Investing in the exponential: AI in Sciences
The future is here: AI in Biology, Chemistry and Physics.
Image generated by DALLE
One thing is clearer than ever: AI will continue to penetrate all areas of our lives — from automating manual tasks to developing treatments for previously uncurable diseases and discovering breakthrough materials for a cleaner future.
As someone who has been investing in AI and deep tech for several years, what excites me most is founders leveraging the latest breakthroughs in AI to solve hard, non-obvious problems with a transformative impact.
Applying Generative AI toward scientific discovery, i.e., Biology, Chemistry, and Physics, is exactly one of those areas.
Last year, Microsoft Researche released a paper titled The Impact of Large Language Models on Scientific Discovery: a Preliminary Study using GPT-4. This paper explores the role of GPT-4 in scientific research across various fields and evaluates the model’s proficiency in understanding concepts and performing tasks in these domains. It also provides a helpful framing for what AI for Scientific Discovery comprises, including but not limited to drug discovery, biology, computational chemistry, materials design, and partial differential equations (PDEs).
Source: Microsoft (Microsoft Research AI4Science and Microsoft Azure Quantum) Report on The Impact of Large Language Models on Scientific Discovery: a Preliminary Study using GPT-4.
3 reasons why I see breakout potential in this space
Scarce Talent: Top talent equally conversant in AI and Biology, or AI and Chemistry, is extremely rare. Scarcity creates value, and these teams are able to create a moat purely by the nature of their unique expertise. Strong multi-disciplinary teams can productize and commercialize their innovations within accelerated timeframes, beating out their competitors.
As such, I believe technically differentiated teams prioritizing domain expertise will build generational companies in this space.
Many of these teams are leaving leading big tech labs, research institutes and industry to start companies in this space. Others have already begun executing a strategy to win in the market.
Examples of best in class teams building companies in this category are Orbital Materials, founded by DeepMind alumnus Jonathan Godwin, and Nabla Bio, founded by Surge Biswas, author of the 2019 UniRep paper that pioneered the application of language models to protein engineering, & Frances Anastassacos, Ph.D. in Biological and Biomedical Sciences from Harvard.
Increased Industry Appetite: Industry is beginning to mature in its appetite for AI solutions. Life sciences, for example, has been a pioneer of sorts in adopting AI solutions.
Today, pharmaceutical companies have existing discovery and joint development arrangements to work with AI companies, enabling early-stage AI companies with best-in-class talent to test and scale their solutions. Examples of these agreements include Genesis Thereapuetic’s $670M collaboration with Eli Lilly and Aspect Biosystem’s $2.6B deal with Novo Nordisk.
Energy, industrial, and infrastructure companies are beginning to follow suit. An analogous example is Occidental Petroleum’s purchase of carbon capture company Carbon Engineering for $1.1B. I am confident we will see more, and the creative business models of the Life Sciences industry will serve as a good precedent.
Exponential Value Creation: I believe that AI applied to the sciences is poised to create some of the largest value categories in the future. Companies in this space will be category creators, so it isn’t easy to wrap our heads around what these categories will be and how to value them. But, to put some numbers out there, Schmidt Futures pegged the bio-economy between $4T to $30T globally. These are trillion-dollar industries waiting to be created, and the work has already begun.
Today, we’re at a point where the enabling technologies are in place, and precedents have been set. Top labs like Google DeepMind have set the stage with their significant contributions to AI-driven scientific discovery, with models like AlphaFold focused on protein folding and GNoME focused on materials discovery, amongst others.
The timing to build these companies is better than ever. While exciting, these businesses are complex, capital-intensive, and risky.
But where’s the fun if the combination of human intelligence and computer intelligence can’t solve previously unsolvable problems?
I’ll close with a quote from Radical portfolio company founder Charles Fisher (CEO of Unlearn). He says:
“Society will get a bigger lift from creating AI to solve problems human beings can’t. That is the domain of narrow superhuman intelligence.”
“While there’s a lot of AI research on automating things humans do pretty well, there’s much more to be gained by using AI to solve unsolved problems. I think we’ll soon find that some of the problems we thought were so complex that they are impossible aren’t that difficult for a narrow superhuman intelligence.”
If you are building or working at the intersection of AI and Sciences, please reach out to connect, provide feedback, brainstorm, and collaborate!
3 resources for additional reading:
Google DeepMind’s podcast covers much of the organization’s contribution to accelerated scientific discovery.
Microsoft Research Report on The Impact of Large Language Models on Scientific Discovery: a Preliminary Study using GPT-4.
University of Toronto’s Acceleration Consortium.
It's always interesting to see the kind of content Venture Capital writes in service to their trending industries. Granted Toronto will be a major hub of AI at the intersection of Healthcare. If you want to write a guest post about that on my AI Newsletter, let me know.
Sanjana - nice overview. totally agree. have a fusion solution for the "biomarker problem" with transformative implications. Kevin - kevinh@liquidbiosciences.com