise perspective, at some point, AI comes down to
26:16
Speaker 1: quality data inputs going in. And as I've been having
26:19
Speaker 1: these conversations with a lot of folks, Where we're getting
26:23
Speaker 1: that data from, we can't have it be a barrier.
26:25
Speaker 1: I had a very good conversation with the regulatory folks
26:28
Speaker 1: in the UK. And without saying it, they were saying like,
26:33
Speaker 1: we've got the NHS, our data centralized. Whereas in the US,
26:38
Speaker 1: is it going to be very hard for us to pull?
26:40
Speaker 1: If we're thinking of single arm phase three trials that
26:42
Speaker 1: are going to be driven a lot by computational models,
26:46
Speaker 1: do we need to pull the insurers to the table
26:48
Speaker 1: for claims data? Do we need to pull the hospitals
26:50
Speaker 1: for claims data? Is there an opportunity to look at
26:53
Speaker 1: failed clinical trials and pull that data out of there?
26:56
Speaker 1: And frankly, is that going to be really hard or
26:59
Speaker 1: really expensive? Because getting hospitals and insurers to part with
27:02
Speaker 1: that data, is that something where we need to be
27:05
Speaker 1: stepping in so that we can keep pace around the world?
27:08
Speaker 4: Yeah, I mean, for sure.
27:09
Speaker 2: That's got to be I mean, I'm not saying it
27:11
Speaker 2: can be answered or solved with policy, but it certainly
27:14
Speaker 2: has a role to play because, I mean, just a
27:16
Speaker 2: few examples, 100% negative data from a clinical trial is,
27:21
Speaker 2: in some respects, as valuable as positive data. But the
27:24
Speaker 2: sponsor has no incentive, in fact, the opposite incentive to
27:27
Speaker 2: ever actually release those data, right? Like as soon as
27:29
Speaker 2: something comes up negative, oftentimes the sponsor is trying to see,
27:32
Speaker 2: is there a subgroup that's potentially positive or instructive for
27:36
Speaker 2: you know, continuing this trial in a different patient population
27:38
Speaker 2: or subpopulation, like their first response is not like, okay,
27:41
Speaker 2: you know, we got to go back to the drawing board,
27:42
Speaker 2: but we're going to release these data publicly so that
27:44
Speaker 2: folks can learn from our negative result. That's just not
27:47
Speaker 2: the way things work. There's no real, you know, financial
27:50
Speaker 2: incentive to do so. You know, same thing on the
27:53
Speaker 2: FDA side. I think, you know, it was for a
27:56
Speaker 2: brief period of time, but releasing CRLs is a great thing,
27:59
Speaker 2: I think, both for like just humans reading to learn,
28:01
Speaker 2: but also potentially trading models. I mean, there's a bunch of,
28:04
Speaker 2: and this is why I think pharma, has an inherent
28:07
Speaker 2: advantage relative to early stage biotech. They can afford to
28:10
Speaker 2: pay for very experienced, expensive regulatory expertise and consultants to
28:15
Speaker 2: understand about their regulatory strategy with the FDA. If you
28:18
Speaker 2: could open source much of the clinical trial documentation that's
28:21
Speaker 2: required for filing NDAs and BLAs and INDs, if you
28:25
Speaker 2: could open source more of that, you can train AI
28:27
Speaker 2: models to help you understand what should be your regulatory
28:30
Speaker 2: strategy or what should your data package look.
28:32
Speaker 4: Like, et cetera.
28:33
Speaker 2: So this stuff is hard to do because there's not
28:35
Speaker 2: inherent financial incentive for the open sourcing of those things,
28:38
Speaker 2: which is why I say that policy could potentially play
28:40
Speaker 2: a role there. But I think that is as impactful,
28:44
Speaker 2: if not more impactful, generating those types of document type
28:48
Speaker 2: data sets for clinical or regulatory stage companies than it
28:51
Speaker 2: is to develop another protein data bank, which enabled AlphaFold,
28:55
Speaker 2: for example. It's all important, but there's less of a focus,
28:57
Speaker 2: I think, on the former.
28:59
Speaker 3: Yeah, I agree with that less of the focus.
