“Science is a way of thinking much more than it is a body of knowledge.”
Nobel Price in Chemistry 2024
The Nobel Prize in Chemistry was awarded half to David Baker and half to Demis Hassabis and John Jumper.
Baker, from the University of Washington, won the award for his work in “computational protein design.” Hassabis and Jumper, of Google DeepMind’s artificial intelligence lab, were recognized for their contributions to “protein structure prediction.” Congratulations to them all. Sir Demis Hassabis is truly a remarkable individual and one of the greatest innovators of our generation—he’s still only 48.
AlphaFold in 2020
When DeepMind released AlphaFold in late 2020, we recognized it as a groundbreaking achievement, potentially worthy of a Nobel Prize in the future. We certainly didn’t expect it to gain mainstream recognition as quickly as it has. AlphaFold is an open-source tool/library that leverages machine learning to predict how protein molecules fold based on amino acid sequences.
Scientific breakthroughs are like art: sometimes you instantly recognize an original masterpiece—like AlphaFold. Other times, you're left wondering what the big deal is or what you might be missing. For example, the 2003 Nobel Prize in Economics awarded for the ARCH/GARCH models. In that case, I don’t even see a crowd gathered around the exhibit which suggests there's something to look at. If there are any options traders out there who regularly use or refer to the ARCH/GARCH model, please let us know. One day maybe!
AlphaFold is also an example of technology done right—it empowers scientific researchers, helping them enhance their productivity and efficiency. The technology doesn’t drive scientific breakthroughs, but the people using it do. The master-slave relationship is the right way round.
Like all machine learning techniques that use large neural networks and data to make statistical inferences, AlphaFold is not 100% reliable. Nor is it a current obsession or priority of developers to achieve total accuracy in a field that is not entirely absolute. The ultimate goal is personalized medicine tailored to an individual's unique DNA. There are also legitimate concerns that AlphaFold could be misused, such as the development of biological weapons.
While the proliferation of powerful open-source tools theoretically lowers barriers and increases competition, in practice, the greatest benefits are realized by domain experts with years of accrued knowledge and experience. In the field of bioinformatics, some mistakenly believe that access to tools like AlphaFold and genome sequencing data alone is sufficient to achieve breakthroughs. This is akin to thinking that having a 3D printer with genome data as a blueprint is all that’s needed.
Professional Intuition
To counter this view, we can refer back to the development of the Oxford-AstraZeneca COVID-19 vaccine by Professor Sarah Gilbert and Dr. Catherine Green. The Oxford-AstraZeneca vaccine was not developed using modern mRNA technology but a more traditional viral vector (adenovirus) platform. However, the core principles remain the same. The vaccines were not designed as perfect matches for the SARS-CoV-2 DNA sequence, first published by Chinese scientists on January 1, 2020. Instead, the developers considered many imprecise and unknown factors, such as the human immune system and potential viral mutations, to maximize efficacy and coverage. This part of vaccine development was as much art as science and cannot be replicated by AI experts and data scientists alone.
Better risk reporting tools also don't make one a better portfolio manager.
A couple of quotes from a Times article in July 2021:
“I had an email from a parent of an eight-year-old girl who said she’d seen me on television and didn’t realize that women could be scientists, and now she wants to think about being a scientist.”
So far, so inspiring? Not quite. Gilbert thought, “Well, what on earth have you been teaching her? For heaven’s sake.”
Brilliant! It's funny when superficial displays of empathy are expected of people who are closer to the ‘aspey' end of the spectrum. But she's right!
“And many have, indeed, failed. That they have done so is not somehow a reflection of their lesser skill, but simply the sheer complexity of a human body and what happens when you tweak its immune system. One vaccine trial stopped because it caused people to wrongly test positive for HIV. The vaccine interfered with the tests—astonishing bad luck.”
Subsequently, many have concluded that mRNA vaccines are superior to traditional forms. However, this may be a premature conclusion, as it depends on the course of viral mutations and the most prevalent form at any given time.
Unfortunately, the information domain is dominated by people who either don't believe in science or don't truly understand it, viewing things in absolutes. Religious ideologies are out of scope of the discussion.
