George Magrath: AI-native Biotechs Have Arrived

Ram Yalamanchili (00:04.129)

Hey, George, I want to welcome you back on our podcast. For those who haven't heard or met George, he is a force of nature, amazing individual. I think I'm super excited about all the success he's had since we last spoke, but George, maybe you can introduce quickly what you do, little bit about your background, and then we can jump into some of the topics I'd like to

George Magrath (00:29.816)

Sure. Yeah, thanks, Ram. And it's great to be back and talk to you guys. I'm the CEO of Opus Genetics, which is a gene therapy company developing novel gene therapies for different types of inherited retinal degenerations, which means children who are born visually impaired. And we've worked with Ram and the Tilda team now for a couple of years.

and it's been a great, great experience and glad to be able to talk about it.

Ram Yalamanchili (01:05.421)

Yeah, and so I think I'll jump off into the reason why I was super excited to bring you back here. I think we had our first podcast where we spoke about your mission with Opus. I want to say it's between nine to 12 months ago. So, you know, it's been a little while and I distinctly remember your thesis and how you plan to do this at Opus. And I'll recap here, but I'd love to get

you to weigh in on that. What I remember at that point was you saying, hey, there is this entirely unmet need for developing therapies for these certain class of diseases. They're rare, they're small in population, but they're economically unviable to fund. And it's very hard to build companies in that space, right? And what was the silver lining is you and us talking about how we can bring

AI and some of these technologies to actually make it viable and build a successful case. And we'll see where that went. think our audience will be excited to see where you've taken this thesis. But yeah, I'd love to hear, did I describe that right? And anything else to add there?

George Magrath (02:25.806)

Yeah, no, absolutely. mean, it's so so the bottleneck was not the science, right? The bottleneck was the the ability to organize the resources and and and execute in a way that made sense. You know, the science we have is all from University of Pennsylvania, from Dr. Gene Bennett, who was the inventor of Luxterna, which was the first approved genetic medicine, which was for a type of inherited

eye disease. And that was a decade ago and nothing else was approved since then. And so we're trying to take, know, there are a lot of these things out there. We're trying to take some and get them across the finish line. so, yeah, so since we last spoke, you know, the thesis really was if we could use an AI enabled

platform, along with other things, right? We're being very creative on the regulatory side. We're being very creative on how we do clinical development. But is there a way to really innovate across the board to be able to develop drugs faster, more efficiently, and with the same high quality that's demanded by patients and by regulators? So since we've last spoken, it has taken off, right? So we've...

So the business plan is now fully funded to develop seven of these gene therapies for different kids, for different diseases affecting kids to the tune of approximately $250 million that we've raised in the past year. We've executed on the first program, the LC-5 program. It's a rare condition and affects kids, around 200 kids in the United States.

And we treated the first six of those. And the second patient we treated was a woman in her 30s who had been blind since she was a child. And she had a remarkable improvement in her vision after treatment. She was featured on Good Morning America in November because of how impactful this was for her. We've also started the second program in a disease called BEST Disease. We've treated the first patient there.

George Magrath (04:44.948)

Similar results, a woman who was using a cane, she was in her 60s, she'd been blind since she was in her teenage years, was able to put the cane down, has significantly improved visual acuity, things like that. So really impactful stuff. using the TILDA platform, using AI in particular, has made a big difference in our ability to execute at a high level.

with these programs in a more efficient manner. So it's been great. mean, and happy to dive into some of the specifics on that.

Ram Yalamanchili (05:27.105)

Yeah, I mean, George, first of all, this is like unbelievable. Like just, just everything you're saying is like what I think any, any person would want to wish if you're in like a clinical development carrier, right? You're talking about multiple individuals getting essentially get to see the first time in their, in their lifetime, right? That's what you're essentially saying. Like they're, they're blind blind. It's amazing. and I think I feel very fortunate. I mean, I'm sure all of our team at Tilda, we were

George Magrath (05:37.742)

Yeah.

George Magrath (05:47.212)

in.

Ram Yalamanchili (05:54.154)

So excited that we were able to help. were able to bring these programs to the table and execute on them. think it definitely proves that AI teammates are real. They're actually able to do the real work and deliver value to customers and innovators like yourself. What I think I'd like to understand or just recap is why did you say these type of companies? mean, clearly now it's changed, but you said about nine, nine to 12 months ago before you raised all this capital. know, I think.