29:02
Speaker 1: And it's something we're sort of trying to call attention
29:04
Speaker 1: to is we are not prepared for the AI capabilities
29:09
Speaker 1: just from a data harmonization and cleanliness standpoint. And I
29:13
Speaker 1: do think this is a place where countries with centralized
29:16
Speaker 1: health governments and trials that, yes, it's in a black box,
29:21
Speaker 1: but they're going to have access to the entire black box.
29:24
Speaker 1: We're going to be seeing if there's a marketplace for
29:26
Speaker 1: this data. And at the end of the day, especially
29:28
Speaker 1: from the insurance claim side, a lot of it's patient data.
29:32
Speaker 1: And so talking with patient groups about how can they
29:35
Speaker 1: seize back some of that.
29:38
Speaker 3: But I think... you've said something else about.
29:42
Speaker 1: Policy that really stood with me, and that is we
29:46
Speaker 1: need a voice from the ecosystem or voices that really
29:51
Speaker 1: are championing this at the senior most levels of government.
29:55
Speaker 1: And you had put something out that really stood with
29:58
Speaker 1: me that we, AI and crypto sort of got a
30:02
Speaker 1: czar at.
30:03
Speaker 3: The White House and David Sachs.
30:05
Speaker 1: And you look around and there's some great people working
30:08
Speaker 1: on biotech and pharma in Washington, but we don't have
30:14
Speaker 1: that star power that certainly David Sachs brought in. Talk
30:20
Speaker 1: a little bit more about your thoughts on what role
30:24
Speaker 1: that person needs to play and what the benefits of
30:27
Speaker 1: it could be?
30:28
Speaker 2: Yeah, I think it's definitely important in my mind. I
30:31
Speaker 2: think it's the star power of David Sachs. The piece
30:34
Speaker 2: I wrote was called Biotech Needs Its Own David Sachs,
30:36
Speaker 2: which I genuinely believe. I think in part, I mean,
30:39
Speaker 2: certainly the celebrity of Sachs helps in his personal relationships
30:43
Speaker 2: with the administration. I think it's more than that. I mean,
30:46
Speaker 2: two things that I think biotech would need. One is
30:49
Speaker 2: like a boots on the ground political operative, like somebody
30:52
Speaker 2: that just knows how to work, understands biotech, but also
30:55
Speaker 2: understands how to interface with people on the Hill, interface
30:58
Speaker 2: across all of the relevant agencies, both scientific, but also otherwise,
31:03
Speaker 2: knows how to work with the administration. It is clear
31:07
Speaker 2: that I think over a while, the white papers and
31:09
Speaker 2: the writing that I and you guys and many others
31:11
Speaker 2: do are important.
31:13
Speaker 4: I don't think it's going.
31:13
Speaker 2: To solve our desire for more change and policy kind
31:20
Speaker 2: of initiative within Washington. I mean, my sense, let me
31:22
Speaker 2: just take biotech today. I think right now, the political
31:26
Speaker 2: will in DC to do some things in biotech shouldn't
31:30
Speaker 2: be stronger than ever before. I mean, one of the
31:32
Speaker 2: things I wrote in that piece is that one of
31:34
Speaker 2: the things that's been frustrating to me is you could
31:35
Speaker 2: take many of this administration's executive orders from the first
31:38
Speaker 2: few months of Trump's presidency and basically just like control
31:44
Speaker 2: F replace like rare earths with biotech or AI with biotech.
31:48
Speaker 2: And like, it would, it would, it would exactly play
31:50
Speaker 2: like it essentially us, you know, the importance of us
31:52
Speaker 2: maintaining our supply chains, the importance of the health of
31:55
Speaker 2: our nation. Like, I just think the same principles apply here,
31:58
Speaker 2: but it just, there hasn't been, I don't think I've
32:01
Speaker 2: talked to a bunch of folks in DC. I know
32:03
Speaker 2: you guys very much do too as well, trying to
32:05
Speaker 2: understand why has there been so much more traction for
32:09
Speaker 2: certainly AI more than anything else, but even like rare
32:12
Speaker 2: earths and some other areas too, than biotech. And it
32:15
Speaker 2: seems like it's much more like those just dominate the
32:17
Speaker 2: conversation right now. It's not anything explicitly anti-biotech. It's just
32:20
Speaker 2: not really occupying the mindshare of this administration. So I
32:24
Speaker 2: think that's what would be required of this person, somebody
32:26
Speaker 2: that ideally knows the administration or at least knows how
32:29
Speaker 2: to interact with them and motivate them and motivate all
32:32
Speaker 2: the people in DC required to you'll get more policy kind.
