Leading The Parade

Background

Terminology: Counterfactual Impact vs “Leading The Parade”

Y’know how a parade or marching band has a person who walks in front waving a fancy-looking stick up and down? Like this guy:

The classic 80’s comedy Animal House features a great scene in which a prankster steals the stick, and then leads the marching band off the main road and down a dead-end alley.

That is not the guy who’s supposed to have that stick.

In the context of the movie, it’s hilarious. It’s also, presumably, not at all how parades actually work these days. If you happen to be “leading” a parade, and you go wandering off down a side alley, then (I claim) those following behind will be briefly confused, then ignore you and continue along the parade route. The parade leader may appear to be “leading”, but they do not have any counterfactual impact on the route taken by everyone else; the “leader” is just walking slightly ahead.

(Note that I have not personally tested this claim, and I am eager for empirical evidence from anyone who has, preferably with video.)

A lot of questions about how to influence the world, or how to allocate credit/​blame to produce useful incentives, hinge on whether people in various positions have counterfactual impact or are “just leading the parade”.

Examples

Research

I’m a researcher. Even assuming my research is “successful” (i.e. I solve the problems I’m trying to solve and/​or discover and solve even better problems), even assuming my work ends up adopted and deployed in practice, to what extent is my impact counterfactual? Am I just doing things which other people would have done anyway, but maybe slightly ahead of them? For historical researchers, how can I tell, in order to build my priors?

Looking at historical examples, there are at least some cases where very famous work done by researchers was clearly not counterfactual. Newton’s development of calculus is one such example: there was simultaneous discovery by Leibniz, therefore calculus clearly would have been figured out around the same time even without Newton.

On the other end of the spectrum, Shannon’s development of information theory is my go-to example of research which was probably not just leading the parade. There was no simultaneous discovery, as far as I know. The main prior research was by Nyquist and Hartley about 20 years earlier—so for at least two decades the foundations Shannon built on were there, yet nobody else made significant progress toward the core ideas of information theory in those 20 years. There wasn’t any qualitatively new demand for Shannon’s results, or any key new data or tool which unlocked the work, compared to 20 years earlier. And qualitatively, the gap between Shannon’s discoveries and Nyquist/​Hartley seems quite wide: Shannon’s theorems on the fungibility of information both pose and answer a whole new challenge compared to the earlier work. So that all suggests Shannon was not just leading the parade; it would likely have taken decades for someone else to figure out the core ideas of information theory in his absence.

Politics and Activism

Imagine I’m a politician or activist pushing some policy or social change. Even assuming my preferred changes come to pass, to what extent is my impact counterfactual?

Looking at historical examples, there are at least some cases where political/​activist work was probably not very counterfactual. For instance, as I understand it the abolition of slavery in the late 18th/​early 19th century happened in many countries in parallel around broadly the same time, with relatively little unification between the various efforts. That’s roughly analogous to “simultaneous discovery” in science: mostly-independent simultaneous passing of similar laws in different polities suggests that the impact of particular politicians or activists was not very counterfactual, and the change would likely have happened regardless. Another type of evidence to look for here is policy changes being made at earlier times, but having relatively little real impact—for instance, Wikipedia’s timeline of the abolition of slavery mentions in 1542 “The New Laws ban slave raiding in the Americas and abolish the slavery of natives, but replace it with other systems of forced labor like the repartimiento.”.

On the other hand, there are probably cases where political/​activist work was counterfactually impactful. While I’m less familiar with the history here, one natural place to look would be polities which had very unusual outcomes relative to other similar polities—think e.g. the East Asian growth miracle, especially Japan and South Korea.

(Aside: at least some economists have been thinking properly about causality and counterfactual impact for a while, so this is something you can probably find literature on especially in the context of the economic effects of policy.)

Business

Business “success” does seem-to-me to typically track counterfactual impact at least somewhat better than scientific “success”, insofar as business success requires outperforming one’s competition. Plenty of people built home computers in the 80’s; Jobs and Wozniak got famous for building better computers. And that continues to Apple later on: Apple’s high status among businesses is mostly about its products being better than their competitors’. Now, we can certainly debate the extent to which that reputation is accurate (I personally am no fan of most Apple products since the early iPod shuffle), but it is at least “about the right thing”—i.e. Apple’s high status among businesses is largely about how well their products compare to the next best. That’s a reasonable proxy for counterfactual impact.

