How Cradle’s AI Beat 130 Teams to Design a Cancer-Targeting Protein | Eli Bixby, Head of ML

I want to jump into the plumbing of this in a second and talk more about the models and the kind of software and the technology that you've developed here.

I think where my mind goes with what you're currently doing is how you wrestle with the liability and the morality of it.

So if we can put a few different components together, right? Obviously, most of these pharmaceutical companies are incentivized to do two things.

One, make money, and they probably make money by making drugs and products that are the most useful to the end consumer, right? Which bring a positive benefit.

Assumedly so, that's probably the incentive.

When you work with these companies, they pay to use your tool to develop something new and then sell it to the market, right? Assumably so, like it's a positive thing, brings a positive kind of impact on the environment and to people and to the human race.

There's also the opportunity that it goes the other way, right? They develop something that's not necessarily beneficial or only beneficial for a few people.

How exactly do you wrestle with the morality of that?

No, yeah, that's a great question.

I mean, I think...

This is really a question that is, I think, every technologist more broadly should be asking themselves.

How responsible am I for the fruits of my labor and how much control can I exert over that? And I think there are some categories, like we are not going to work on bioweapons, for example.

The...

And more broadly, are there systemic problems in the pharma industry? Absolutely.

I think, honestly, a lot of those problems are problems at the intersection of pharma and capitalism.

And they're problems that are made better if... that it's viable to be a smaller competitor.

I think a lot of those problems are due to the crazy concentration in the pharma industry and the fact that you have this oligopoly.

And so I think the, the, what I hope at least is that we have a business model that is kind of encouraging a little bit of change in that direction within the industry.

And, and that that that for me is, is worth doing.

I think also we get to ensure that a lot of great science gets done along the way and that there are downstream benefits that we may not know about about that science getting done aside from the sort of the product that gets created.

So I think those are that's kind of, you know, to your broader question, how I think about it.

But yeah, certainly I think there are moments when you're like, this system is not behaving the way we'd like it to.

Yeah,

I mean, fundamentally, you're democratizing access to these tools to be able to develop different proteins and drugs and things like that.

Like, how exactly would that work? Like, say, for example, I came to you tomorrow and I'm like, I want to use the wet lab to like run an experiment and design something new.

I want to jump into the plumbing of this in a second and talk more about the models and the kind of software and the technology that you've developed here.

I think where my mind goes with what you're currently doing is how you wrestle with the liability and the morality of it.

So if we can put a few different components together, right? Obviously, most of these pharmaceutical companies are incentivized to do two things.

One, make money, and they probably make money by making drugs and products that are the most useful to the end consumer, right? Which bring a positive benefit.

Assumedly so, that's probably the incentive.

When you work with these companies, they pay to use your tool to develop something new and then sell it to the market, right? Assumably so, like it's a positive thing, brings a positive kind of impact on the environment and to people and to the human race.

There's also the opportunity that it goes the other way, right? They develop something that's not necessarily beneficial or only beneficial for a few people.

How exactly do you wrestle with the morality of that?

No, yeah, that's a great question.

I mean, I think...

This is really a question that is, I think, every technologist more broadly should be asking themselves.

How responsible am I for the fruits of my labor and how much control can I exert over that? And I think there are some categories, like we are not going to work on bioweapons, for example.

The...

And more broadly, are there systemic problems in the pharma industry? Absolutely.

I think, honestly, a lot of those problems are problems at the intersection of pharma and capitalism.

And they're problems that are made better if... that it's viable to be a smaller competitor.

I think a lot of those problems are due to the crazy concentration in the pharma industry and the fact that you have this oligopoly.

And so I think the, the, what I hope at least is that we have a business model that is kind of encouraging a little bit of change in that direction within the industry.

And, and that that that for me is, is worth doing.

I think also we get to ensure that a lot of great science gets done along the way and that there are downstream benefits that we may not know about about that science getting done aside from the sort of the product that gets created.

So I think those are that's kind of, you know, to your broader question, how I think about it.

But yeah, certainly I think there are moments when you're like, this system is not behaving the way we'd like it to.

Yeah,

I mean, fundamentally, you're democratizing access to these tools to be able to develop different proteins and drugs and things like that.

Like, how exactly would that work? Like, say, for example, I came to you tomorrow and I'm like, I want to use the wet lab to like run an experiment and design something new.