32:36
Speaker 4: Of inertia going here.
32:37
Speaker 2: There's certainly great stuff in Operation Trial Blazer, kind of
32:40
Speaker 2: on the domestic reform side. There's been a bunch of
32:42
Speaker 2: fantastic things happening at HHS and the FDA, but there's
32:46
Speaker 2: more to be done for sure.
32:47
Speaker 1: Well, if we're going on that there is more to
32:51
Speaker 1: be done, if you could, a question I always love
32:54
Speaker 1: to sort of close out with, and you've been very
32:57
Speaker 1: generous with your time, is if you were waving a
33:00
Speaker 1: magic wand and we have these various policies out there,
33:03
Speaker 1: are there a couple that are at the top of
33:06
Speaker 1: your list, whether it's out of Trialblazer or others that.
33:10
Speaker 3: You just wish this could happen.
33:11
Speaker 1: Tomorrow and you would already be seeing dividends from a
33:15
Speaker 1: policy change?
33:17
Speaker 4: Yeah.
33:17
Speaker 2: So I'll answer your question and then answer a bit
33:19
Speaker 2: of a question you didn't ask, but I think it's
33:21
Speaker 2: an important one. I mean, I think the thing that
33:23
Speaker 2: I would do, and I think everybody in our industry
33:25
Speaker 2: would agree with on the domestic reform side is I
33:28
Speaker 2: would implement almost everything that's in Operation Trial Blazer. I
33:32
Speaker 2: do think that there's a few things in there, which like,
33:34
Speaker 2: I definitely are good ideas. I'm a little bit skeptical
33:37
Speaker 2: that they aren't gonna introduce more complexity. Like there's these QRIs,
33:41
Speaker 2: qualified research institutions that like, it seems like it introduces
33:45
Speaker 2: additional complexity and additional institutions that might make things more difficult.
33:49
Speaker 2: But in general, I do think we do 80% of
33:53
Speaker 2: what is in there. We're gonna cut like 12 months
33:55
Speaker 2: off the time to IND and First in Human in
33:58
Speaker 2: the US. There's other things, one thing that I think
34:00
Speaker 2: is very important that's not in there. I've been thinking about,
34:03
Speaker 2: I would like to fund a company or even build
34:05
Speaker 2: a company that does the following because I think it's
34:07
Speaker 2: really critical too. Even if Trialblazer is maximally successful, we're
34:11
Speaker 2: still going to be 10 months slower than China. Part
34:15
Speaker 2: of the reason why, in addition to structural advantages that
34:17
Speaker 2: China has, many more people than we do, treatment naive,
34:21
Speaker 2: et cetera, they can run both the ethics review, the
34:25
Speaker 2: IRB review, in parallel with a lot of their equivalent
34:29
Speaker 2: of the IND filing and their FDA regulatory review.
34:33
Speaker 4: We do that in sequence.
34:35
Speaker 2: And the reason why you don't have to, but it
34:36
Speaker 2: doesn't make financial sense to go in parallel. Because if
34:39
Speaker 2: something trips up in IRB review, you have to redo
34:42
Speaker 2: all of your IND documentation.
34:43
Speaker 4: That's very expensive. So something should exist.
34:45
Speaker 2: It should probably be a private company, a private company
34:49
Speaker 2: that basically will take on the risk of like, say
34:51
Speaker 2: across 20 programs or companies, allows all 20 to go
34:55
Speaker 2: in parallel. One fails will basically kind of eat that
34:58
Speaker 2: risk in exchange for a margin across all those companies.
