Would You Rather…
You are asked “would you rather live in a time of stability but low economic growth, or a time of instability but the chance at high growth?”. Which option would you choose?
Perhaps some would choose the first right away. Others might need more detail. How is growth defined? What’s the severity of low or high growth? What’s the degree of instability? I too would need to address these questions to provide a measured answer. But I’ll be honest: unless instability is very low and the chance at growth pretty high in option 2, I’d go with option 1.
It’s reasonable to claim that instability compounds itself. The more chaos in a system, the more likely chaos will reign in the future…right? “Growth” is not always that great anyway — think negative externalities like pollution, or infringement on human rights— and even if it is, “well-rounded” growth seems more sustainable when society is not on a chaotic trajectory. So, if given the choice, in most cases I’d go with option 1. Final answer.
Of course, you can’t choose the cards you’re dealt. As 2021 rolls on, it seems the hand that the world was dealt looks more like option 2. Let’s be real — the lived experience of option 2 so far is awful. To anyone stuck in quarantine or dealing with the death of family members due to the pandemic, there needn’t be a reminder.
However, if option 2 is in fact what we have on our hands, is there a chance at high growth? Is there a reasonable chance at greater progress?
“Progress” is not the first word that comes to mind after nearly a year of living in a pandemic. I don’t have to be specific as to why that is: overall death, economic downturn, political instability, as well as incompetence, negligence, gridlock amongst our leaders. If anything, it would seem like the pandemic has set us back — both the developed world and the developing world perhaps more. So what case could there be for impending high growth?
Growth At A Standstill
Well, it turns out that “progress” has been quite hard to come by for the past 50 years. You can see this in the lackluster economic growth statistics in the US (especially TFP) and in the slowdown in new patented technology. I’ll limit this discussion to only the economic and technological dimensions. Social and political matters deserve a separate discussion, although I don’t think many are keen to argue that social and political progress has fared any different.
I’d wager that if I asked someone born after the 1970’s what the greatest technology is that they’ve seen come to pass in their lifetime, it would be the internet, the smart phone or the personal computer. Who could argue with that? The internet and the digital space have changed our lives — from the way we shop, to the way we learn, to the way we communicate and develop socially. The internet has only grown in importance during the pandemic, and I think it’s a safe bet to say that the internet will remain a bigger part of our lives in a post-pandemic world than before the pandemic.
But all this change is concentrated in the digital space. What about the physical space? Well, we have electric cars now — but adoption is still slow. In the US, infrastructure looks about the same it did in the 1950’s. Planes look the same and are not faster — though domestic and international flights were the cheapest they’ve ever been before the pandemic. The physical space of our world hasn’t been revolutionized in the way we expected it to — we were promised flying cars, after all. There’s a famous quote by Robert Solow (in 1987!) that rings true: computers can be seen everywhere except the productivity statistics.
Let’s get back to the central idea here. Is there a reasonable chance at better growth? Central to this matter is the question of whether or not crisis breeds innovation. The devastation of WWII incited several advancements (penicillin, atomic energy). Of course, there are plenty of counter examples — especially in recent times. We didn’t learn much from the tech bubble in the 2000’s, and we didn’t learn much from the financial collapse. 20 years since 9/11 US forces still kill people and still get killed in the Middle East.
So, it seems like the crisis -> innovation logic is shaky at best, especially in recent times. But perhaps the potential of the 2021’s lies not in the idea of crisis leading to innovation. As it turns out, this sorry state of growth isn’t unique to the past 20 years, though now it seems particularly bad. When we zoom out a bit more, this lack of growth is part of a much longer period some call The Great Stagnation.
However, many claim that the despair of the pandemic is exactly the event we needed to spur change for the new decade. Several leading voices claim that the 2020’s show unique promise for a way out of this period of economic and technological stagnation.
