Parasite Load
Every institution is meant to produce something. This might be something concrete, like cars or pants, or it might be intangible, like law or knowledge. An institution that succeeds at this task then accumulates wealth. The company tasked with producing cars accumulates wealth by selling them at a profit. Then the company is taxed and the taxes are given to the government agency tasked with administering law. Then the company’s CEO donates a hundred million dollars to the university tasked with producing knowledge.
In theory this all works great. As everyone does their jobs, civilization ascends higher and higher, with more material wealth and better-administered laws and more knowledge than before. The institutions, and the people within the institutions, accumulate money and power because they do useful jobs well. The physical wealth, the cars and pants and couches, get distributed to people in return for making contributions to society.
The car company is only able to make cars and sell them at a profit because it’s in a society which has lots of law and lots of knowledge, so it makes sense that the producers of material wealth will end up sending some of that material wealth to the institutions that produce these intangible goods, through some combination of law and custom. That’s what it means for everyone to participate in a functional society. If you squint a little bit and force everything into an econbrain framework then you can even see it as a kind of trade.
The problem is that piles of wealth attract parasites. If you fill your pantry with sugar then ants will try to get in. For an institution, this happens both internally—you are giving out steady high-paying salaries, so people try to get those salaries by doing work which superficially appears to help your company, or by lying about doing work and pocketing the money—and externally—you are giving money to institutions that educate the youth, so institutions try to get that funding by doing work which superficially appears to educate the youth, or by lying about doing work and pocketing the money.
Bigger piles of wealth and power and prestige attract parasites in greater numbers and with greater sophistication. I’ve seen small hobbyist communities attract clumsy affinity scammers who run away with a few thousand dollars in donations for fake cancer treatments. I’ve seen fashionable research organizations attract smooth-talking job applicants angling for no-show sinecure jobs and connections in the scene. I’ve seen the feeding frenzy of half-baked grant proposals when the FTX Future Fund sprayed out $100 million in donations before its funders were arrested for fraud. I’ve seen brilliant hucksters spin elaborate webs of lies and apparently-working “AI” technology to try to scam tens of millions of dollars from investors.
In isolation this is inconvenient, but not a disaster. You spend a fraction of your resources on detecting and purging parasites from your system. Your detection won’t be perfect, so parasites will consume a fraction of your resources on top of that. It’s a source of drag, and will kill some projects on the margin, but it doesn’t fundamentally change what’s happening. Every big group in the history of the human species has had some amount of this going on.
Things get really bad when the parasites become politically organized. This is the point where things metastasize. Once an institution contains a powerful faction who are appropriating the institution’s resources at the expense of the institution’s mission, rather than by fulfilling the institution’s mission, it has entered a terminal spiral which is nearly impossible to reverse. Isolated individuals stealing stuff can only do so much damage, and the amounts they can steal are limited by the need to avoid detection. On the other hand, factions can bend the institution’s priorities to boondoggles which are easier to steal from, and in advanced cases such changes in strategy crowd out even the attempt to do real work.
For example in 2021 San Francisco, where I live, launched the Dream Keeper Initiative, which describes itself as a “citywide effort to reinvest $120 million over the next two years from law enforcement into San Francisco’s Black and African-American communities”. In fact it was largely a pretext for city officials to solicit bribes in return for sending city funds to their friends. The former head of the program, Sheryl Davis, is currently being tried for conflict of interest and misappropriating public funds. Most brazenly, Davis funneled millions of dollars to Collective Impact, a nonprofit organization she founded and formerly led, which paid the executive director salary of Davis’s live-in boyfriend James Spingola while the two shared bank accounts and expenses. As of now, most of the corrupt nonprofits involved have not been charged, such as the Homeless Children’s Network, which received $3.5 million from the program while paying kickbacks to Davis’s son.
The point of redirecting the money is precisely that it makes the money easier to steal, since the police are supposed to produce an output which can be checked, and the NGOs are not. (San Francisco police, like most American police departments, are notorious for making schedules which give officers large overtime payments as a matter of course—in 2024 about a third of officers received over $100,000 in extra overtime pay—but the officers do actually work a lot of overtime in order to receive this.) Many similar schemes were launched across the country around the same time, in large part pushed by people intending to take a cut for themselves and eager to distort the ostensible purpose to that goal.
Of course, when people are organizing to loot an organization, they don’t say so in those words. They’ll tell a nice-sounding story to everyone outside their faction, and cooperate to cover for each other—this consists less of explicit conspiracy and more of tacit cooperation on the basis of shared professional interest and ideological affinity, although there is some explicit conspiracy, and it’s disproportionately important when it does happen. Only when the fraudsters have triumphed completely and have no remaining fear of punishment will they openly defend the fraud without any reference to the organization’s ostensible purpose.
Another example, less lurid but as important as the looting of government budgets: Recently Arxiv announced a policy of banning authors for one year if they submit papers containing “incontrovertible evidence that the authors did not check the results of LLM generation”, including “hallucinated references”. Lots of academics attacked the policy. Some did so by lying, claiming that it is wrong to punish authors for making mistakes formatting their references, as though it were being applied to people who mistyped an author’s name rather than people who claimed to base their arguments on works which do not exist. More interestingly, though, many attacked the policy on the basis that it will hurt the careers of the many academics who routinely cite references without reading them, that “it’s very common for professors just to copy citations they found in other papers and put them into their own papers because they need a lot of citations to look credible” and therefore fake LLM citations should not be punished. Very obviously, inserting fake citations (whether done by human or LLM) is a practice which cannot possibly occur in a work that is constrained by attempting only valid arguments and is trying to say things only if they’re true. Very obviously, works which fail such a basic brown M&M test do not advance society’s knowledge. In fact they retard the pursuit of knowledge by drowning out real work. And Arxiv’s detractors do not contest this! Instead they argue that the policy will hurt the careers of people who acted in accordance with academic norms. Which is true so far as it goes, and would be a good reason to oppose the policy if the purpose of publishing papers was to provide comfy make-work jobs to people who abide by academic norms. But if the purpose of publishing papers is to produce and disseminate knowledge, then of course Arxiv should purge fraudsters who jam the frequencies with antiknowledge, most especially when such practices are the norm. (I suspect it’s not a coincidence that Arxiv made this move two months after announcing its intent to leave the university system and become an independent organization.)
From this, we can see that the cancer is more advanced in academic publishing than in San Francisco exostate “non-government” organizations. In both cases, the fraud rings have achieved enough internal control to steer the project to their own ends rather than the ostensible purpose. In both cases attempts to fix the “dysfunction” by pouring in more money make the problem worse because the people causing the problems are the ones who get the money—it’s like trying to get the ants out of your pantry by pouring in more sugar. The more money San Francisco gives to organizations claiming to treat drug addiction, the more addicts die of overdoses. The more money is given to academic historians or psychologists or Alzheimer’s disease researchers, the worse the outputs of their fields become. The institutions are too corrupt to turn resources into progress towards their ostensible mission. But in the case of San Francisco graft, people still have to lie. If they’re caught then a few of them will go to jail, and a few more will lose their government checks, at least until they move to a new 501(c)3. Meanwhile the publishing academics are so far gone that the corrupt ones don’t have to lie anymore.