35:01
Speaker 2: And now all of a sudden you've just reliably shaped
35:03
Speaker 2: another 10 to 12 months off of the, so that's
35:06
Speaker 2: another thing that I would kind of wave a magic
35:08
Speaker 2: wand and do. But the question you didn't ask that
35:12
Speaker 2: I still think is important to answer is I think
35:15
Speaker 2: any policy, there's more and more panels and events going
35:19
Speaker 2: on in DC and Boston and elsewhere that I've been
35:22
Speaker 2: going to.
35:22
Speaker 4: I'm sure you guys have too, which are great.
35:24
Speaker 2: I love that the biotech industry is as a collective
35:28
Speaker 2: having these conversations to figure out what we should be
35:30
Speaker 2: advocating for. The thing that I worry about, and I
35:33
Speaker 2: think it's the most important question is the domestic reform
35:36
Speaker 2: stuff is easy. Like it's easy to get on a
35:38
Speaker 2: panel and we'll all agree that we should do stuff.
35:40
Speaker 2: We should get our own house in order. We can
35:42
Speaker 2: all pat ourselves on the back. The hard question is, what,
35:45
Speaker 2: if anything, should we be doing in terms of protectionism?
35:48
Speaker 2: It's like a dirty word that people... It certainly makes
35:50
Speaker 2: me uncomfortable. I've never in my life even advocated for protectionism.
35:56
Speaker 2: But I do think that's the hard question to grapple with.
35:58
Speaker 2: And the thing that I worry about, and I'll tell
35:59
Speaker 2: you my view in a second, but the thing that
36:01
Speaker 2: I worry about is that if we don't actually... have
36:04
Speaker 2: a conversation around what we think should be done, if anything.
36:07
Speaker 2: And the answer may be, we may align as an
36:09
Speaker 2: industry that nothing should be done. What's going to happen
36:11
Speaker 2: is that in DC, I do think in the next
36:13
Speaker 2: couple of years, it is probably decently likely that something
36:17
Speaker 2: will be handed down to us with or without our input.
36:20
Speaker 2: And we probably, we're a complex industry. I think we're
36:22
Speaker 2: misunderstood for many reasons. We may not like what is
36:25
Speaker 2: handed down to us from DC rather than having a volunteer,
36:28
Speaker 2: our own approach. I think the semiconductor industry did this
36:31
Speaker 2: very well, right? It's like the SRI, I think is
36:33
Speaker 2: the national semiconductor kind of trade association. They put forward
36:38
Speaker 2: a bunch of policy proposals in the lead up to
36:40
Speaker 2: the coins, sorry, the CHIPS Act in 22, many of
36:44
Speaker 2: which were reflected there. So there's a bunch of opinions
36:48
Speaker 2: in the industry from we should do nothing protectionism to
36:51
Speaker 2: we should add biotech to coins and we should restrict everything.
36:54
Speaker 2: I think what is easier to align on is something targeted.
36:58
Speaker 2: This is probably a much longer conversation, so I can
37:00
Speaker 2: keep it short, but I can expand on it if
37:02
Speaker 2: it's of interest, where Any deal, larger deal that is
37:06
Speaker 2: more than just a simple transfer of, say, access to
37:11
Speaker 2: a basic CRO or even in or out licensing of assets,
37:15
Speaker 2: but is much more about like broader capabilities that are
37:17
Speaker 2: being transferred or trained from the West to the East.
37:21
Speaker 2: And therefore, we may be losing some capabilities here. Those
37:23
Speaker 2: should be at least reviewed, potentially conditioned and maybe even restricted.
37:28
Speaker 2: Those things I think are more, and there's case studies
37:30
Speaker 2: that I've kind of studied over the last 15 years
37:33
Speaker 2: where I think historically there's been those types of deals
37:36
Speaker 2: that have resulted in, I think, some atrophy of capabilities
37:38
Speaker 2: here and enhancing capabilities in China. But I think that
37:41
Speaker 2: is the difficult question, is what we should be doing,
37:44
Speaker 2: if anything, in terms of protectionism. Anyway, I could talk
37:47
Speaker 2: more about it if it's of interest, but it's a hard,
37:49
Speaker 2: important question.
