The life sciences are bottlenecked by the speed at which they can gather data. Progress will come from speeding the translation from atoms to bits and back again.
Atoms → Bits
In the life sciences, one strategy is ground-up simulations. I hear that OpenWorm, the attempt to simulate the body and brain of C. elegans, is considered a notorious boondoggle. If so, it was an audacious boondoggle in exactly the right direction. The many AI-driven protein folding projects are another example. A third is the recent development of a programming language for biocircuits.
Bits → Atoms
Another strategy is bioprinting and high-throughput roboticized labs. Tissue culturing is a slow, tedious, and delicate process with a hard limit on what’s technically achievable by hand. Automating much of that work will not only free up workers for other projects, it will massively increase the amount and speed of physical data collection. A second example are biobanks, which specialize in large-scale medical data collection and digitization. Being able to order detailed medical data on 500,000 subjects for a few thousand bucks is a great business model.
In my opinion, the problem starts with undergraduate education. We need better advice so that students can advocate for cross-disciplinary training and find some of these ideas for themselves. Right now, the old guard is capturing impressionable students and preparing them for 20th century science.
The life sciences are bottlenecked by the speed at which they can gather data. Progress will come from speeding the translation from atoms to bits and back again.
Atoms → Bits
In the life sciences, one strategy is ground-up simulations. I hear that OpenWorm, the attempt to simulate the body and brain of C. elegans, is considered a notorious boondoggle. If so, it was an audacious boondoggle in exactly the right direction. The many AI-driven protein folding projects are another example. A third is the recent development of a programming language for biocircuits.
Bits → Atoms
Another strategy is bioprinting and high-throughput roboticized labs. Tissue culturing is a slow, tedious, and delicate process with a hard limit on what’s technically achievable by hand. Automating much of that work will not only free up workers for other projects, it will massively increase the amount and speed of physical data collection. A second example are biobanks, which specialize in large-scale medical data collection and digitization. Being able to order detailed medical data on 500,000 subjects for a few thousand bucks is a great business model.
In my opinion, the problem starts with undergraduate education. We need better advice so that students can advocate for cross-disciplinary training and find some of these ideas for themselves. Right now, the old guard is capturing impressionable students and preparing them for 20th century science.