Scaling enterprise AI: Fireside chat with Eli Lilly’s Diogo Rau and Dario Amodei

Dario Amodei, CEO and co-founder at Anthropic, sits down with Diogo Rau, Chief Information and Digital Officer at Eli Lilly and Company, to discuss building enterprise AI for regulated industries like life sciences. In his role at Lilly, Diogo is responsible for setting the pharmaceutical leader’s AI strategy, including how organizations use models like Claude to power clinical research and drug development. The two discuss Anthropic’s approach to building more steerable and reliable AI for enterprise deployments, our commitment to creating more skills for life sciences use cases, and the importance of building specialized models to power industry-specific solutions. Learn more about what Claude can do for life sciences: https://claude.com/solutions/life-sciences

Dario AmodeiguestDiogo Rauhost

Oct 20, 20255mWatch on YouTube ↗

EVERY SPOKEN WORD

5 min read · 1,262 words

0:00 – 1:00
Why enterprises need AI optimized for truth, not engagement
1. DADario Amodei
  Let's have faith in the pace of progress of the technology. Because if the models get good enough to do it end-to-end a year from now and only then you start deploying it-
2. DRDiogo Rau
  That's right
3. DADario Amodei
  ... there'll be another two-year delay, and that's, you know, that's two years during which all the work that you're doing-
4. DRDiogo Rau
  That's right
5. DADario Amodei
  ... to benefit patients is not happening.
6. DRDiogo Rau
  Hello, everyone. My name is Diogo Rau, and I'm Chief Information and Digital Officer of Eli Lilly and Company. I'm joined here with Dario, who is the founder and CEO of Anthropic. Dario, thanks for joining me today.
7. DADario Amodei
  Thanks for having me, Diogo.
8. DRDiogo Rau
  I know you're, you're spending a lot of time now thinking about how do we- how do you work better with enterprises. What's your enterprise strategy and, and how do you see Anthropic different from other providers?
9. DADario Amodei
  Yeah, I mean, you know, I think we made a number of choices that are different, right? So if I think about the incentives given by consumer AI, they're- folks are in a competition for engagement and growth, right? And so that drives a lot of behaviors of the AI that I think are not ideal from an enterprise perspective.
10. DRDiogo Rau
  Right.
1:00 – 1:58
Sycophancy as a real enterprise risk (and why it matters in life sciences)
1. DADario Amodei
  For example, there's this idea of model sycophancy where the model tells you-
2. DRDiogo Rau
  Yeah
3. DADario Amodei
  ... whatever you say is a good idea.
4. DRDiogo Rau
  Right.
5. DADario Amodei
  Right? And even on the consumer side, that can, you know, cause problems. We've seen stories of people who are like, "Oh, yeah, I've discovered a new fundamental theory of physics," and-
6. DRDiogo Rau
  That's right
7. DADario Amodei
  ... model's like, "That's great." Um, and maybe you don't want it to [laughs] maybe you don't want it to say that. Um, but, but I think, uh, you know, of course, on the enterprise side, the, the prob- you know, the, the problems are, are much greater and clearer with that, where, you know, you really don't, you really don't want the model to say, "Oh, yeah, this drug compound's great."
8. DRDiogo Rau
  [laughs]
9. DADario Amodei
  "You should spend millions of dollars to," you know. I just think this is ... You know, I think your idea's great. I think it's really promising. Like, you, you want truth.
10. DRDiogo Rau
  Yeah.
11. DADario Amodei
  Um, uh, and, and so I think that incentive has led us to design our models in a different way, right? I think it's more compatible with making the models smarter, making them better at a wide variety-
12. DRDiogo Rau
  Yeah
13. DADario Amodei
  ... of economically valuable tasks, and it causes us to put a premium on accuracy and reliability.
14. DRDiogo Rau
  Sure.
1:58 – 2:25
What enterprises value: deeper domain knowledge and economically valuable capability
1. DADario Amodei
  One experiment I, you know, that I, I, that I give to everyone, although it's, it's particularly relevant 'cause I'm talking to you, is I say, you know, let's say I improve the model's knowledge of biochemistry from undergraduate-level knowledge to graduate-level knowledge. You know, if I go to consumers and say that, 99% of them are gonna say, uh, you know, "I, I didn't know what you were talking about before. I don't know what you're talking about now."
2. DRDiogo Rau
  That's right.
3. DADario Amodei
  But if I go to you, like, you're gonna-
4. DRDiogo Rau
  Yeah
5. DADario Amodei
  ... that you care about that-
6. DRDiogo Rau
  We appreciate that
7. DADario Amodei
  ... a lot.
8. DRDiogo Rau
  Yes, yes.
9. DADario Amodei
  Like [laughs] you know-
10. DRDiogo Rau
  That's a big difference
11. DADario Amodei
  ... that's very important. [laughs]
2:25 – 3:00
From general models to “skills” and specialized Claudes
1. DRDiogo Rau
