James Kotecki (00:01): 

CES. It's where the healthcare and medical community converge to explore the role technology plays in advancing and reforming medicine, healthcare, and consumer wellness. It's where business gets done with more than 14,000 attendees at last year's show focused on digital health solutions. Join us in Las Vegas, January 9th through the 12th. Go to ces.tech/register. This is CES Tech Talk. I'm James Kotecki. The world's most powerful tech event CES 2024 brings the future to Las Vegas January 9th through 12th. Today, we preview the future of medicine with a company you may already be very close to. In fact, you may already have Moderna's COVID vaccine coursing through your veins right now. But the company hopes to feeding a global pandemic is just the beginning, and it's pushing for more healthcare breakthroughs with mRNA and AI. Here to explain those acronyms and what they might mean for the future of medical miracles is Brad Miller, the chief information Officer of Moderna. Brad, welcome to the show. 

Brad Miller (01:19): 

Hey James, thanks for having us. Really excited to talk to you today. 

James Kotecki (01:22): 

Brad, you're the chief information officer of Moderna. You're coming at this from a technology perspective, but I think we do need to understand a little bit of the medical science, the biology, to give us some context in this conversation. So can you start us out with an explanation of what mRNA means and how it was used to actually develop Moderna's COVID vaccine? 

Brad Miller (01:45): 

Yeah, sure. Again, thanks for having me. Really excited today and to start with, mRNA, messenger ribonucleic acid, as I'm sure many people out there know, is a single stranded molecule of RNA that really does just represent the genetic sequence of a gene. The body takes that in and there's this little component in the body called the ribosome inside of cell, and it reads that mRNA and kicks off the process of the synthesis of proteins. Those proteins then go into your body and create an effect, either a protective effect or a effect that helps the body cure itself. And so that at a high level is how more of the medical language would read. The way I read it is it's an information molecule that we're passing the body a message and then the body just acts on that message, much like computer systems in terms of feeding it code, and then it executes a program and the program runs on your machine. And that's how I relate the two together to what I do and what our amazing biologists and scientists do. 

James Kotecki (02:53): 

It makes sense that you would give it that analogy. And of course, the program is to generate a resistance to the virus, right? In this case, COVID-19. 

Brad Miller (03:02): 

Yeah, for COVID-19, yes. But for areas like cancer, it's actually a therapeutic to treat it and minimize it. So it's either protective or therapeutic, is how we see it moving forward. 

James Kotecki (03:15): 

That's interesting. And I know a lot of people first became familiar with even the term mRNA through COVID and because of the vaccines that were being developed. Is it fair to say that COVID was if not the first time that this technology had been deployed in vaccines, the coming out party for popular awareness of this ability? 

Brad Miller (03:36): 

Yeah, certainly it's the at scale moment of the company. But previous to that, the company had been doing a lot of research obviously with the use of the mRNA platform, across other therapeutic and research areas. And then because we're a platform company, we were able to use the mRNA platform that was built for previous research to help with COVID ... Well, to have a massive impact on COVID-19 in terms of the speed at which we delivered that vaccine to market. 

James Kotecki (04:10): 

And this is another area where I think there's some cross-pollination terminology wise between technology and scientific medicine. We talk about being a platform company. Can you unpack what you mean by that term in this context? What does it mean to be a platform? 

Brad Miller (04:26): 

It's one of the triggers that got me really excited about coming to Moderna because the vernacular, the language was identical between how I think about building software systems at scale and how Moderna thought about building medicines. And so if you think about the mRNA platform as the operating system that medicines get built upon, we reuse that mRNA platform for all the different medicines that we would bring to market, with COVID being one of them. But flu is another one and RSV is another one. And those are in our respiratory portfolio, and they all leverage the same mRNA platform. So when we go to develop a medicine to market, we don't have to rebuild the platform, which is why we can move so fast. In addition, because that is the basis of how our medicines work, our therapeutics work, our vaccines work, when you go to manufacturing, we don't have to stand up a whole new manufacturing plant. It's the exact same plant. It's the exact same process by which we run through to build, whether it's COVID manufacturer, whether it's COVID, flu, RSV, or even our cancer therapeutics or our INT business. 

