There is so much more I don’t know now than when I started as a portfolio manager twenty-four years ago. I don’t wrestle with this fact, I love it. Research is my favorite part of the business. I learned curiosity beats conviction.
Taking a deep dive into something I know nothing about is easier than ever before, for two reasons: 1) all the data is bigger, and 2) original sourcing is closer.
Nowhere are those two technological advances going to make a greater impact than in science, which is about to change the world and make it a little better once again.
I stumbled into some research down a rabbit hole that started from a completely different project I also knew nothing about. That is the beauty of embracing ignorance – questions compound and lead to the most amazing places you never knew existed. I don’t know where this one will end, but it’s impossible to not be excited about the planet shaking for good, so I wanted to share some notes. To be clear, I am not a subject matter expert. The only reason I write this stuff down is to make certain I read more, learn better, and to open collaborations with razor sharp people I didn’t know.
“If we knew what we were doing it wouldn’t be called research, would it?” -Albert Einstein
Before we shatter its record pace below, Moore’s Law deserves some love for all of its remarkable unintended consequences. Gordon Moore predicted in 1965 how fast computing power would accelerate and costs decline – exponentially, for many years. Moore’s Law allowed for technology innovation to explode in unimaginable directions, ever since.
Moore’s Law remains the most accurate long-term forecast I’ve ever witnessed, with more profound moon shots launched off of it than all the rest of predictions combined, so its importance cannot be overstated. New careers, companies and entire industries were born. A generation of students dreamed of being involved in some way, one day.
Some of the brightest minds realized the best way to leverage that increased speed and decreased cost was to build platforms. I am neither a technologist nor scientist, so I call it a sandbox. There is no better way to see something new be discovered than to let all ages, of all the smartest people, from all different backgrounds jump in the same sandbox. Throw some new tools in there and see what the most creative builders, who play well with others, can come up with.
Three guys wanted to make video games and multimedia more enjoyable to play on computers. They started a company called Nvidia to make a graphics chip. They kept experimenting and got pretty good at it. In 1999 they made the first GPU (graphics processing unit).
At the Game Developers Conference the following year, Bill Gates revealed Microsoft’s first Xbox, and Nvidia’s GPU was inside.
The groundbreaking graphics features started multiplying, as did the uses for designing more than just video games. Nvidia was developing different GPUs for serious research challenges, at the same time it won an Emmy for their role in the entertainment industry. Some GPUs were being used in drug discovery and medical imaging, while others were being placed inside Apple’s first MacBook and also Android tablets. GPUs were reducing PC laptop power consumption by 30x, at the same time they powered up the first 3D ultrasound for healthcare. Robotics and autonomous vehicles used their GPUs, while almost every Academy Award for “Best Visual Effects” included other GPUs.
More recently, the most powerful GPUs are being used for deep learning and A.I. discoveries, accelerating the world of data science. Nvidia offers an open-source platform for its widest ever sandbox to play in.
Startup Garages Are Now Labs
Genome sequencing costs have been plunging and new medical tools keep getting added downhill. If the solutions hatched from Moore’s Law were capable of adding unimaginable technology advancements, it will be even harder to imagine what the exponential improvements in healthcare might be. Curious, I looked at a couple and have never been more astounded.
A little startup called Parabricks from Ann Arbor, Michigan was playing in Nvidia’s GPU sandbox. New scientific data for everybody is being generated faster than anybody could analyze it, causing bottlenecks. Nvidia observed at this rate, in less than five years we will generate 20 ExaBytes of data – that’s more than Twitter, YouTube, and all of astronomy combined.
Parabricks used Nvidia’s GPUs to build a faster analysis of whole genomes. Then, they used Nvidia’s sandbox with more tools to accelerate the analysis of sequence data by more than 50 times. All three billion base pairs in human chromosomes previously took many days to analyze. Parabricks got it down to less than one hour.
