Renormalizing hype

Gaming the game theory of scientific risk taking.

It’s probably safe to say that a fair bit of the hype surrounding quantum computing has been damped over the past few months. Recent re-analyses of experiments that claimed Majorana fermions have been observed as emergent degrees of freedom in certain semi-conductors have found these claims to be premature. This is noteworthy, because Majorana fermions (neutral electron like particles that are their own antiparticle) are thought to offer one of our best hopes for large, stable and scalable arrays of qubits, necessary for quantum computing technology to ever move from laboratories into the real world. If that weren’t bad enough news for quantum computation enthusiasts, a recent pre-print purports to have shown that Google’s claim of having achieved `quantum supremacy‘ — that is, of having an implementation of quantum computation that surpasses anything practically achievable by classical computers — may also in fact, have been premature.

So where might this leave things? It is important to step back and take everything in context. `Hype’ is not just a seemingly unavoidable side-effect of our present form of capitalism, it is perhaps an essential component of fostering innovation and investment in new ideas within this context*. This is because the purpose of hype is to understate, if not obfuscate risk. This is absolutely crucial to encourage initial investment, be it in terms of resources, labor, or less tangible expenditures such as risk to career prospects or perceived status accrued by junior researchers who choose to work in a fledgling field. If everyone were cynical of the possibilities from the outset, no one would ever take the chance to devote time and resources to an endeavor whose success is far from certain, and the field might never get off the ground. It is only after achieving a critical mass of suspended disbelief that highly speculative endeavors can stand a chance of getting anywhere, even if this critical mass takes the form of a small, dedicated band to begin with. This to me, appears to be the core functionality of hype within science, at least initially.

Of course, none of this happens in a vacuum. Almost always, hype requires a firm center of gravity to begin with, typically the initial backing of a small number of established, high status individuals within a field, and perhaps some accompanying seed funding. This serves the purpose of offering at least short term cover for the immediate career prospects of junior researchers embarking on this path. In this way, hype within the scientific enterprise appears to take on the same role that complexity does for financial products — by obscuring risk, and thus reducing people’s apprehensions to parting with their hard earned money. It is worth reading through the linked to blog post by Steve Randy Waldman, which this set of musings borrows heavily from. The meta-process underlying both sets of observations is evidently that of Nash equilibria.

Although it might seem like a caricature of any real life situation to consider, in a two player game of `invest, or not to invest’, a very transparent dynamic immediately plays out. If neither player invests, things remain as they always were. If one player invests and the other doesn’t, the result is a likely loss of return for the player who does. If both invest, both players can count on some form of payoff. As a canonical example, consider starting a car factory in a hitherto agrarian society. There’s no point building cars if other people aren’t building gas stations, spare part factories, roads etc. One can’t just decide on one’s own to start manufacturing cars and expect any return on the investment. A host of other agents have to be investing as well for anyone to gain anything from it. The corresponding situation in the two player game is that there are two stable Nash equilibria — neither invests, or both do. The purpose of freely available credit and limited liability is to eliminate the lower equilibrium. In the ecosystem of scientific ideas, hype is evidently the device that eliminates the lower Nash equilibrium of collective hesitance to embark on a new endeavor.

Problems begin, if and when hype outlives its utility, especially if a self referencing and self reinforcing universe of subjective value, however compelling in the minds of its proponents, ends up taking on a life of its own. One could argue that there are a number of such asset bubbles in physics, as there are in the economy. There are very likely entire sub-fields of various theoretical sciences that may come and go in the long run without leaving much in the way of lasting impact, other than having gainfully sustained a bunch of scientists over some duration. All of this perfectly fine, if not inevitable if we’re being true to our ambition as a species. However the important caveat here is that this is only fine so long as we don’t end up investing in too many bubbles that end up imposing an opportunity cost on potentially more viable lines of inquiry. Within the context of theoretical physics, there are various commentators who feel that the past thirty years of high energy theoretical research has succumbed to this fate. My own view is that it’s too early to tell, and moreover, it behooves us as scientists to explore a particular meta-logic for discovery that has culminated in the modern edifice of particle physics to its logical conclusion. Furthermore, because theory is cheap, the monetary opportunity costs, if any, are not particularly high, for now — although the opportunity cost in terms of time and careers spent may be worth pausing to consider. All of this of course, comes with huge caveats, which I might explore in future posts through the analytic lens of the (other) Frankfurt School.

With regards to quantum computation, even if the sheen of the initial hype has started to wear off, it has served its purpose in actualizing the field beyond the confines of theoretical speculation, to something we’re nevertheless still heading towards, if on a longer timescale. It is almost certainly true that if quantum computing ever becomes workable commercial technology, it will have a major role to play in the fourth industrial revolution that is evidently imminent. Or perhaps that too will turn out to be unsubstantiated hype?

*Of course, this is a highly contingent statement. There are many plausible alternatives within the market based context (some, not even that radical a departure from the economic systems currently prevalent in the western world) where innovation can be incentivized differently. It has certainly been thought about and developed within the socialist context, which contrary to parlor room conceptions, can also embrace markets as part of its implementation.