37:50
Speaker 3: It absolutely is.
37:50
Speaker 1: And I think the way you– I'm glad you brought
37:53
Speaker 1: it up in the waving of the magic wand. I
37:55
Speaker 1: think it's the delicate waving that we are sort of
37:58
Speaker 1: struggling with. I mean it's– for several years, it's sort
38:01
Speaker 1: of kumbaya. Yeah, price controls are going to slow innovation
38:05
Speaker 1: where we really don't like them. The IRA causes the
38:08
Speaker 1: pill penalty. We're all on the same page there. I
38:11
Speaker 1: think this question about China has divided in almost.
38:18
Speaker 3: The.
38:18
Speaker 1: Stratifications are somewhat bizarre and also somewhat predictable. I mean,
38:22
Speaker 1: I think if you have more exposure to China, you
38:24
Speaker 1: probably want a little less protectionism. But I want to
38:27
Speaker 1: go a little bit deeper on... A question I come
38:32
Speaker 1: back to is, is having the U.S. market enough? I
38:36
Speaker 1: think about Celengene and I think... From a policy perspective,
38:43
Speaker 1: we did not, as a country and as Congress and
38:47
Speaker 1: multiple administrations, I don't think we embraced cell and gene
38:52
Speaker 1: from a reimbursement, clear policy, FDA-friendly. A lot of great
38:56
Speaker 1: initiatives happened, but I don't think we gave a huge
38:59
Speaker 1: green light saying the U.S. is going to be home
39:01
Speaker 1: to every new cell and gene.
39:05
Speaker 3: During that, markets.
39:07
Speaker 1: Developed elsewhere, capabilities developed elsewhere. Does the fact that the U.S.
39:12
Speaker 1: will likely, for the foreseeable future, be the largest future
39:16
Speaker 1: market of medicines, does that offset some of your concerns
39:20
Speaker 1: of atrophying on the skill development? Look, as long as
39:25
Speaker 1: we're forcing people to come here and to do clinical
39:28
Speaker 1: trial data here, they've got to stand up a new
39:30
Speaker 1: co or something, and they've got to do some stuff
39:33
Speaker 1: in country.
39:33
Speaker 3: Is that enough to offset your concerns or not enough?
39:36
Speaker 4: Yeah. It depends on what you care about.
39:38
Speaker 2: So if you care about access to medicines, there's two
39:41
Speaker 2: things you can be motivated by. You could care about
39:43
Speaker 2: loss of leadership or the US and the West in
39:46
Speaker 2: biotech globally, which I think everybody does and should care
39:50
Speaker 2: about in biotech. I think, to be honest, in Ohio,
39:53
Speaker 2: where I'm from, you talk to somebody in my hometown
39:56
Speaker 2: that's not a biotech person, most of them probably don't care,
39:58
Speaker 2: to be honest. But we should talk about that in
40:00
Speaker 2: a second. Access, everyone cares about. You're going to take
40:03
Speaker 2: away my pills. You're going to take away my life-saving medicine. That,
40:05
Speaker 2: people care about. I think we do have more leeway
40:10
Speaker 2: here because we are the biggest market in the world.
40:12
Speaker 2: You can't become the biggest, you can't be BYD to
40:15
Speaker 2: take the automotive case. BYD is about to become the
40:17
Speaker 2: biggest automotive manufacturer in the world, despite not selling here.
40:20
Speaker 2: It's not possible in drugs, right? You would never be
40:21
Speaker 2: able to do that. But I do think it's a
40:24
Speaker 2: little bit overly simplistic to think that despite, like China
40:29
Speaker 2: does not always act according to their own financial interest.
40:32
Speaker 2: In fact, oftentimes they don't. And we also forget, like,
40:35
Speaker 2: I think there's this tendency in biotech to have a
40:38
Speaker 2: overly myopic vision of just our industry and not others,
40:41
Speaker 2: there could be like, again, just to lay the obvious example,
40:44
Speaker 2: China invades Taiwan. We respond with our own sanctions or
40:48
Speaker 2: whatever else. And then China decides that they're going to
40:51
Speaker 2: restrict or weaponize the supply chains they have access to.