Nvidia bought Parabricks last year. It could have been anybody in that sandbox who figured it out. And, no doubt, somebody different will discover something better or might just build on top of it, someday soon. Remember, nobody knew what was going on in those now famous garages of technology startups that thanks to Moore’s Law changed lives forever. Since those tech garages are now science laboratories, the results will be bigger changes for even greater uses, so I call it MORE’s Law.
From Nvidia’s Developers Site: “Users experience orders of magnitude faster analysis for the whole pipeline, while generating 100% equivalent results. All the tools are containerized and available on major public clouds or on premise for wide availability.”
Parabricks was proud to join the sandbox operators. “We plan to provide accelerated tools for major established genomic analyses, continue supporting all major versions of these packages in the long term, and include invaluable enterprise grade support to many major institutions. Now, as part of Nvidia, we can scale our efforts to provide these analyses to everyone looking to explore new and impactful ways of analyzing genomic data to find new insights,” said Mehrzad Samadi, Parabricks co-founder and CEO.
Nvidia offered a free license to use Parabricks for any researcher in the world fighting Covid-19. After that giant problem is solved, it will be back to the sandbox for more personalized medical breakthroughs affecting more lives. MORE discoveries than any of us can imagine will become law.
“If you want to go quickly, go alone. If you want to go far, go together.” -African Proverb
Bigger Than Texas-Sized Sandbox
Moving several miles southwest in the sandbox, Jason McLellan works just off the main drag of campus at the University of Texas in Austin. I walked by this spot hundreds of times as a student, never stopping, focused on a career in investments. There was a pay phone across the street I used to mash numbers corresponding to letters to dial up stock tickers for my little account. I told Jason that my strongest hunch down this rabbit hole is the next generation of the brightest minds are dreaming of jumping in the science sandbox, more than technology or finance. I know one, and am heavily biased by a young lady much smarter than I was. My daughter is studying biomedical engineering at LSU’s Honors College.
Jason McLellan is a Professor of Molecular Biosciences at UT-Austin. His team has the potential to be one of the most important you have never heard of. So, I wanted to fix that (and I needed a Longhorn team to believe in again). The McLellan Lab put a bullseye on the back of the enemy causing Covid-19, and also one on the front, both sides, top and bottom.
To start with, many copies of the virus molecule were frozen on a thin layer of ice. Then, they were imaged through an electron microscope to create 100,000 two-dimensional projections. The McLellan Lab was playing in Nvidia’s GPU sandbox and found some tools to accelerate a solution for the density of the entire molecule. Twelve days later, their raw sample became the very first atomic scale map, a three-dimensional bullseyes for the best scientists in the world to start shooting at.
You are looking at the 3D model of the spike protein right before it infects a healthy cell causing Covid-19.
“This 3D map allows researchers to design antibodies or small molecules that could combine into little cavities to prevent this from binding to the host cell membrane to be really useful for vaccine development,” explained Jason McLellan.
The McLellan Lab’s sandbox is open to the world in collaboration with Dr. Barney Graham at the Vaccine Research Center of the National Institute of Health (NIH).
“There are many different groups trying to make vaccines. There are many different methods for making vaccines. But almost all of them will contain this spike in some form…Our immune system will raise antibodies against it, get ‘em ready to attack. For a normal vaccine development, the entire process would take 10-20 years to go through all the various steps. We’re trying to do this in 1-2 years. It’s really exciting we’re already starting to enroll a phase 1 clinical trial,” Jason McLellan said.
Whenever we get a vaccine, from anybody playing in this sandbox or another one I know nothing about, will never be soon enough. But, after reviewing the NIH data, I would note that it will almost certainly shatter any other timetable record.
For the next phase of their lab’s work, McLellan plans to use the molecule to isolate the naturally produced antibodies from patients who have been infected and successfully recovered. He does not need any more motivation, but I promised to bring ice-cold Shiner Bocks to pour in some disinfected frosted beakers. We will be celebrating a breakthrough from another brilliant team he has likely never met before, creating something new and better.
The greatest technology applications will be science, for its uniting purposes multiplying into the most inclusive projects ever. The changes for good will be unimaginable…except in more of a generation’s brightest minds now dreaming about science a lot more.