Brad Feld

Tag: eric norlin

Paul Kedrosky and Eric Norlin of SK Ventures wrote an interesting and important essay titled Society’s Technical Debt and Software’s Gutenberg Moment.

The abstract follows. I encourage you to read the full essay.


There is immense hyperbole about recent developments in artificial intelligence, especially Large Language Models like ChatGPT. And there is also deserved concern about such technologies’ material impact on jobs. But observers are missing two very important things: 

  1. Every wave of technological innovation has been unleashed by something costly becoming cheap enough to waste.
  2. Software production has been too complex and expensive for too long, which has caused us to underproduce software for decades, resulting in immense, society-wide technical debt. 

This technical debt is about to contract in a dramatic, economy-wide fashion as the cost and complexity of software production collapses, releasing a wave of innovation.


Since mid-March, I have received endless letters from companies and funds I’m an investor in with their thoughts on the Covid crisis. One of the best was from Paul Kedrosky and Eric Norlin of SK Ventures (one of our Partner Fund investments).

Paul and Eric have given me permission to repost it here. 


(First published May 15, 2020.)

Greetings-

To start, a few quotations as markers:

Then he heard the sand rumbling. Every Fremen knew the sound, could distinguish it immediately from the noises of worms or other desert life. Somewhere beneath him, the pre-spice mass had accumulated enough water and organic matter from the little makers, had reached the critical stage of wild growth. A gigantic bubble of carbon dioxide was forming deep in the sand, heaving upward in an enormous “blow” with a dust whirlpool at its center. It would exchange what had been formed deep in the sand for whatever lay on the surface.
– Frank Herbert, Dune

Chigurh: Just call it.
Shopkeeper: I didn’t put nothin up.
Chigurh: Yes you did. You’ve been putting it up your whole life. You just didn’t know it.

– Cormac McCarthy, No Country For Old Men

Unfortunately, most warning systems do not warn us that they can no longer warn us.
– Charles Perrow, Normal Accidents: Living with High Risk Technologies

Crises usually accelerate real trends in society and technology; they don’t create or refute them. 
– Gary Kasparov

The opposite of fragile is something that actually gains from disorder.
– Nassim Taleb


“There are decades where nothing happens, and there are weeks where decades happen.” That is Lenin’s line, and it has felt right in every way and, likely, in almost every country in the world these last eight weeks. And people—investors, in particular—are falling all over themselves trying to understand what it means. We all want to try to explain something this wrenching, and to explain how it feels. 

We want to believe that we just lived through weeks where decades happened, as Lenin said. Except he didn’t say that. And as near as quote investigators can tell, he never said anything like it: the first example of the phrase only appeared a few decades ago. It has caught on partly because it’s well put, but mostly because it captures how we feel about what it’s like to have something come exploding into our consciousness and force us out of our usual amniotic now. We want an explanation, and we want it to explain where things go from here. 

The reality, however, is that wildness has always lurked just beneath the surface. A combination of willful blindness, homeostasis, wishful thinking, and luck have let us skate past the holes in modernity’s ice and pretend nothing lurks beneath it. We have been making bets on smooth, thick ice for decades, and we stopped noticing, even if cracks open anytime in the thickest ice. Pandemics are a crack in our preferred reality, but they are nothing new, even if many countries, like the US, lack recent experience with them, and so pandemics hit harder and longer.

So, what changes? Post-pandemic, in the short-run, and contrary to many, we think very little changes, at least very little that is materially different from what we thought before. Rather than being a break with the past, we think people’s desperation for a return to normalcy—shopping! travel! work!—creates immense pressure to return to the recent past faster than anyone expects. There is inherent human-driven homeostasis, an almost inexorable need to bring things back to where they were before. 

We think the biggest short-term effect will be an acceleration of existing trends. More things will go in the cloud; more things will be virtualized; more things will happen at the edge; more buying, selling, and entertaining will happen online: and so on. These trends will simply speed up.

What about, you wonder, the bigger changes people chatter about, like the death of commuting to work, the end of globalization, the collapse of professional sports, and the like? Not so much. Sure, we will see a paroxysm of people fighting the last war, much like how we armored commercial airliner cockpits after 9/11. In that light, expect a continuing run on contract tracing apps, thermal scanning, work from home chatter, N95 mask technologies, and that sort of thing. But that is extrapolative and impermanent, armoring metaphorical cockpits, rather than thinking about what this episode has taught us about the wildness that lurks beneath modernity. 

