Defending Goliath

Biggest guy is best guy

Epistemic status: Not actually sure how to measure the epistemic status for this one. This one came out more Ayn Rand-y than I intended - I want this to be a mental tool people have, not a replacement way to measure value.

A few weeks ago, the F35 at long last got its first manned air-to-air kill¹, shooting down an Iranian Yak-130 training jet. Reactions ranged from “huh neat” to “don’t brag about it, of course a top-of-the-line fifth gen fighter jet could shoot down a 90s-era training jet, talking it up is a bad look”. There’s something general here - the claim that cheering for the overdog and punching down is a bad look - that I want to get more into, because it’s the basis for a lot of well-intentioned mistakes.

In general, people like David vs Goliath stories and instinctively side with David for two good reasons².

The first is ingenuity. David is smaller and weaker than Goliath, so he needs to have some special trick or advantage to fight back³. This makes for a more interesting story, but it’s also more useful for us to hear. Most of us can’t just copy Goliath’s size and strength, but we can learn tricks and ingenuity from Davids. And from the outside view, there’s alpha in finding Davids, a Moneyball-style way of knowing how to find non-obvious advantages. If we just bet on Goliath every time we’d be adding approximately zero insight to the problem.

Still, it’s worth remembering that this is mostly wrong. Usually Goliath wins, because being bigger and stronger and frankly is a professional and experienced soldier. The Iranian pilot flying the Yak-130 was probably outclassed in skill as well as equipment, given western air force training standards. It’s worth keeping an eye out for the possibility of Davids, but unless we’re discerning enough to see them coming, we should embrace the judgement of the rock that says Goliath always wins.

For a subtly different version of this - Sometimes you get a behemoth incumbent like Google or Ford successfully challenged by an OpenAI or Tesla (or an OpenAI challenged by Anthropic), and it’s important that this can happen because big behemoths get slow and bureaucratic and clunky and sometimes you just need to burn them and replace them with something new⁴. But most new startups aren’t plucky Davids ready to sweep in a new world. For every successful Tesla there’s a dozen Nikola-style scams or failures⁵.

The second reason to side with David is a more utilitarian argument that David might need our support more. There’s a few versions of this - the virtue argument that we should help the weak, the utilitarian argument that marginal aid goes further to those who have less, and the cynical argument that providing support for David will be more meaningful to him and will inspire greater reciprocity in return.

These are all often good and useful arguments. But they need to be situational rather than absolute.

The cynical argument is the easiest to address, in that I don’t really have anything against it. It can be true or false depending on the situation (helping out your neighbor will generally get you more reciprocal goodwill than either helping out a movie star or helping out a homeless stranger). In general people seem to deal with this tradeoff okay in everyday life but badly in internet fights or politics, where they have a systemically poor sense for what constitutes an attention-starved underdog⁶.

The one about virtue is the hardest. Humility is often considered a virtue, but weakness itself does not correlate with virtue, and is in general anticorrelated with it. This isn’t to say the weaker parties can’t be more virtuous (the original Goliath was just a big guy in a way that doesn’t seem tied to any particular virtue).

But in general the correlation goes the other way, because being virtuous is good for you. Being hardworking and honest and reliable and empathetic are all virtues that also make you stronger and more competent and will, in general, correlate with the ability to win in competitions. There’s some individual exceptions - an unscrupulously dishonest politician might be able to win elections by lying about his ability to solve problems, for example - but those generally weaken society overall (a society led by lying politicians will be weaker), so in terms of societal virtue they’re still net negatives⁷. So there’s a hard local/global tradeoff problem here - you need to reconcile what’s virtuous for you with what’s virtuous for society.

So I’d argue we should draw the line around charity in this way. Helping people out is good if you see a way for it to actually improve things into a better state of affairs. Blindly giving without thinking about how or why will not, in general, lead to a stable state of affairs.

The utilitarian argument is the most likely one to be true but also the one we have to be most careful about. Helping people is often good⁸, but doing it systemically provides bad incentives to act or appear weak. Especially if weakness is correlated with worse behavior, this incentivises people to act worse in order to be given resources, which is destructive to society⁹.

I don’t exactly have a full solution here. A healthy society does need ways to support the weak and downtrodden, for both the reasons mentioned above. It just needs to do it without actively incentivising them, and without forgetting that good things succeeding is good¹⁰.

1. The F22 meanwhile, despite being planned as the ultimate air superiority fighter, is still at zero unless you count chinese weather balloons :(

2. A third, more straightforwardly bad reason is recall bias - because we hear so many stories about David defeating Goliath, we come to expect it to happen by default.

3. There’s an interesting comparison here to DBZ: The stereotype is that Goku is just strong enough to out-punch whoever he’s fighting. But in an actual arc he typically starts out weaker and has to find ways to power up, and a typical fight will usually involve a bunch of tricks and special moves. Even in famously punchy anime, having the story actually come down to who punches who harder would be boring. Superhero movies tend to work similarly.

4. Hey, remember when Google was the cool new kid with the hot talent and the usable products compared to the stuffy old incumbents like Microsoft? And probably Microsoft was that compared to IBM at some point, though this would be before my time.

5. In case you don’t remember who Nikola was, they were the guys who couldn’t get their electric truck to actually run, so they filmed it rolling downhill for their promotional video.

6. One skill of a good politician is to make voters feel like they care about them personally, and one method of doing this is to play yourself as an underdog campaign with few supporters even if you’re getting millions of votes. Voters are mostly good about being cynical about establishment politicians but less so about upstarts - remember, the single biggest bottleneck of a politician caring about your issue is how much total attention they’re getting, not how institutionally captured they are. Once a politician has enough support to be on the news, you don’t have much of an impact on them.

7. There are some harder exceptions. Local pollution is unvirtuous and generally pretty harmful to the society that pollutes, but greenhouse gas emissions are pretty much only harmful globally and don’t disadvantage defectors in particular. So this correlation is broadly useful but nowhere near universal.

8. There is the “give a man a fish” style counterargument about this encouraging dependence, but that’s mostly orthogonal to the David vs Goliath issue here (it would also affect praise for Goliath), and we’ll put that aside for now.

9. There’s a Sherlock Holmes story about a reporter who discovered he could make more money as a beggar, so he took up begging while lying to his wife about it. In order for this to work he had to dress himself up as the most wretched beggar in London (he notes that normal beggars don’t get nearly as much money as he does), showing a case where the charitable incentives promoted ugliness in a very literal sense.

10. Remember to thank your local F35 today!