The other day I saw that Opus is like in the top five stocks in the biotech index or something. it's unbelievable. mean, you guys have done is clearly amazing. So I guess what my point is that prior to knowing all of this, right? Like nine, 12 months ago, before all the success came your way, why did you say that? And has that changed? Are you seeing a path towards like now these kinds of companies can be founded? Can people can actually think about programs like these?

George Magrath (06:28.878)

That's not a good...

George Magrath (06:49.224)

Sure. Yeah, so the traditional knock on rare disease, right? And the reason that these things are hard to capitalize is because it's, you know, traditionally, drug development is long, expensive, laborious, you know, it just has not had much innovation over the years. It just is very prescriptive and regulated, which you want to some degree.

but in rare disease, you need the ability to be creative, right? there's no reason, you know, when I make a batch of product, it makes 10, basically 10,000 doses. And so for a disease where there's 200 patients, there's really no practical need to be able to, to make three or four batches of that product before you go commercial. So it's like common sense, things like that, that you have to work across the board in order to get the.

timelines down and the costs down to make these things viable. The other aspect to them is that to make these viable, you can do multiple programs. So if you look at what we've done at Opus, we have seven programs and they stack on each other. So if each of the seven programs is treating 2,000 patients, by the time you do seven of them, you're treating 14,000 patients. So is there a way that you can get efficiencies?

by stacking programs. And that's one of the things that we've really used the AI teammates from Tilda for is because these trials are very similar. It's really just switching out which therapy we have in it. And so if you can reduce the cost and time on that end, you know, and make it so that each trial, while it's similar, can be started a little faster than the last one.

on that template, then that's a big win. And that's exactly what we've done, right? I mean, one of the mantras we had when we raised the financing round was that we can develop these things basically in three years for $30 million. And that was, that's kind of the business model that's been funded and we'll see if we can execute. And so far we have, right? I mean, and there's long ways to go, but so far it's working out and it is a template. mean, it definitely,

George Magrath (09:16.952)

can be used not just for the eye, but for other programs as well.

Ram Yalamanchili (09:22.507)

Yeah, and I want to talk about where this is going. think just from my vantage point as well, get your take on it. But just hitting on the learnings from our side, from Tilda's side, right? We understand the heterogeneity of these operations, but there's also homogenous competence to your protocol, various things which you could stack on, like you said, layer them on. And I think the industry problem which we've seen before, AI teammates are like some kind of an AI

George Magrath (09:45.347)

Mm-hmm.

Ram Yalamanchili (09:49.998)

agent coming in and doing this work is you have start and stop. that is, it's just a lot of those problems where you can't really reduce the cost any further. There's always a base cost to like basically starting and stopping. And I think that's kind of been the industry model, right? Like a CRO would, if you go to them, say, I have two studies, but they're almost identical. It's just kind of, you know, letting it on.

To them, it's still two studies. It's still, there's a minimal amount of like commitment of resources which need to come in to be able to like price this. And I think that models really like change with these AIs because you can, for the first time, essentially use these AIs learnings and continuous learning loops to say, okay, I'm going to, I'm going to, just a continuous process, right? I'm going to have the AI do this extra work or the new work. And it's not completely new programmers system extension to what I've been sort of doing all along.

George Magrath (10:40.472)

So that's why the, I mean, that's the root of why the CRO industry was invented, right? Was because if you were a pharmaceutical company and you kept, you really had a resourcing problem, right? Because if you start a trial, have to hire all these people. And then if the trial goes away, now you've got, you know, a lot of people and you want to be able to treat everybody right. And so by outsourcing that to a CRO, they in the...

the original model was that they could arbitrage that labor force across multiple programs at the same time and avoid churn of employees basically at pharmaceutical companies. And so I think that what you're describing with AI teammates is just taking that to the next step, right? It's just, it's where the original model of pharmaceutical companies where you had a ton of internal development and then you would have people

you know, who had a lot of downtime or you'd have to ramp up really fast was that was essentially that logistical problem was essentially outsourced to the CRO industry. And now the CRO industry is taking the next step and pharmaceutical companies are taking the next step to use AI teammates to help with that. it's obvious.

Ram Yalamanchili (11:50.135)

Yeah.