40:54
Speaker 2: They're not thinking myopically about biotech. They think about biotech.
40:56
Speaker 2: They think about rare earths. They think about EVs, semiconductors,
40:59
Speaker 2: et cetera. So it is like one point in many.
41:02
Speaker 2: And so the bottom line in terms of what is
41:05
Speaker 2: the financial repercussions of for China to restrict medicines in
41:08
Speaker 2: the U S which could be punitive because like they
41:11
Speaker 2: wouldn't be selling their medicines in the biggest.
41:12
Speaker 4: Market in the world.
41:14
Speaker 2: It could be fine if they have like a bunch
41:15
Speaker 2: of other things and other industries to offset it. So
41:17
Speaker 2: it's like one piece of the puzzle, I guess, is
41:19
Speaker 2: my point. So I don't think it, I think it's
41:21
Speaker 2: a bit simple minded to assume that just because we
41:24
Speaker 2: have the biggest market, there's never going to be an
41:26
Speaker 2: issue with restricting medicines in the U S. Um, I
41:29
Speaker 2: think the second thing here is that, um, Everyone in
41:32
Speaker 2: our industry, and I think most people, if they just
41:36
Speaker 2: kind of hear out our arguments, should care about US
41:39
Speaker 2: leadership in biotech for many reasons. I mean, one of
41:41
Speaker 2: the big ones is that I think so much of
41:44
Speaker 2: the norms around informed consent and clinical trial regulation and
41:50
Speaker 2: regulatory norms and all the things that go into medical ethics,
41:53
Speaker 2: those flow from the leader of the world in biotech,
41:55
Speaker 2: which has historically been the US. If that becomes China,
41:58
Speaker 2: there's been more of a checkered recent past, in my opinion.
42:01
Speaker 2: for China on those topics. I mean, many of these
42:03
Speaker 2: tragic deaths lately within IITs in China, they're a good example.
42:07
Speaker 2: So I think most people that care about patients everywhere
42:09
Speaker 2: should probably care about the global industry where China is
42:13
Speaker 2: absolutely a superpower in biotech. I think that is net
42:15
Speaker 2: a good thing, as long as it is counterbalanced by
42:18
Speaker 2: Western and US leadership and many other ecosystems in Japan
42:21
Speaker 2: and South Korea and many other places. That's what I
42:22
Speaker 2: care about. I don't want to be misunderstood. I don't
42:24
Speaker 2: care about the US dominating everything and everyone else being
42:27
Speaker 2: service providers, essentially. I think what is best for everyone,
42:30
Speaker 2: including us, is a global industry where there are a
42:33
Speaker 2: few superpowers in biotech that are all very good. What
42:37
Speaker 2: I worry about is that on our current default path,
42:40
Speaker 2: for many of the trends that we've now seen, not
42:42
Speaker 2: just outsourcing many of our low-end capabilities like CROs, et cetera,
42:47
Speaker 2: but a lot of this in-licensing boom and new co-formation,
42:51
Speaker 2: it's just steadily putting on this path of just outsourcing
42:53
Speaker 2: the R & D industry completely to China, where basically
42:56
Speaker 2: we are just a late stage clinical development and commercial
43:00
Speaker 2: engine here in the US and everything else is done elsewhere.
43:02
Speaker 3: I think it's really well said.
43:04
Speaker 1: And I think based on the dozen or people we've
43:07
Speaker 1: had on the show to talk about this issue on
43:10
Speaker 1: all ends of the spectrum, I think you're sort of
43:13
Speaker 1: landing in the Look, there are going to be some
43:16
Speaker 1: cases that require merit or require review, if not potentially
43:23
Speaker 1: merit blocking. And we're going to see how Treasury comes
43:27
Speaker 1: out with this any day now. So I think we're
43:28
Speaker 1: going to have to have you back on to unpack
43:31
Speaker 1: what's in the proposed reg, what your reaction is going
43:34
Speaker 1: to be to it.
43:35
Speaker 3: And I think it's going to be something we continue
43:37
Speaker 3: to talk about.