James Kotecki (05:40): 

So the platform is like the physical and scientific infrastructure by which this mRNA code is delivered. And then the difference is in terms of ... You talked about having a different portfolio, the different things in the portfolio, the different diseases or ailments that you're trying to target, are basically just different pieces of code that's delivered through the same way through the same platform. Is that fair to say? 

Brad Miller (06:03): 

That's right. It's specific genetic sequences that respond to a specific biological need in the body. And that's where AI comes into it in terms of how you sequence and find the right sequencing for AGC and U, to be able to affect the body and give the right message to have the right response by your body, to build a protein that's going to go and affect the body. 

James Kotecki (06:26): 

Well, I'd love to get into the AI bit a bit more. Thank you for enabling that transition so seamlessly. So when we talk about what AI is actually doing, am I thinking about technology that's basically just trying a whole bunch of sequences, just throw on a bunch of stuff at the wall and just seeing what sticks, obviously in a simulated environment? What does AI really mean in this context? 

Brad Miller (06:48): 

Yeah, well, we use AI pervasively across the company. But it all started way back when. So it's not like we're not new to the use of AI. It's been foundational to identifying the right sequence to any issue that we're trying to solve. And so it is exactly as you mentioned. There's a lot of permutations of four and the sequencing of AGC and U that can actually come together to have the right effect on the body. And so it's really figuring out that right sequencing and then testing it and then iterating on it and using that data feedback loop back into the algorithm to strengthen the algorithm, and get very more specific about what is the right answer. Which is why a great example of that is with COVID-19, once the Chinese government published the sequencing for COVID-19, within two hours, with the use of AI, we had the answer to what we had to go and produce and manufacture it to help save lives. 

James Kotecki (07:51): 

So say that again. And I remember hearing something about this. There was a really rapid initial state of being able to get some answer here. I didn't realize it was as quick as two hours. So the genomic sequence is published. 

Brad Miller (08:06): 

That's right. 

James Kotecki (08:06): 

And then you type that in, you plug that into your machine and the machine crunches some stuff and two hours later spits out ... Okay, here's the mRNA code that you need to deliver through this platform in order to combat this. That's what happened? 

Brad Miller (08:24): 

That's correct. Now, obviously, it took us 11 months to go end to end and go through trials and make sure safety protocols. But in terms of getting the answer, that's how powerful the platform is in terms of the reuse and the use of AI to get to the right gene sequencing, to create the right action in the body that we wanted, the right message to be delivered. 

James Kotecki (08:48): 

And I don't mean to put you on the spot. What I'm interested in is when we think about ... Well, actually let me ask this question first. Do you think about what you're doing as generative AI? Do you use that term? 

Brad Miller (08:59): 

No. Well, we do do a lot of gen AI, but this specifically is machine learning. This is writing algorithms, having the algorithm learn to get to the right answer, versus processing a ton of data from literature or whatever else from a gen AI perspective, and having a conversation with the data. I do think we'll be able to use a lot of gen AI capabilities from a opportunity of what medicines we could go into by analyzing a lot of the research and documentation. And we've started doing some of that with gen AI. But for this specifically, it's the basis of our machine learning platform that we have for research, that we have taken and extended now across the rest of the company. 

James Kotecki (09:48): 

And it sounds like what you're talking about with machine learning, it's like there is an answer. There is an optimal combination, there is a code to unlock here, and that's what you're doing as far as the mRNA for vaccines. We'll get into- 

Brad Miller (10:01): 


James Kotecki (10:02): 

I wanted to get more into gen AI in a second. But the reason I like to get into these questions of what's on the screen and what's the actual user experience, I definitely don't mean to put you on the spot, but it's more about understanding how people actually interact with AI and what AI can mean in different contexts. Because as you know, I think generative AI has such a hold on people's imagination. Many people became aware of AI as a concept seemingly just about a year ago when ChatGPT first came out to the public. But of course, as you know, AI can mean so much more than that. Have you had to put frameworks in place internally at Moderna for understanding and even communicating what AI is, what different kinds of AI are, and how those operate inside Moderna? 

