Life sciences and industry have an inverse problem. While the former are staffed with armies of creative artisans who labor in labs producing bespoke solutions to complex problems, the latter are the exact opposite. They typically employ highly educated engineers, whose task it is to deliver products and services with high reproducibility and structure.
Going forward, this must change. In order to capture the power of computational methods, life sciences must nurture an engineering culture within their ranks. A shift of focus is required towards generating the data necessary to train foundation models and harness the power of AI for drug discovery, diagnostics, and care. The key word here is reproducibility. Check out Alan’s presentation on this topic. According to Alán Aspuru-Guzik, an interesting character who holds chairs in chemistry, computer science, chemical engineering, and materials science at the University of Toronto, the key to future progress in life sciences lies in harnessing the power of Moore’s Law. This can be done by automating the workflow from creative exploration to experimentation. Check his talk for details. He outlines a roadmap to turn chemistry into an engineering science.
For this to happen, people and culture must change. Just look at the average science student enrolled at University today and ask yourself whether they are ready and capable of being exposed to the rigors of an engineering career. I doubt it. My conjecture is that most people choose science because they have a romanticized misconception of the field. They dream of sitting in the lab late at night and discovering something truly impactful. Those types of people usually don’t like to be constrained by equations and highly structured workflows. But that’s exactly what’s needed to drive life sciences into the fast lane of computer science, machine learning, and Moore’s Law.
Recently, I came across two talks by Silicon Valley people, who both emphasize the coming shift in life sciences from bespoke to engineering disciplines. One is Nvidia CEO Jensen Huang at the JPM 24 Pharma conference in San Francisco. Huang talks about in-silico drug design, which means that pathways for drugs can first be simulated and then, only when the heavy lifting is done, actually produced in the lab. Another interesting talk is Vijay Pande’s interview at Software Engineering Daily. Pande is an investor at the Silicon Valley Primus VC firm Andreessen Horowitz and focuses on the crossroads of life sciences and AI. In his talk, Pande emphasizes the importance of a cultural shift in biology from bespoke lab creativity to structured engineering work.
It’s interesting how the problem is the exact opposite in industrial companies. Here, firms are in desperate need of artisans. Take the car industry, where engineers have been trained for decades to work within highly structured processes. In fact, this goes back to Henry Ford, who introduced the assembly line, a production method that doesn’t just apply to blue-collar workers. The assembly line permeates all through the car industry. In fact, the assembly line is synonymous with automotive industry culture. Today, however, the industry needs more artisans. Creative engineers, who can solve problems on the fly, are needed.
But how can you nurture a culture of creative exploration in an industry where the ultimate arbitrator is physics? Nuts and bolts must hold and withstand any forces, whether benign or not. This is not software where, like with ChatGPT or video games, mistakes just get erased by resets.
The solution to this problem is surprisingly simple. You don’t need to change people. What is required is a culture of iteration. Let the engineers do their work. Let them compute, think, and test whatever they do. But make sure they rapidly iterate, try, fail, and fix errors. This is the key. Iteration brings innovation. Compound returns are not a function of time but of iterations. In a culture of iteration, failure is not a bug but a feature.
Iteration requires a reset in corporate culture. Notions of failure, promotion, career advancement, and success require reformulation. New words must be invented. What is a process engineer? A professional who attempts to solve a problem by iteratively fine-tuning a process. There is no right or wrong, only improvement. Failure is not when machines break, but when engineers fail to explain why a certain problem cannot be solved.
Tesla has introduced this type of culture into the car industry. Elon Musk has also revolutionized the space industry by introducing a culture of rapid iteration at Space X.
Conclusion
Industry needs more artisans, and life sciences need more engineers. Both industries are going through a fundamental reorganization driven by innovation and competitive opportunities. Companies must harness the power of AI and data management to migrate their industries towards Moore’s Law. While life sciences need more automation and structure, industry requires the exact opposite. The car industry is a case in point. Here, more artisans are needed to creatively tackle and solve problems.