Week 4: AlphaFold and Stripe

1. DeepMind made a major breakthrough in biology, solving a problem that scientists had been working on for 50 years. DeepMind, the UK based artificial intelligence (AI) company (acquired by Google in 2014), has said it’s AI system AlphaFold can predict the structure of proteins. The progress has apparently come “decades” before people in the field thought it possible, and is a “once in a generation” advance.

   Consider this: a protein is estimated to fold in 10^300 different combinations. For context, it is estimated that the number of atoms in the known universe is around 10^78 to 10^82. That works out to between ten quadrillion vigintillion and one-hundred thousand quadrillion vigintillion atoms. In 1969 Cyrus Levinthal noted that it would take longer than the age of the known universe to enumerate all possible configurations of a typical protein by brute force calculation. Now, with the help of AI, we can predict the outcome of a problem with 10^300 combinations with a high degree of accuracy in a few weeks with a moderate amount of computing power.

   This could dramatically speed up the discovery of new drugs for disease treatment, and potentially solve issues like finding enzymes that break down industrial waste. This is a triumph of AI and machine learning (ML), and is getting no where near the amount of attention it deserves. Some thoughts and facts:

   • There will be dozens of new biotech companies that will use the AlphaFold software in the coming years to try and discover new drugs.

   • AI/ML is being used to make big breakthroughs in science but it is also used in our lives in ways we don’t think about.

   • Google has used DeepMind to drive YouTube viewing hours per user up by 3-4x, and has apparently driven $15bn in incremental YouTube revenue through its algorithms. One can be sure that Netflix, Facebook, Twitter etc. use AI in similar ways. Yes, some of the smartest people in the world, using cutting edge technology, are working to make you consume more right wing / left wing / whatever content, as per your consumption patterns.

   • Apparently Elon Musk “begged” the founders of DeepMind to not sell the company to Google in 2014, and has called DeepMind his “top concern” for humanity (he means AI in general, but because DeepMind is at the forefront of AI, he has focused on it specifically).

   • What happens if something like this ends up in the wrong hands? Just like there are infinite potential ways for it to help humanity, there must be potential for it cause untold harm (e.g. biological weapons).

   • Who is regulating this? Do regulators even understand it? Do governments allow private companies and individuals to work on uranium enrichment without oversight? Then why allow something like this, which has the same potential for mass destruction, to go unregulated? I don’t know the answer, but this scares and excites me in equal measure.

2. Stripe, a payments platform, has introduced “banking as a service” feature on its platform. It will seamlessly allow companies to “enable [their] customers to hold funds, pay bills, earn interest, and manage cash flow”. Fintech and finance are evolving in dramatic ways. I don’t think this will disintermediate banks (Stripe has banking partners for this product), but surely it is making banking, as a business, more and more commoditized. I think a lot of how this evolves will depend on how regulations (in various countries) move. What is possible in the US may not be allowed in India and vice versa.