Brad Miller (10:53): 

Yeah, very much so. Because of the explosion of gen AI, everyone hopped on the AI term of that, which impacts the umbrella of AI and how we talk about its company. And so we had to go through some good education process as well in terms of underneath the big umbrella of AI. Yes, we have machine learning, and yes, there's gen AI. And I would even argue as I was talking about mRNA and how we're trying to define the right sequence of nucleotides using machine learning. We're doing that to get to a finite answer of the right answer of that sequencing of those nucleotides. 


And yet, there's other uses of machine learning that are continuously learning and updating. A good example would be the data that we bring in from the Southern Hemisphere of flu versus the Northern Hemisphere, and what areas are being impacted greater than others. And we have big algorithms that look at that information that are continuously learning, versus getting to a specific answer like we do with our sequencing of nucleotides for our mRNA development. And then that is complimentary to gen AI, which is completely different using large language models to take in large amounts of content and then be able to have a conversation with that content seeking information and more knowledge. And those are just very different use cases in my mind. 

James Kotecki (12:25): 

And I understand that ... I believe half of Moderna's employees are using generative AI on the job. What are some of the interesting things that you're doing with that internally? 

Brad Miller (12:33): 

First, I think it's good to roll back the clock to somewhere around February, March. And with the explosion of open source availability, large language models, what I'm really proud about is the way that Moderna ran towards it, not ran away from it. And we enabled the use of Gen AI to our employee base by actually creating our own emChat. So there's ChatGPT externally. We didn't want our people going there because we weren't clear about what kind of content they were uploading to ChatGPT or how it was being treated within OpenAI. And so we built our own version of ChatGPT called emChat, Moderna chat, internally. We built our own application. We obviously are not building large language models ourself. But thanks to the good open source community, there's a ton of use of APIs to gather the utilization of these large language models. 


And we partnered with OpenAI to have API access to their backend doorway, to their large language models. And so we created our own emChat. And the reason we did this is if you're not using gen AI to help improve how everybody works every single day, you're going to miss the boat on this one. And so we wanted to enable early access, get a lot of learnings from our workforce as to how they would go use it. And so with emChat, we came out with our base version of it and people started playing with it. Everything from making it talk like a pirate at first to actually doing really good work and managing business use cases for us where we remove unwanted work or bottleneck work from our environment. And so it's just been an explosion where we now have over 60% of our employee base using ChatGPT or emChat on a daily basis. 

James Kotecki (14:26): 

And as a leader here in terms of incorporating this into your business, and really it's been, you said roll back the clock to February or March, it really wasn't that long ago. We're talking about less than a year to actually make or begin to make somewhat potentially fundamental shifts in the way that people work and businesses operate. Did you set any internal guidelines or frameworks for how to use it? Obviously people do want to play with it and make it talk like a pirate. Obviously, you did put some standards in place to make sure that you had some more control over it. But how do you manage people's use of it or expectations of it? What kind of best practices or frameworks did you put in place? 

Brad Miller (15:09): 

Luckily, because we've been so entrenched in AI to begin with, and we have such a learning mindset at Moderna. And so we leveraged our AI academy. So we have this notion of within Moderna, it's almost like it's a university of learning inside of it. And we have different academies by which you may learn. We have one which is specific to AI, and we've been doing it for multiple years now in terms of bringing our employee base along of understanding AI. And so what we did is we leveraged that platform internally to teach and learn our employee base about gen AI and about the use of emChat. But then we went beyond that. And you need to go beyond that when you're doing a fairly significant workforce, how you work transformation, so that people can adopt it and understand the power of it beyond making it chat like a pirate. 


And so we put together a small, very nimble, very directed transformation team of two people actually, who went to every single town hall and talked about our version of ChatGPT, emChat, the uses of it. We actually ran a prompt engineering competition, and so we taught prompt engineering. So after you first learned the fun of it, we then moved to prompt engineering. And then we've got from there into deeper use cases that have business impact. And out of that, we started building small versions of our emChat for specific use cases like legal chat, like how to analyze documents from a legal perspective or how to look at information coming in from the medical community. And so that's the journey we've been on. And as they say, I can bring you a technology, but it doesn't mean everyone's going to use it. 70% of the effort of uptake of that is really about transforming how people work. And that's where we've been focused. 