… but there’s still lots of loopholes. For instance, in any “natural monopoly” industry, a business can easily become disproportionately “successful” merely by being first (and thereby grabbing the monopoly). Then we have the same issue as in science: the first mover ends up with high status merely for leading the parade, even if someone else would have done the same thing soon after. Think Facebook or Bell Telephone.

As another example, consider the sort of businesses which produce mediocre products but survive through strong sales teams. Big defense contractors, most large B2B businesses, that sort of thing. There again, business success comes decoupled from counterfactual impact.

(We could also make an economics-101-style argument here. Insofar as markets are efficient, business success should track (positive) counterfactual impact. And we have some standard economic conditions under which markets are efficient: most notably competition and informed consumers. When those conditions fail, business success doesn’t track counterfactual impact so well.)

At a personal level, there’s also the contentious question of how counterfactually impactful the business founder is specifically. The obvious types of evidence relevant here are:

  • Does the founder found other comparably-successful businesses?

  • Does the company perform much less well after the founder leaves?

Picking some contemporary examples: Elon Musk is a go-to example of someone who has founded multiple ridiculously-successful businesses, and is therefore probably not just leading the parade. (And as an added bonus, those businesses are usually not natural monopolies—e.g. Tesla and SpaceX both had to outcompete entrenched competitors.) Another example: Steve Jobs is arguably someone whose company performed much less well in their absence, and was therefore probably not just a parade-leader.

Status is a Terrible Proxy for Counterfactual Impact

Consider our example above of the invention of calculus by Newton and Leibniz. There was simultaneous discovery, so clearly neither of them had very much counterfactual impact. Even if both of them had done other things instead, calculus would very likely have been discovered shortly after—after all, if two people discovered calculus around the same time, it would be rather surprising if nobody else was on the brink of discovery around that time.

… and yet, Newton and Leibniz are among the most famous, highest-status mathematicians in history. In Leibniz’ case, he’s known almost exclusively for the invention of calculus. So even though Leibniz’ work was very clearly “just leading the parade”, our society has still assigned him very high status for leading that parade.

More generally… the part of history I’m most familiar with is the history of science and invention. And it sure seems like leading the parade is the default for high-status historical scientists. Not the universal default—e.g. there’s still Shannon, as a probable counterexample. But at least the majority of historical “major scientific discoveries” involve some strong evidence that the discovery was not counterfactual. Most often, that evidence is one or both of:

  • Prior or simultaneous discovery (often ignored due to lack of demand)

  • The discovery coming very shortly after some prerequisite is available, like e.g. some new data or instrument

Of course, there’s still a continuum in “how counterfactual” research was—really the question is not “counterfactual: yes or no?” but rather “how many years would the discovery have been delayed?”. But the main point still stands: it seems to me that a majority (though importantly not all!) scientific discoveries would likely have occurred within a relatively short time, even if their original discoverers had done something else.

I expect this generalizes to politics. In business, I expect that counterfactual impact is instead the norm, especially among relatively high status small-to-medium businesses, but parade-leaders are still a large minority—especially among the very large businesses which tend to be natural monopolies.

A Missing Mood

I’ve been writing this in a relatively neutral tone so far, but this all implies a mood. A mood which is extremely skeptical and unimpressed by mainstream status by default… but conversely more impressed with evidence of counterfactual impact.

If someone is like “Bell Labs used <management technique>, and they were super successful, we should emulate them!” then I’m like… yeah sure, if your goal is to achieve high status by frontrunning the parade. Now, if Shannon specifically were counterfactually downstream of that management technique, then you’d maybe have an argument. But most of Bell Labs supposed accomplishments, as far as I can tell, were not very counterfactual.

Look, man, I just don’t really give a shit about who’s leading the parade, except insofar as they’re the noise I’m trying to sort through when looking for signal in history books. I’m aiming to play a different game than that.

On the flip side, insofar as most supposedly-high-impact people were leading the parade, that means high counterfactual impact is even rarer than our high-status-focused instincts make it seem. Therefore, we should pay proportionately more attention to those rare cases when we do find them. It also means that we should give proportionately more credit relatively-small-seeming counterfactual impact—e.g. making a discovery ten years earlier than it would otherwise have occurred may not sound like much, but if that’s the high end of what’s been achieved historically, then those are key data points to pay attention to!