Ram Yalamanchili (11:54.79)

And that's actually interesting, right? Because maybe at a certain time in certain point in history, it was possible to take a resource and have them context switch between different projects and essentially get value. But I think over the course of time, at least my perspective is workflows got so complicated and there's been new things which got added that that sort of context switch has a heavy cost. And so

George Magrath (12:06.563)

Mm-hmm.

George Magrath (12:13.474)

Yeah.

George Magrath (12:19.726)

Totally.

Ram Yalamanchili (12:20.269)

original model doesn't work anymore. It's not how it used to be 10, 15, 20 years ago, perhaps. And I think that's what I'm seeing with essentially how AIs are able to help the CRO industry. As you know, we work with quite a few of the CROs out there. just to be clear, you are also using a CRO. It's just that you're using our platform and the CRO to help this happen. So I think it's interesting that maybe I'd love to hear a little bit about

George Magrath (12:25.571)

Yeah.

George Magrath (12:41.228)

Isn't it?

Mm-hmm. Yeah, totally.

Ram Yalamanchili (12:49.581)

What was the decision-making process you took, of course, considering risk and reward in terms of picking the right CRO? You obviously bet on an AI-based approach with us, with our platform and our AI teammates, but I'm just curious, what's the right ingredient of the CRO which you looked for and ultimately found which helped out?

George Magrath (13:00.92)

Mm-hmm.

George Magrath (13:08.206)

Well, there are a couple of pretty simple things I cared about, right? So number one was I cared about the technology infusion that you guys brought. That's obvious. But number two was I cared about the domain expertise. So I wanted people who knew what they were doing in our world, in the ophthalmology world. And then number three is I wanted a team I could trust, right? And that was a big deal with it too. And so, you know.

The CRO that we chose, Mentra, had a lot of people that I knew and had worked with in the past and trusted. And so those three things really created a really, really nice, you know, nice foundation for this. And look, you know, our company is 30 people and we have, you know, this time next year, we're going to have five active clinical trials, which is two of them will be pivotal, which is an amazing amount of

efficiency for a small group and it's made possible by, yeah, I don't think it would be possible with the original CRO model.

Ram Yalamanchili (14:17.805)

Absolutely, I agree. Yeah. And I think that's what's most exciting, right? I think we are absolutely very quickly entering that phase where I think this is going to become norm. Obviously companies like us are powering a lot of this and it's really fun to build and actually make an impact. I think for us, the big exciting part is like we're seeing that impact from some of earliest customers, including you guys. So very rewarding to be working in this space. I think one thing I'd like to talk about is perhaps like where things are going.

George Magrath (14:32.878)

Mm-hmm.

Ram Yalamanchili (14:45.837)

I'm sure you've seen a lot of focus on new types of foundation models, biology-based foundation models, which are potentially going to be the way to create new types of molecules, new types of target therapies. Without getting into the specific names, there's been a lot of funding which is going into these companies right now. Do you have a sense of like, where is that going? mean, let's just assume a bunch of them are successful. They're able to bring

a high throughput way of discovering new molecules which can potentially cure diseases. I think as you're saying, you have five programs active next year. I can also see some of these companies, I think their roadmaps actually are calling out for 50, 100, 500. I mean, I've seen all sorts of numbers, right? So let me take a look at like, you are in the trenches. I'm just curious what you're doing, dreaming about.

George Magrath (15:33.154)

Yeah.

George Magrath (15:36.623)

So, I mean, it's amazing what they're doing, right? I mean, they're basically turning into a world where there will be no undruggable target in the future, right? I mean, every target that you can think of will be, you know, with the assistance of AI or probably the lead of AI, will be able to figure out how to drug it. That squarely puts the bottleneck on the clinical development.

Ram Yalamanchili (15:46.093)

you

George Magrath (16:05.024)

And I think that we actually are there, right? I think right now there are way more fantastic candidates out there than there is the capability to get them through clinical trials. And I think that that's gonna be solved. It's gonna have to be solved by innovation through technology, also through regulation. I think what the FDA is doing around this has been fantastic so far in the past.

you know, since Dr. McCarrie has been in the job. But I think it's just gonna, it's gonna have to speed up because it's gonna have to be able to keep up with this new pipeline, this extended pipeline of potential products. You know, we have five active clinical programs. We're in ophthalmology. I don't know of any other company our size that has five active programs like this. know, it's just most companies as you,

you guys know are a single asset and the whole team, a team of 30 people is working on that one program. We're gonna have to figure out how to get five, 10 times that number in order to, and it does a couple of things. Not only does it get more science, gets it shot on goal to show whether or not it's effective for people and helpful for people. Gets patients drugs faster, it creates a different model.

in the investment thesis of biotechnology. Right now, biotechnology is typically you bet on a company that has a single asset, that company will run a trial that will read out in a number of years. And it's typically either positive or negative, right? It's a very binary industry. If you have multiple products, it becomes less binary. So at Opus, we've got seven programs, five of them in the clinic. You know, if any combination of

Ram Yalamanchili (17:52.268)

Yeah.

