Individual computational biologists will replace entire pharma companies.
Pharma alone is 1.5% of global GDP, and healthcare as a broad category is closer to 10%. Crypto will enable a (highly) profitable decomposition of pharmaceutical problems (and companies) into maths and software problems. Om nom nom.
Two recent innovations (PCR and Crispr) allow scientists (and mathematicians) to read and write DNA, like computer source code.
The flywheel and Wright’s Law are hard at work, and the cost of sequencing a complete human genome is approaching $100 today, down from $25mm in 2006. Meanwhile, for better or worse biohackers are using Crispr to edit the DNA of their own, living cells.
Since we can read and write DNA, we can treat genetic diseases as math and logic bugs in software, and we can apply computer science and mathematics techniques to find solutions. Covid gives countless examples of AI/ML being applied to understanding the virus, vaccine design, predicting drug and vaccine effectiveness, predicting potential mutations and their behaviour, and more.
Imagine having a rare genetic disease, something too rare for pharmaceutical companies to focus on.
In the near future, a genetic disease is a software problem, that can be solved with data, maths and insight.
All you need is a single computation biologist (sitting on a beach in Thailand, sipping her post-dive Chang) with access to the data. With the right data, algorithms and insight, she will take your DNA and synthesize a 1 of 1, custom Crispr therapy that you hope will cure you of your disease.
The data to save you exists. There are databases with many other human genomes, annotated with longitudinal health and treatment data. There are studies and trials of previous therapies.
In the 2020s, the data often existed somewhere, sitting in a file in a server somewhere in the world. But it wasn’t available, either because regulations kept things private (nobody wants their doctor selling identifiable medical records) or because there wasn’t a market for the information (so neither a “price” for it, nor an easy way to share it).
But new flywheel technologies (AI and mindbending maths with names like homomorphic encryption and zk-SNARKs) allow machine learning models to be trained on datasets, without without revealing those datasets. Those who hold vast troves of sensitive data can now make it available, without concern of private sensitive information leaking, and can monetise this data. This was an economic inflection point. These technologies came of age in the mid 2020s and accelerated another wave of AI improvements because of the orders of magnitude more data that was made available, and the many more people who could build atop these new data sources.
The program our computational biologist writes is the aggregation of many pieces, some she wrote, some open source, some are smart contracts that are paid directly. There’s the clinical trial data (used to train the ML model) that’s dynamically selected and licensed based on your DNA. There’s the “DNA Printer Driver”, that converts the code of your new therapy into the instructions that the DNA printer will use to make the thing you’ll inject into your body. There’s the homomorphic wrapper you apply to your DNA, so that you’re never revealing anything too private. All these other bits of code (and by extension, their creators) get paid, predictably, when you pay the biologist.
You upload your DNA and some Bitcoin, and in return you receive a file (and a few days later, a vial) with your custom gene therapy.
After the FDA collapse in 2030, the USCF Computational Biology lab released StaySafe, a free and open source project that used machine learning to predict the safety of any particular therapy with any particular patient DNA. This project quickly became the biggest open source project in the world, with more contributors than the Linux kernel.
Before their collapse, FDA, CDC, WHO and others fought tooth and nail against ideas like this. “People will hurt themselves! What about scams and ponzis? We can’t let the hoi polloi make decisions about important things like genetic augmentation.” In the end, the societal value provided by regulators is fundamentally trust, and the regulators neither learned nor recovered from covid.
In the place of regulators, cryptographic technologies like zk-SNARKS provided the building blocks necessary for a whole-scale re-invention of trust. You only think your computational biologist is diver in Thailand because of her profile picture. Maybe she’s one of the many nomads from the great migration of the 2020s, but you’ll never really know. Instead, when you look for her reviews, what you find are cryptographic proof that other people have used her therapies before, recommend her strongly, and you might even see your genetic similarity vs each of the reviewers.
The 2020s were a time of great change, with entire concepts like trust requiring reinvestigation: what is trust in a pseudonymous world?, we wondered.
Technologies like zk-SNARKs, which allow you to cryptographically prove an assertion (e.g. “I have a valid driver’s license. I have brown eyes. I have a Y chromosome.”) without revealing identity are a fundamentally new technology, and an important building block in the re-imagining of trust in a pseudonymous world.
The vial arrives, and it deserves a moment’s reflection. There’s no FDA logo to speak of, no pharma company either. Nobody to sue if it doesn’t work. There’s only a tamper-resistant seal with a hologram and a QR code on it. You scan it and it verifies this is what you’ve ordered. Your phone’s flash taps out a message, and nano-particles in the fluid confirm the DNA printer’s cryptographic signature.
You fill the syringe, look at it, and smile, thinking of your grandparents era, when drug costs (let alone rare drug costs) could bankrupt people, and their faith that - on balance - the FDA was keeping them “safe”. Provided the illusion of safety, instead of the raw data itself, accepting glacial progress as “how it’s always been.” Quaint.