So where can we expect to see this miraculous growth? I — like many — am very interested in this question. Pulling together innovations from diverse sources, below I discuss:
- the technological areas already changed by the pandemic
- the ones most ripe for opportunity to push us out of this stagnation
- my assessment of their likelihood of coming to fruition
Pandemic Gains
It’s not surprising that the biggest productivity gains in the pandemic — outside the obvious mRNA vaccine, which I’ll talk about separately — are in the digital sphere. For many occupations (including my own), remote work is the new normal. Without unproductive time spent in burdensome commutes, workers are able to focus more intently on work. There’s even evidence to suggest people are working longer than they were before, which may be good for productivity and not-so-good for mental health. In my eyes, however, the gains of remote work are the additional hours people reap by spending time not working and not commuting. Though remote work only effects certain kinds of workers, for many others the idea of remote work is becoming less of a novelty. I think a balance of remote and in-person work is the best for productivity and psychological health in the long-term.
In addition, firms were forced to make investments in cloud computing and video conferencing. These overhauls and investments will continue, and are an important step in making the digital space and the internet more central to our economy. Investment in digital and video-based education is a natural progression of this initial investment, though there are of course growing pains. To make sure that the transition to learning online isn’t hardest on rural communities, we should focus on making broadband internet as ubiquitous throughout the US as it is in Switzerland or in Taiwan.
Bio-Tech
The most publicized breakthrough during the pandemic is the first widely-used mRNA vaccine. The newest vaccines work in similar ways. Scientists isolated the coronavirus' RNA that produces the corona-spikes and created a mRNA that gives our bodies instructions to build these corona spikes. Our immune system then learns to attack cells with these corona spikes and gives our immune system the chance to kill the virus before the viral cells replicate.
Future vaccines of this type, which can be rapidly developed and distributed (as long as the viral protein is sequenced), are the future of epidemic-fighting technology. Gene-editing using CRISPR, which seems to be slow in generating a market-ready product, also seems to be picking up in terms of useful applications. The method is used in this new covid testing application, but it’s plausible that using CRISPR for gene repair could soon target specific genetic disorders such as sickle-cell anemia and cystic fibrosis.
This breakthrough follows a slew of similar potential breakthroughs. A universal flu vaccine is now in our headlights, as is a credible malaria vaccine. The progress on these fronts could be understated, but the investment surge in biotechnology is not. Large investments of this nature will have an outsized impact on our society if one of these opportunities comes to fruition.
Machine Learning
People outside of the machine learning industry also tout the recent improvements in two areas of research: natural language processing and protein-folding. GPT-3, a recent development in generating human-like speech, has received a lot of press. It performs well on several language tasks. For example, if you give it half a sentence, it will tell you which words come next. Or, if you ask it a factual question, it will give you an answer.
Its results seem impressive. Unlike other models, GPT-3 does not query any outside data at inference time (i.e. when it is asked for a prediction). Yet, for many NLP tasks, GPT-3 outperforms models that do query outside data (like Wikipedia) for factual information at inference time. Still, a question arises: did the model learn anything?
The more one looks under the hood, the less convincing it seems. What makes it so interesting is how big the model is. It uses 175B parameters, whereas the largest languages model before it only used 1.5B parameters! With so much training data encoded in these parameters, people like Yannic Kilcher seem to think that it performs so well on language tasks simply because the text corpus it was trained on is so large that it already has seen an instance like it that is encoded in these parameters. This changes the problem from a generalization problem to a lookup problem.
Still, what amazes me is not GPT-3’s performance but the fact that a model of its size can even be used. It’s a marvel that this model can be trained in the first place, and model size may in fact be the future of language learning models. Still, I don’t think the model is all that innovative. It’s not a new technique — it’s just really large. So let’s be frank: it didn’t learn to think or reason. It just has access to a larger corpus than any other algorithm before it had.
In my opinion, a more promising development is Alphafold. The protein-folding problem is a notoriously difficult one. Proteins are defined by their amino acid chains, which are the building blocks of DNA. There are over 200M identified proteins, but the full amino-acid structures are known for a only fraction of them. What Alphafold does is it takes an amino acid sequence as an input and it outputs a specific protein.