George Magrath (18:02.38)

two or three or even one of those works, then it's a massive return for our investors. And so they're not just betting on one program, they're betting on a matrix essentially. And the risk is very different with that. And a lot of times that risk is mitigated at the capital allocator sector. So that's what the venture capitalists do, right? Is they place

multiple bats, obviously. But if you can do it within a single company and that company can work efficiently enough to be able to execute it, then that's where it's best allocated. The reason that it's not typically been done there is because it's so resource intensive. Like not just capital, like people, bandwidth, you know, all the things that go into development, expertise that you need to get these programs done, that it's a lot to have more than one program.

Ram Yalamanchili (18:48.193)

Hmm.

George Magrath (19:01.908)

And I think that's a big thing that technology can enable, is the ability of a small team to really execute against multiple programs simultaneously in parallel.

Ram Yalamanchili (19:13.421)

Right. Yeah. And as you're saying it, right, I'm kind of having sort of like a deja vu moment because, you know, we are obviously a software company. are, you know, we're bigger than 30. We've got around a hundred people here. But,

George Magrath (19:28.654)

Yeah.

Ram Yalamanchili (19:31.352)

But I could imagine the amount of work we're doing, the amount of scale which we have in terms of customer base, it probably should be about four five times the size. any given, five years ago, that's probably where it would have been. But our generation has been such a prolific innovation in terms of LLMs, being able to do co-gen. And I think it really has a tremendous effect on the software engineering business right now. You can do a lot with the limited number of resources, right?

And you can really focus and you can actually, I think on a net basis, you can be more efficient and much more focused and you can actually deliver better because you just, you're just managing things in much more sharper way. And I'm kind of, I guess what you're also saying is maybe that's what's happening in your business and potentially where it'll go. if you can, on the on-look, it sort of slings, right? Like how many people are actually there? Yeah.

George Magrath (20:21.102)

Totally. yeah. It's all about just getting people the ability to do the real work, right? The real value add work. And then all of the other stuff that comes along with it, trying to take that off their plate as much as you can. And that's a big part of what we're trying to do with Mentra and with Tilda is to just automate away a lot of these.

Ram Yalamanchili (20:31.149)

Let's work out.

George Magrath (20:47.476)

these things that people typically in the past had to spend hundreds of man hours on. It's kind of crazy to think about, but I mean, like TMS, mean, yeah. Yeah. Yeah. And you're right. It's only a part of it, right? So it does drive down cost, which is a great thing, but cost is only part of the program, right? Speed is just as important to me. Like we have got to move fast.

Ram Yalamanchili (20:52.843)

Yeah. Yeah, yeah. And the quality and the rework and all this other stuff which catches up, right?

George Magrath (21:16.616)

And we've done that with you guys, with Stite Startup and things like that. We've seen real world examples now that we have of how this makes a big difference. And then quality is the third leg of the stool, obviously. And the quality is, I think the way I think about it is,

is that you really are taking the quality to a whole different level. Because even if you have people doing quality checks, having the computer do them is faster, more efficient, and quite honestly, it can be better. The computer can pick up things that people miss.

Ram Yalamanchili (22:02.477)

Yeah, you're adding a real time component to all of this, right? You're sort of moving away from discrete monitoring or discrete task completion and execution to continuous. And that is a paradigm shift in the way I think AI helps. So, you know, it enables a lot of interesting models. think we're just scratching the surface of how I see it in QOPS. I think we're just sort of in the second, third innings as far as I'm seeing.

George Magrath (22:31.118)

I would agree with that. think that we've only, I mean, we've used AI teammates in a number of areas, but there's a whole lot more to go. I mean, we've we still have pavement processing, we have some of the CRF work, you know, we still have a lot of things that we can get, that we're looking forward to working with you guys to roll out and not just in future.

