We’re always running out of time, but that’s something Demis Hassabis reckons with more than most.
Visit sportnewz.click for more information.
Hassabis regularly works until 4 AM, and that’s partly the nature of leading AI at Alphabet. The Google DeepMind cofounder also has another company: Isomorphic Labs, the AI drug discovery startup he spun out of Google in 2021.
I profiled Hassabis and Isomorphic earlier this year, and caught up with him recently as the company hit a new milestone: This month, Isomorphic raised a new $2.1 billion Series B, led by Thrive Capital, with the intention of “turbocharging” its next phase. That means scaling compute, generating more data, and building out programs. (Though we covered the raise in Term Sheet, it only seemed right to bring you more of the interview.) So, I asked Hassabis: What can you “turbocharge,” and what will always involve friction in the grueling drug discovery process?
“Wherever there’s friction, we don’t really accept that, for starters,” he told Fortune. “But near-term, the obvious thing is more compute… It’s not the same amount of compute an AGI lab needs, but with more compute, the more experiments we can do, the more inferences we can run.”
As Hassabis thinks about the logistics of the future, he’s also thinking about Isomorphic’s infrastructure and the possibility of automated labs: “I’ve no doubt that, at some point, we will go out and do that, but we want to get the timing right,” he said. “There’s still more research to be done on the automation piece, and on the specific repeatable steps you’d want to automate.”
Drug discovery is, historically, a scientific process that involves a spectacular amount of trial and error. And often, even when a drug beats the low odds of making it to clinical trials, it’s rarer still for a new pharmaceutical to make it to full FDA approval. Isomorphic has yet to bring a drug into clinical trials. Hassabis declined to offer a timeline, but the implication is certainly that the goal is sooner rather than later.
“What I can say is that we’re in pre-clinical with a bunch of our programs, and it’s going very well,” he said. “The new funding, the new partnerships [including with J&J and Novartis]… there’s enough data points if you want to extrapolate how things are going.”
The thing about Isomorphic is that the measure of success won’t be one drug.
“It’s actually about building the platform to solve the overall problem, solving all disease, right,” said Hassabis. “We don’t measure ourselves by one individual program.”
The name Isomorphic is derived from the mathematical principle of two entities that look different but have the same underlying structure.
“It’s biology on one hand, and information science on the other hand,” said Hassabis. “I think biology can be thought of as information. Cells can be thought of as information structures. So, in my opinion, AI is the perfect tool or description language for biology—to understand the isomorphic connection between information structures, and biological structures.”
There’s also a sly nod to the movie Tron: Legacy, Hassabis adds. Isomorphic employees call themselves “Isos,” and in the Tron sequel, the protagonist also refers to Isos as “amazing programs that come and solve disease,” Hassabis laughs. For all his ability to wrangle data and AI, and articulate those futures, there’s at least one thing he’s uncertain about.
“If you asked me, what’s the number one mystery about reality, I’d say it’s time,” said Hassabis. “What is it? It’s kind of crazy that we don’t know, other than a bit of hand-wavy: ‘well, entropy increases in one direction, and it doesn’t the other way.’ That’s not a satisfying answer to the conscious experience of time. It’s the number one thing I want to understand. I also feel that life’s too short and we don’t have enough of it. That’s maybe why I’ve been in a rush my whole life.”
See you tomorrow,
Allie Garfinkle
X: @agarfinks
Email: [email protected]
Submit a deal for the Term Sheet newsletter here.
Joey Abrams curated the deals section of today’s newsletter. Subscribe here.
This story was originally featured on Fortune.com