So is this useful? Right now scientists discover structures of proteins through years of painstaking labwork. With alphafold’s predictions, however, drug makers can rapidly deduce protein structures in new pathogens (like SARS-COV-2) and find ways to attack them. This application is a great example of how useful machine learning can be when applied to diverse fields such as biology or medicine.
Hypebeasts often overrate the promise of machine-learning medical diagnostics, but in this example the excitement is justified. The protein-folding problem is a 50 year-old problem, and the fact that many experts now consider it solved is a major breakthrough for the field and demonstrates great promise for the intersection of machine learning and biology.
Improved Technology
I talked about breakthroughs in bio-tech (which are on everyone’s mind in a pandemic) and in machine learning (which are on my mind). But one can make the case that the 2020’s will also witness several existing but improving technologies.
It seems that two technologies poised for prolonged success in the renewable energy space are solar power and new batteries for electric vehicles. Solar is already a success story. The cost of solar power plummeted in the past 20 years, and is now the cheapest way to produce energy in many places. Given the influence of the Green New Deal on the US Democratic establishment now in power, big investments in renewable energies like solar are on the horizon. In the next years we may see solar adoption at a scale we’ve never seen.
One can say the same for electric vehicles. The cost of a hybrid car has decreased steadily since the new millennium. Large investments in electric vehicles from the Biden administration coupled with innovations in new batteries like these produced by Toyota may speed up the adoption of electric vehicles. Greater adoption can only be good news for a world that must deal with the consequences of global warming in the near future.
Of course, there are also the tech giants. Apple hasn’t dawdled their time away in the pandemic. They are set to release their M1 chips in the next generation of MacBooks. From all preliminary speed-tests, it seems we can expect to see competition in the personal computing technologies surge, with the M1 chip leading the way for affordability and performance.
And who can forget the most recent successes of SpaceX. Besides making space cool again, the aerospace manufacturer is set to add to their resume of impressive accomplishments. The milestones due up in the 2020’s include launching Tom Cruise aboard a spacecraft to shoot a movie as well as a crewed mission to Mars, which Papa Musk claims could happen as early as 2024 (though this seems implausible). Private firms like SpaceX are leading the way in investments in space engineering and transportation, and their progress in the 2020’s could finally get the public interested in space-related R&D. Their success may even convince people that governments should again invest in more ambitious research and space-related science programs.
Hope and Chance
I want to make it clear that I don’t think it’s likely that these technologies will speed up growth. I, like many people right now, think that there are many more ways for us to screw up the unique opportunities at rebirth that this pandemic offers us than to use them to progress forward.
I’d also go long on many of the “hottest” technologies that many claim will lift us out of stagnation. If there are gains, it’s likely they won’t be as large as we think. I don’t want to fall into the trap of picking a few emerging, speculative technologies and saying “these are going to revolutionize the world” — they probability won’t.
Still, with investment and discovery in these areas of opportunity gaining speed at once, perhaps the chance we stumble into grand changes is not as small as it seems. That’s why I’m a fan of greater spending on R&D at scale — the more lines we have in the water, the more likely we are to stumble into these grand changes.
In fact, it may be that the chance at high growth is the highest it’s been in a long time. I’m convinced there’s never been a time in recent history when there’s been such a large mass of unhappy people dissatisfied with the status quo. To quote Anton Chigurh from No Country for Old Men:
If the rule you followed brought you to this, of what use was the rule?
It’s never been more critical for this unhappy mass of people to be vocal about NOT wanting to return to how society was. When there’s change, there’s also the chance at progress. So, to sum up — the chance at real growth is low, but perhaps the highest that it’s ever been. It’s going to take a lot of work to make this a reality. Of course, a little wishful thinking doesn’t hurt. Hope can help us bridge the distance from chance to reality.