Ram Yalamanchili (23:00.459)

Yeah, we're working hard. We'll be there. So I have a couple of questions on where you are as far as like the future, right? There is a lot of, I would say concern about the future of the pharma services industry, CROs included. And, you know, we get this question always like, you know, like, so what does this mean when AIs are doing a lot of this work better, higher quality, faster, potentially cheaper?

George Magrath (23:02.41)

Yes.

Ram Yalamanchili (23:29.749)

And I think I've taken the view, I think you're going with in what you just said, which is when the risk reward changes, the opportunities and the number of shots are going to go, you know, exponential. It was hard to see that. mean, you know, like I really don't, don't see a lot of people just like very quickly internalizing that. But given you understand the space well, you understand ophthalmology, the, the, the void, I suppose, which is like the unmet need, right.

George Magrath (23:40.92)

Mm-hmm.

Ram Yalamanchili (23:59.15)

diseases which have not been yet addressed, what does that look like? mean, are we thinking like 2x the volume, 10x the volume? mean, in a world where AI is increasingly getting more more efficient and we have more productivity in the industry, I would assume, yes, costs might come down on a unit basis, but I would assume the number of units will go much, much higher. It should, right? Like that's what's happened in, least in software engineering, that's what's happened, think, with the number of, you know.

startups which created has gone astronomical. So I'm just curious, like what is the roadmap you think of the future for biotech and number of programs which will be probably initiated as things go to start?

George Magrath (24:28.546)

Black.

George Magrath (24:39.724)

Yeah, yeah. the kind of old axiom of you've got three legs to the stool, you've got time, money, and then quality, and you can't have all three. We've got a decent shot at actually breaking that with technology. Right.

Ram Yalamanchili (24:58.189)

Yeah, absolutely. That is no longer a... I don't think that can be used anymore. We need to throw that out.

George Magrath (25:04.84)

Exactly, exactly. think that that's so I think that what you're going to see is all three of those go up. And when that goes up, I think what I care about the most is that more science gets a shot on goal to see if it actually helps people. what that looks like in the real world right now to us, like I can give our example, is that our people are able to spend their time doing the work that actually is very specific for each

disease and program rather than more routine work. So right now where we've seen our efficiencies is in things like quality checks, like speed to get sites activated, costs have come down, that all is true. But the number of people in my company keeps going up.

And what we are doing with that is really focusing people on the nuances of each disease. So what I want the people that work at Opus thinking about is, okay, how do we connect with those patients? How do we get those patients? How do we get them into the trials? I don't want them thinking, okay, how do we set up the site? How do we get all the infrastructure in place? We want that site to be set up.

Ram Yalamanchili (26:09.197)

Mm-hmm.

Ram Yalamanchili (26:26.945)

Yeah, how do we manage the regulatory completeness and all this like nuanced detail, right? Which an AI can just as easily do better.

George Magrath (26:34.709)

Exactly. we need them, we need them thinking about, what are the exact end points that we need to be looking at? And how are those a little different than other diseases? And how do I get the patient's perspective? How do get the physician's perspective? How do I mold all that together into like the perfect protocol for this particular disease? And so I've got all of our people, all of our people are working on

elevating the quality of the studies by really, really getting deep into it. And that's what people typically do, right? Like that's nothing new. What's new is the ability to arbitrage it over multiple programs in a small company. And that I don't think has been, that I think is novel. That's why most companies my size have one program is because

you typically spend most of your time, you if you're building the infrastructure and you're adding the nuance that goes along with every drug, every drug is like a baby. You know, you have to nurture it. You have to learn about it. You have to mold it. You know, if you, if you are doing that with the team and you have to do all the foundational work too, then you really cut down your ability to divide yourself across multiple programs.

So that's where I think the arbitrage really is.

Ram Yalamanchili (28:03.277)

That's a, I can absolutely see where you're going with this. I think the way I look at it, going back to that question of the new age biotech, biopharma companies with say latest in class foundation model, which is able to find targets and drug them, right? know, let's just say they're able to do that. I think one other skill we're also saying is most of the clean ops space in this industry does not have the ability to focus on more than a handful of drugs at any given time.

more than handful of programs at any given time, right? Even if you had a stack of these like NCEs or drug candidates, which are ready to go into your clinical programs, the scale of both is just not there. But I think you're essentially saying given how you have set up your organization and your operations, that is something you're able to prove at least to an extent of like five programs for 30 people. And maybe that'll expand. We can do like, I don't know, 30, I mean,

George Magrath (28:40.93)

Yes.