43:39
Speaker 1: I suppose I would say my fear is actually a
43:42
Speaker 1: little bit the reverse of what you originally said when
43:45
Speaker 1: you brought up China is the domestic stuff we can
43:48
Speaker 1: all agree on.
43:50
Speaker 3: And I agree with that. But in Washington, it's also
43:54
Speaker 3: about political will.
43:55
Speaker 1: And I do worry that the oxygen consumed by China,
44:00
Speaker 1: we do something on China, radical or not, protectionist or not,
44:05
Speaker 1: And it scratches the itch and takes the political will
44:08
Speaker 1: away from codifying trial blazer and having, frankly, not sexy
44:13
Speaker 1: conversations about modernizing our whole drug development process. So I
44:17
Speaker 1: guess I look at it from the opposite side of
44:20
Speaker 1: I absolutely want these FDA things to get done.
44:22
Speaker 3: I want the data readiness for AI. I want all that.
44:25
Speaker 1: And I have a little bit of fear that if
44:28
Speaker 1: we make China just the tip of the iceberg, we
44:31
Speaker 1: forget about everything under the water.
44:33
Speaker 2: That's fair, but I think they are related in an
44:36
Speaker 2: important way. I mean, the first thing to say is
44:38
Speaker 2: I personally do not think all the things that have
44:41
Speaker 2: been done in the domestic foreign sign over the last
44:42
Speaker 2: two years would have happened in D.C. had it not
44:45
Speaker 2: been for the pressure from China. Like, it just wouldn't
44:47
Speaker 2: have been the political will. And importantly, too, just look at, like, again,
44:51
Speaker 2: I study these other industries to understand what we should
44:52
Speaker 2: be doing here. In the semiconductor industry with the CHIPS Act,
44:56
Speaker 2: there's a bunch of domestic incentives for manufacturing here paired
44:59
Speaker 2: with targeted protectionism. So for example, if you were to
45:01
Speaker 2: get money from the federal government to do semiconductor manufacturing here,
45:05
Speaker 2: you have to commit to not standing up new manufacturing
45:08
Speaker 2: capacity in China. So I think that is a thoughtful
45:11
Speaker 2: way of not completely fracturing both countries, but keeping things
45:17
Speaker 2: related such that we can build our own supply chain here,
45:19
Speaker 2: but not destroy both economies.
45:21
Speaker 1: Yeah, and I think that's where I hope to see
45:23
Speaker 1: whatever comes out of this. I think that push and
45:26
Speaker 1: pull of incentives, I just don't want to forget either
45:29
Speaker 1: side of the equation. And I think we'll have no
45:32
Speaker 1: choice but to absolutely have you back. We'll do this
45:34
Speaker 1: next one live and in person to unpack what is
45:39
Speaker 1: proposed and what gets left on the table. But what
45:41
Speaker 1: I also really like that you brought new to the
45:43
Speaker 1: show is some of the specifics about Trialblazer that might
45:48
Speaker 1: need to be tweaked and then– You know, I think
45:51
Speaker 1: it's time for me to hang up the podcasting hat.
45:53
Speaker 1: We go into business and we start this company that's
45:57
Speaker 1: going to mitigate risk of dual tracking IRBs and INDs.
46:00
Speaker 3: I think that's brilliant.
46:01
Speaker 1: And someone listening to this is probably can go spin
46:03
Speaker 1: this up and then come ask you for money.
46:05
Speaker 3: And that's the power of media, I suppose.
46:07
Speaker 4: Please do. That'd be great. That'd be good.
46:09
Speaker 3: Well, we're so grateful. Thank you, Patrick.
46:11
Speaker 1: You've been very generous with your time and we'll definitely
46:14
Speaker 1: be having you back soon.
46:15
Speaker 4: Thank you. This is fun.
46:17
Speaker 1: That wraps up this episode of the Making Medicine podcast.
46:20
Speaker 1: A huge thank you to Patrick for joining us on
46:22
Speaker 1: today's episode. Be sure to subscribe on YouTube and follow
46:25
Speaker 1: us on other channels as we round out the How
46:27
Speaker 1: America Makes Medicine series.
46:29
Speaker 3: Thanks for listening. And as always, keep innovating.