James Kotecki (17:04): 

I am so excited to keep following along with the journey here that you are on, because it sounds like you've really taken a leadership role in putting this out into the company and making sure that people are using it in the right way. Based on what you've seen so far, and obviously this is so hard to predict, but that's what this podcast is for, making fun predictions about the future, where do you see the biggest gains, the biggest potential leaps for where gen AI is not just right now, but where it could plausibly go in the next year or so? 

Brad Miller (17:35): 

When I think about it from Moderna, I break it down into components. So if you start at the beginning of our pipe, from an R&D perspective, it's all about decreasing the time to drug discovery and finding out what is the right drug to meet the need of the science. And so there's a lot of research that goes into that. Our scientists are reading a ton. So how do we help with gen AI, their ability to analyze data, analyze report, analyze research, and get to a faster answer? And then you move into clinical trials. And when you think about clinical trials, it's about going through phase one, phase two, phase three. And those are very costly and very time-consuming and very data-driven. And so how do we squish that time down with having faster analysis of the reports and the data that are coming in? 


It's like it's ripe for a gen AI use case. And for both moving out of research into clinical and clinical into manufacturing, there's regulatory process we must follow. And what's really cool about those is they're very fixed documentation that must be met at a high bar and a high quality. And so feeding them and billing them in real time using gen AI is an advancement we really believe will shorten the amount of time we can get to answers, the right answers. 


And then obviously manufacturing, there's a ton of use for it in terms of the quality there. And then in commercial, analyzing all of the data that's coming in from our customers so that we can obsess about them, our patients. And even pharmacovigilant. How do we respond appropriately? How do we get the data? There's tons of reading work that's in our industry, and there's such an opportunity to condense the time down and enable people to take in more information, the right information. And so that's what we're focused on, those real time AI use cases that enable productivity. Yeah, I think it's going to dramatically change our workforce. If we keep adding people the way we are, the value by which they will add will be diminished. But if we can bring the technology to impact the value in a positive way, I think we'll be able to build a smaller workforce that's more powerful. 

James Kotecki (19:55): 

Well, this has certainly been really fascinating, and I think there are lessons here in what you've built inside Moderna for companies that are even beyond the healthcare and medical space. And as we talk about bringing different industries together and sharing lessons, what is Moderna planning to do in Las Vegas at CES 2024? 

Brad Miller (20:17): 

Yeah, so our president, Stephen Hoge, will be there speaking on the morning of January 11th as part of the digital health track. And he'll be discussing AI and its impact on our pipeline. We're also going to be sponsoring a lounge with digital health space at CES. So I encourage everyone to come check it out, come talk with us. I'll be there as well. Be great to see everybody. And just I'd like to learn what everybody else is doing too, so that I can accelerate our use of AI at Moderna as well. I think when our jobs are all about saving lives, it's not about what's good for me. It's about how we can share as much as we can and really embrace the community aspect of it so that all of us can have a greater impact on humanity, which is ... That's our mission, is how do we have the greatest impact on people with the use of mRNA? 

James Kotecki (21:05): 

Well, I can't think of a better way to wrap up the conversation than what you just said. Brad Miller, chief information officer, Moderna. Thank you so much for being on the show today. 

Brad Miller (21:14): 

Thanks, James. 

James Kotecki (21:15): 

And that's our show for now, but there's always more tech to talk about. So if you're joining us on YouTube, be sure to hit that subscribe button, leave a comment. And if you're listening on Spotify, Apple Podcasts, iHeartMedia or wherever you get your podcasts, be sure to hit that follow button. You can get even more CES and prepare for Vegas at ces.tech. That's ces.T-E-C-H. Our show is produced by Nicole Vidovich and Mason Manuel, recorded by Eric Kirkland and edited by Third Spoon. I'm James Kotecki, talking tech on CES Tech Talk.