A Project I’d Like To See

Note again the last sentence of the previous section:

e.g. making a discovery ten years earlier than it would otherwise have occurred may not sound like much, but if that’s the high end of what’s been achieved historically, then those are key data points to pay attention to!

Is that the high end of what’s been achieved historically? I don’t know, because I’ve never seen someone systematically study the question over a wide range of historical cases.

That’s just one question among many which could all be answered with the same project. Here’s what the project would look like:

  • Take a big list of discoveries and inventions

  • For each of them, look for some standard types of evidence of counterfactual impact or its absence—like e.g. independent simultaneous discovery or a wide/​short time-gap between the prerequisites becoming available and the discovery/​invention occurring

  • Ideally, make an order-of-magnitude estimate of how long each would have been delayed had the discoverer/​inventor worked on something else. (Months? Years? Decades?)

  • Put it all in a big table and sort it.

People often say “there’s no obvious recipe for high impact research/​invention”, but I’ve never actually seen someone do the work to sort by counterfactual impact. So… I have no idea. It’s the sort of exercise which I’d expect to reveal surprising patterns.

(An aside: I’m sure some of you are thinking “let’s use a large language model to do this!”. Not a bad idea. I did experiment a little, and found that for basically every discovery, GPT-4 says the discovery was not very counterfactual, and usually gives a fairly generic argument with very similar structure—“Simultaneous discovery, or the idea of ‘multiple discovery,’ suggests that scientific breakthroughs often occur independently and almost concurrently by different people”, blah blah blah. It does give different estimates for quantitatively how long discoveries would be delayed if the original discoverers/​inventors hadn’t worked on it, and it is a good source of potential prior/​simultaneous discoveries to check. But it does very much require checking.)

What If Leading The Parade Grants Counterfactual Impact Later?

People occasionally come up with plans like “I’ll lead the parade for a while, thereby accumulating high status. Then, I’ll use that high status to counterfactually influence things!”. This is one subcategory of a more general class of plans: “I’ll chase status for a while, then use that status to counterfactually influence things!”. Various versions of this often come from EAs who are planning to get a machine learning PhD or work at one of the big three AI labs.

A proper analysis of such plans would take a whole post on its own, and distract more than I want from the rest of this post. But since it’s relatively common among my social circles and adjacent to the post topic, I will drop a few quick takes on this sort of plan.

When someone tells me they’re going to optimize for status and then use that status for counterfactual influence, here are thoughts which go through my head:

  • “Reeeeeaaaaalllly? You’re chasing status for counterfactual influence and not because, say, chasing status is inherently rewarding and familiar and a lot easier than directly tackling hard research problems? I have some True Rejection-style questions for you...”

  • “Your stated plan of <getting PhD/​working at big three lab> will not actually get you much status. You’re following a local status gradient, but it’s a local gradient which maxes out with you being a peon. If you want to have nontrivial counterfactual impact via status, then stop being such a wuss and tackle something ambitious enough to get you a lot of status. Go found a startup or something.”

  • <Mental picture of someone marching along at the front of a parade, then veering off onto a side street looking very official, and the rest of the parade totally ignoring them.>

Summary

Status is a shit proxy for counterfactual impact, especially in science. High-status historical figures largely seem to be leading the parade: if they hadn’t made the discovery or invention, someone else likely would have shortly after. That said, at least some cases were probably counterfactually impactful… but high social status isn’t a very good proxy for identifying those cases.

There are simple kinds of historical evidence we can look for, to tell whether someone was counterfactually impactful or just leading the parade. Simultaneous discovery is the clearest—e.g. since Newton and Leibniz discovered calculus at basically the same time, clearly neither invention was counterfactual, and likely someone else would have figured it out soon after even if both Newton and Leibniz hadn’t. On the other side, when there’s a wide gap in time between prerequisites and the discovery/​invention, that does suggest high counterfactual impact.

This all adds up to a mood which is very skeptical of mainstream social status: “look, man, I just don’t really give a shit about who’s leading the parade, except insofar as they’re the noise I’m trying to sort through when looking for signal in history books”. Until someone does the work to check which historical discoveries or inventions were counterfactual, and which were just leading the parade, that’s my default attitude.