  That, that's exactly right. Well, actually, that gets into something else that you've launched as well, which is skills, right? There are a lot of skills that you want in biology or even just skills, like, as a, as an enterprise, how you wanna, how you wanna operate. Uh, is, is that part of the future for you as well?
2. DADario Amodei
  Yeah. I, I definitely think so. I mean, things ranging from skills to, you know, we're in the process of launching various specialized Claudes-
3. DRDiogo Rau
  Yeah
4. DADario Amodei
  ... which are, you know, in some cases will be improvements to the model itself, fine-tunings of the model. But in some cases it'll be something that looks more like wrapping the model with access to particular types of information.
5. DRDiogo Rau
  Right.
3:00 – 3:07
Example: Claude for Financial Services—value of connecting the right data
1. DADario Amodei
  So when we did Claude for financial services, you know, we, we connect it to a lot of the usual kind of in, in- indices and ratings.
2. DRDiogo Rau
  Right.
3:07 – 3:36
Toward Claude for Life Sciences: model intelligence plus scientific databases
1. DADario Amodei
  Um, and so, you know, you'd be surprised how much just making it easy to connect Claude to those things and kind of use it in a way that's aware of that knowledge is valuable. So I think, you know, we're working on a Claude for life sciences that will be-
2. DRDiogo Rau
  Right
3. DADario Amodei
  ... some mixture of making the model inherently smarter and, and wrapping it with various things, right? I don't know exactly what the analogy will be here, but, like, geez, there are zillions of databases of, you know, proteins-
4. DRDiogo Rau
  That's right
5. DADario Amodei
  ... compounds, assays. Like, you know, you probably want that at the model's fingertips.
3:36 – 3:50
Parting advice: start small, but plan for end-to-end transformation
1. DRDiogo Rau
  Well, any parting advice for those of us that are working in this world of drug discovery and development?
2. DADario Amodei
  You know, I would say there is a temptation, and, and it's, it's ... I think it's hard to avoid starting this way, of, you know, what are the small things we can do with AI?
3. DRDiogo Rau
  Yeah.
3:50 – 4:28
Avoid “hill-climbing” that locks you into today’s workflows
1. DADario Amodei
  Um, like, I ... And in a way, you just kind of have to start there. I think one of my pieces of advice is be very, very ambitious-
2. DRDiogo Rau
  Yep
3. DADario Amodei
  ... in terms of where the models are going. I think you can get caught in a mode where there's an existing process. It has 20 parts. You wanna swap in AI to part 5 and part 12.
4. DRDiogo Rau
  Right.
5. DADario Amodei
  And, uh, you know, uh, that can actually be hard because part 12 has to, you know, intersect with part 13-
6. DRDiogo Rau
  Yeah
7. DADario Amodei
  ... and part 11, which are not being done with AI. And, you know, you look at it and you're like, "Well, the AI models aren't w- aren't where they could do, you know, part 0 to part 20 end to end."
8. DRDiogo Rau
  Right.
4:28 – 5:09
Parallelize readiness: prepare now to avoid multi-year deployment delays
1. DADario Amodei
  But in a year they might be.
2. DRDiogo Rau
  That's right.
3. DADario Amodei
  And so you should start thinking now, don't get too seduced by, oh, we can make these, these little hill-climbing gains-
4. DRDiogo Rau
  Right
5. DADario Amodei
  ... by, you know, doing this part and that part. Let's start preparing to do the whole thing end to end. Let's have faith in the pace of progress of the technology. Because if the models get good enough to do it end to end a year from now and only then you start deploying it, there will be another two-year-
6. DRDiogo Rau
  That's right
7. DADario Amodei
  ... there'll be another two-year delay, and that's, you know, that's two years during which all the work that you're doing-
8. DRDiogo Rau
  That's right
9. DADario Amodei
  ... to benefit patients is not happening. Whereas if you go in parallel, if you start-
10. DRDiogo Rau
  Yeah
11. DADario Amodei
  ... preparing now for the, the large change as the models are getting better, then, you know, you may save years of time.
5:09 – 5:25
Planning principle: two-year projects must assume a different AI landscape
1. DRDiogo Rau
  That's right. So don't do two-year-long projects and, and expect that it's gonna be exactly the same way in-
2. DADario Amodei
  Yes
3. DRDiogo Rau
  ... two years from now as it is today.
4. DADario Amodei
  Yes. If you do two-year-long projects-
5. DRDiogo Rau
  Yeah
6. DADario Amodei
  ... plan for where the AI is gonna ... I mean, that sounds like an obvious thing to say.
7. DRDiogo Rau
  Yeah.
8. DADario Amodei
  But, but I-
9. DRDiogo Rau
  Yeah
10. DADario Amodei
  ... but I think it actually, it actually takes a lot of courage and foresight to do that.
5:25 – 5:34
Closing remarks and thanks
1. DRDiogo Rau
  It does, for sure. Well, thanks a lot for taking the time to chat today. Really appreciate it, Dario.
2. DADario Amodei
  Yeah, yeah.
3. DRDiogo Rau
  And, uh-
4. DADario Amodei
  Thank you
5. DRDiogo Rau
  ... see you again soon. [upbeat music]