We think a more useful analysis must go deeper rather than being merely extrapolative—it must be a thick description of how people live and die. This virus has been, both literally and metaphorically, a disease of modernity. Why? Because It attacks via the vectors of modernity: trade linkages, obesity, diabetes, air travel, mass transportation, urban density, social media, etc. Understanding long-run change requires understanding where modernity itself is under threat, and whether those threats will lead to meaningful—and investable—change.

Fundamental to the changing landscape is the realization that people have been shown how brittle their home structure is. For example, surveys show that New York and Shanghai apartment dwellers are realizing that giving up a balcony for a little more floor space in their aeries made them prisoners of quarantine: most buyers newly say they wouldn’t make the same decision again. Similarly, people all over the world are realizing that “preppers” aren’t nuts (at least, in their prepping), that there is merit in thinking in terms of how much inventory of critical things—food, water, and yes, toilet paper—you have. 

Sociologist of risk Charles Perrow, long ago warned against the catastrophic risks created by tight coupling in society. To Perrow, tight coupling was any complex system where changes in inputs ripped quickly to new and unpredictable outputs, without an opportunity for meaningful intervention. Perrow would have called this current episode a reminder of tight coupling’s risks,  and a forced re-introduction to loose coupling—an attempt to make your life less easily whipsawed by abrupt changes in the world around you. In that light, we think people—and companies—will carry more inventory of everything, that the scarring experience here will turn us into proto-preppers, less willing to be caromed around by the vagaries of life. This a big change, one that will ripple through supply chains, housing, travel, technology, education, and health. 

Speaking of health, life sciences is at an unremarked inflection. There is the real potential for multiple new and effective vaccine and drug delivery platforms to emerge at once, something that has never happened in the history of pharmaceuticals. We not only could see multiple vaccines arrive, which is appealing, but, more importantly in the long run, multiple new platforms for delivering drugs, which would vastly increase the drug arsenal, transform human health, and add vastly to societal wealth via decreasing aggregate cost of illness. 

There is also, however, the real potential for multiple massive drug failures setting the industry back decades. Not just because current vaccine efforts could fail, proving that, in economist Robert Gordon’s terms, we are stuck on an undulating plateau of stalled (drug) innovation, but, more insidiously, that multiple billion-dollar vaccine programs could hit the market at once, all lose money, and re-convince pharma companies that vaccines are a terrible business, making the next pandemic even more therapeutically fraught. 

Which will it be? We are optimists, and we strongly believe it will be the former, but it’s important to keep in mind that it is by no means a foregone conclusion. 

Turning to other deeper changes, machine learning and big data are getting a real run-out here, and given our investments, we are glad to see it. In areas like medical imaging where machine learning continues to acquit itself well, throwing ample shade at human experts. This is overdue, important, and necessary. 

On the other hand, naïve application of “big data” models is being shown for the dangerous practice that it is. Epidemiological models continue to acquit themselves poorly, in part because it’s hard, but also and importantly, to abstract away from this pandemic, because most interesting systems involve humans, and humans adapt and change in ways that work to make models’ predictions fail. As the old capital markets saying goes, “at inflections, markets move in whatever direction will cause the most pain to the most participants.” Big data models suffer no better fate at similar points, as people are belatedly discovering. We are hopeful that this new wisdom will lead to better, more flexible, more adaptive, and more useful analytical models, across finance, medicine, sports, risk, and so on.

Overall, we believe we will quickly return to a state much like where we were before recent events. It will be less different than many pundits expect. Under the surface, however, wildness will lurk—our society will merely be subcritical. This will be, of course, normal, not abnormal. Most of human history has been this way, unlike recent times, which were anomalously placid, a state that’s now ending as we return to subscriticality. We think that making this state visible and manageable will be one of the keys to investing moving forward. 

There will be explosive economic, biological, and technological moves, much more explosive than in the recent past, in part because the ground has been cleared for them, but also because our new, over-excited society has collective scar tissue making it predisposed to jump sooner, further, and faster. This will lead to more rapid technology adoption, faster cycles, and great gains for investors willing to embrace the emergency of subcritical society. Platforms and tools that embrace this—enabling looser coupling, warning when legacy warning systems can’t warn, systems made stronger by volatility—are the emerging investments that we will be digging into as we move forward. 

To summarize, here is our current state of thinking:

  • In the short-run, less will change than people think
  • In longer-run, we will see a complete rethinking of risk, slack, and societal coupling 
  • We are interested in investments that acknowledge, track, and even gain from the wildness and disorder lurking under the thin ice of a newly subcritical society.