George Magrath (28:44.301)

Mm-hmm.

Ram Yalamanchili (28:59.981)

So five programs with a hundred people. don't know. But that's really exciting. I think that's what we need, right? Like to kind of get to that level of innovation.

George Magrath (29:03.682)

Totally.

George Magrath (29:07.308)

Yeah, that's absolutely it is is how do you safely and efficiently get these programs into into the clinical environment to to get patients a shot and and and yeah, absolutely. I mean, you're seeing on a micro scale with us and and it'd be super cool to see and I'm sure there are people out there way smarter than me they're thinking about this to these bigger companies would be if we can do

If we can take 30 people and do five programs to your point, what could you do at Roche or Novartis or Lilly? It'd be amazing. Right.

Ram Yalamanchili (29:43.542)

Yeah. Or another, another opus, right? Like someone, I mean, maybe one of our audience is thinking about starting a biotech and they can imagine a completely new world where they're competing with the Roche. And that's happened, right? Startups can hit well above their weight class and they have, and they have disrupted incumbents. And I think when, when you have these like disruptive technologies come in, whoever kind of adopts them first and runs with them essentially has a disruptive capability. And

George Magrath (30:00.546)

Mm-hmm.

Mm-hmm.

Ram Yalamanchili (30:11.821)

At least we've seen that multiple times in the tech startup space, know, incumbents getting disrupted. And maybe I think one of the things which I'm most excited about in the 10 year timeframe is today we have your top 10 pharma companies. know, it's quite likely in 10 years, this list is very different. Maybe this list is going to be people who, mean, companies we've never even heard of. Hopefully Opus is on there. So I kind of think that is quite a possibility.

George Magrath (30:34.648)

Mm-hmm.

Ram Yalamanchili (30:40.589)

I mean, it's really leveling the field, at least on the clinical operations side. And then you have the LLMs and all the other technologies which may come up with better drugable targets, which are also leveling the field, right? So then it's just about who has the best model, the best operational model to do an end-to-end development and prove value. So it's like we're definitely entering into some sort of like a inflection point, sci-fi era.

George Magrath (30:44.962)

Thank you.

George Magrath (30:51.811)

Mm-hmm.

George Magrath (31:02.328)

Mm-hmm.

Ram Yalamanchili (31:09.185)

like drug development is how I take away from what you're saying.

George Magrath (31:14.008)

Yeah, no, it's very cool. Yeah. And we haven't even dove into the ability to, the ability to use AI in clinical science. We've been focusing mostly on AI teammates, which are amazing, but there's, I think there's a whole nother world out there where

the amazing amount of data that's available on patients and how do you use that data to essentially select patients that are likely to respond to a certain drug? And that I think is another sort of holy grail of AI. We did it a little bit with a paper we published in the American Journal of Ophthalmology last year where we were looking at diabetic patients and

the ability to predict the response of their eye disease to different types of injections in the eye, just based on images from before, right? So if could a computer look at the image of a patient's eye and tell you which drug you should select for that patient. And it turns out, yes, they can, better than people. And so it's just the first salvo at this kind of stuff.

But I'm very interested in that too. think that's a really cool area for some of these LLMs as they figure it out.

Ram Yalamanchili (32:39.469)

Yeah, a lot of avenues where I think we will see disruption. think cleanups, I I obviously be focused on clubs that sort of tends to be very myopic for me. But I think from where you're sitting, it's very clear that there's many, many other areas where there's going to be really interesting disruptions. George, I want to thank you for your time.

George Magrath (32:45.614)

Yeah. Thank you. Yep.

George Magrath (32:56.846)

Yeah, totally.

Ram Yalamanchili (33:02.445)

Congratulations on all the success you're having. We are huge fans of what you guys are doing, your team. I think you've built an amazing company, team, and kudos for being such a, I guess, on the frontier edge, right? We've loved working with you guys, so thank you.

George Magrath (33:22.541)

Yeah, no, likewise. And we appreciate what you guys are doing and look forward to getting some approved drugs with you. Awesome. Yeah, thank you.

Ram Yalamanchili (33:29.934)

Yeah, me too. Yeah. All right. Thank you, George.