Suppose almost everyone in the world worked really hard to get better at football as a hobby, instead of wasting their time watching TV or getting angry about politics. The level of professional football games would improve, but there would still be about the same number of professional football games; maybe a little more diversity, but just as today, most people would watch the matches of the top clubs and not the much larger number of bush-league matches. So the overall well-being of the world would increase, but only a little.
Suppose that, instead, almost everyone in the world worked really hard to get good at medicine as a hobby — doing paramedic training, reading case studies, predicting patient outcomes, doing Gwern-style double-blind self-experiments, performing recreational plastic surgery on their pets, that kind of thing. The level of health care would improve dramatically — deaths due to drug interactions the prescribing physician failed to notice would go way down, people’s decisions about when to go to the emergency room would be much more accurate, people’s lifestyle choices would become slightly healthier, and medical science would advance much more rapidly. The overall well-being of the world would increase substantially.
On the gripping hand, suppose that almost everyone in the world worked really hard to get good at fighting with kitchen knives. Perhaps the level of Olympic fencing would improve substantially, which would make for enjoyable TV — though perhaps the sports are far enough apart that it wouldn’t transfer — but it’s also likely that a larger fraction of fights would involve knives instead of just shouting or fists, usually killing one or more people. The training, too, would probably involve a certain number of accidental fatalities, and some fraction of those would lead to revenge killings. More children would be orphaned. The overall well-being of the world would diminish.
On this basis, I claim that it would be better for almost everyone to take up medicine as a hobby than football, and better football than knife fighting; and this is because knife-fighting is a negative-sum game, football is a zero-sum game, and medicine is a positive-sum “game”.
As I remember, sage Dave Long at some point unpacked the Bill and Ted philosophy as follows: “party on” means not to engage in negative-sum “games” — in the very general sense of “interactions with other agents” — and “be excellent” means to engage in positive-sum games. To the extent that you can control which “games” you “play”, it is better to invest your limited time in positive-sum games. (Sage Shakyamuni reportedly took the alternative position that you should not waste your time playing any games, because you are going to die soon, while the sages of Wyld Stallyns instead played games to defeat death itself.)
Of course, a person may find themself in a situation where it is more advantageous to them personally to direct their effort to a zero-sum game or a negative-sum game. Consider the thief who spotted my neighbor, an old woman, at the bus stop last month, and decided to knock her down and beat her to steal $300 from her (about US$7). He chose to direct his efforts to robbing her, and although he injured her in the process, he presumably considered that a worthwhile tradeoff, because it satisfied his desires for food or crack or whatever. A pyramid-scheme participant who recruits downstream members is doing the same thing, but in a less courageous way, and the injury is delivered via deception to the victim’s mind, rather than via a stick to her knee.
From a consequentialist point of view, this is a bad outcome if we consider the thief and their victim to be equally important, so that the loss of money by one is precisely canceled by the gain of money by the other, while the putative injury or deception has no such counterbalancing benefit. If the consequentialist considers the thief’s profit to be more important — perhaps because the thief is poorer and therefore benefits more from the money, or perhaps because they subscribe to a worldview where it is better for the money to go to a deserving thief, because of his bravery, rather than a cowardly victim, or simply because they are racist — then the consequentialist can justify the theft as morally correct. But an egalitarian consequentialist cannot so justify theft or any other negative-sum game.
Neither can a deontologist who subscribes to Kant’s categorical imperative, which is of course (XXX) one of Kant’s first examples; the thief cannot at the same time will that the victim should steal their money back, delivering the same beating to the thief, because that would leave the thief bruised but no richer. Only by some kind of special pleading can the Kantian thief save his livelihood, and the same kinds of special pleading that the consequentialist could use to justify the theft can be used to exempt the thief from the maxim that would otherwise strip his stolen money from him. Again, this logic applies equally well to any negative-sum game.
For an egalitarian consequentialist, the imperative to “be excellent” — to play positive-sum games rather than zero-sum games — is just as strong as the imperative to “party on”, that is, not play negative-sum games. Kant (and most deontologist theories) also regard it as a “perfect duty”, and thus obligatory, to “party on” in every possible way — at least in the specific examples Kant gives — but regard it as a supererogatory “imperfect duty” to “be excellent”.
In theory, since computers only do what you program them to do, you can not program them to accept arbitrary commands from random people anywhere on the internet, instead only programming them to accept arbitrary commands from their actual owners. That is, there is no need for software to contain security holes. Computer security violations are not inevitable. They are the result of undiscovered programming mistakes, which is why the most significant forum for announcing and fixing security holes in the 1990s was called BUGTRAQ.
In the late 1990s I spent some effort finding security holes in software and reporting them to get them fixed. Unfortunately, it became increasingly apparent that the current economic and intellectual environment was going to introduce new security holes faster than we could remove them, so rather than a gradually decreasing pool of still-undiscovered security holes, we have a gradually increasing one, with the disastrous effects on the human right to privacy documented by the Snowden revelations.
In such an environment, exploit users and defenders are in a sort of arms race, a literal race to see who can respond faster. If the defender is faster to patch a newly discovered hole in deployed systems, they win that round; if the exploit user is faster to acquire and employ an exploit for it, they win that round. So whenever one side or the other increases their commitment of resources and gets ahead a bit, the other side has the option of matching that increase to get back to the previous equilibrium.
This sounds like a zero-sum game, but in a larger context, it’s negative-sum: the human effort spent on the vulnerability treadmill is taken from the time available for making music, discovering new algorithms, cooking food, meditating, painting pictures, writing poetry, reading poetry, watching movies, making love, or raising children. Every sysadmin on call who has to patch the production systems within an hour so that exploit users won’t break in with the newly announced vulnerability is spending the irreplaceable minutes of their life on a Red Queen’s race with the spies at the NSA or the FSB. They are falling short of partying on.
Of course, neither side can unilaterally scale back its efforts; that would amount to surrender. If it’s your job to keep your employer’s public-facing systems up to date so they don’t get popped, you can’t just decide it’s not worth the effort. But you can find a new job.
There are other drawbacks as well: it’s no longer a viable option to continue running outdated software unless it’s in a very unusual isolated environment, like a non-networked video game, and the increasingly rapid and frequent response required to new vulnerabilities has the effect of centralizing both patching and exploit use in large organizations. This means everyone is exposed to the risks of deploying untested new software on short notice and to having strangers administer their most intimate computer systems, such as Android hand computers. Also, sometimes patching a vulnerability unavoidably introduces incompatibilities, causing bitrot.
So I stopped spending effort on that.
The nuclear arms race during the Cold War had a similar dynamic: each side constantly worked to preserve a second-strike capability (by, among other things, building enough warheads that some would be likely to survive a first strike from the other side) and to find ways to remove the other side’s second-strike capability. Neither side could opt out of the game, but the result of both sides playing it harder and harder was the decades-long threat that civilization could end at any time, with 20 minutes of warning.
Fortunately there were also people on both sides like Jonas Salk, Norman Borlaug, and Andrey Kolmogorov who were able to dedicate their lives to positive-sum games instead of the negative-sum game of the nuclear arms race.
Commercial competition is, in theory, a positive-sum game, though eventually only slightly so — as more and higher-frequency participants in a stock market means fairer prices, with retail participants having to pay much narrower spreads to the market-makers, more competition in consumer-goods markets should result in goods priced just above the lowest possible marginal cost of production. One reason this doesn’t happen in practice is advertising: consumers buy goods that are advertised rather than equivalent goods that are not, and in many cases are induced to buy categories of goods they wouldn’t have bought at all without advertising.