Autism & the Internet will defeat the Monoculture

Or why the most original ideas in the coming decades will come from the Internet

“Is the Internet the Enemy of Progress?”, asks Ross Douthat in one of his latest columns. The answer is an emphatic Yes. For Douthat this is just a symptom of a more widespread disease: according to him, the Internet has destroyed something deeper and more important: the diversity of human thought itself. He goes on to link this rise of a Monoculture to various negative outcomes of the last decades, from cultural stagnation to falling birth rates. The Internet is turning us all into similar-thinking drones, and anxious drones who want to look like Kim Kardashian at that — but without the high fertility characteristic of the famous family. Truly the worst of all worlds.

There are many accusations that can be levelled at the Internet, and intellectuals these days seem to be in a race to come up with new ideas in this direction. But that of enabling a Monoculture seems to me entirely incorrect. It’s important to note that a mainstream culture would exist with or without the Internet¹. What an online environment adds on top of this is a powerful engine of cultural evolution that makes disruptions of the mainstream more likely. The Internet is prolific, churning out new ideas at a speed that would be hard to imagine for any other medium. And its selection forces are brutal: unecumbered by institutional inertia, it eats its own children alive with no remorse. Perhaps its most important power, however, lies in its human capital. The Internet has attracted a specific type of people that often face difficulties in the “real world”: that is, those on the autistic spectrum. I use the term "autistic" here in the way researchers like Michelle Dawson have been advocating for decades now: not purely as a pathological condition, but rather as an atypical way of processing information, that comes with both challenges and advantages. What Dawson was hoping would come from a more just and wiser society — that is, an attempt to harness the very real and unique strengths of autistics, as opposed to simply “treating” them — has been brought about by technology and the emergence of The Internet. This is a place where the challenges faced by autistic individuals are diminished and their strengths become more prominent. Among these strengths I count their propensity to question mainstream cultural norms: precisely what one needs in order to challenge the Monoculture. While isolation can magnify the struggles associated with their non-conformity, the Internet's ability to connect autistics regardless of geography allows them to spontaneously coalesce. And if the Internet saves autistics by providing them with a refuge, autistic online movements might just save us all from the Monoculture.

What I have described above is not just a thought experiment, but a reality. Say what you will² about rationalism, the niche (and autistic) internet ideology clustered around blogs like LessWrong and SlateStarCodex and its associated movement, Effective Altruism³. What I think is hard to deny is that it represents exactly what Ross Douthat deems as impossible: an innovative, unlikely, “out-of-nowhere” ideology which emerged on the Internet on niche blogs and created an influential culture with real world impact⁴. The AI risk branch of the rationalist movement⁵, is a prime example of this phenomenon. By obsessing over the creation of the “Machine God” long before the mainstream took AI seriously, they catalysed real world action in this direction. Dario Amodei has discussed how the AI safety movement inspired him to start Anthropic; Sam Altman, the CEO of openAI has said (perhaps in a moment of exaggeration), that Eliezer Yudkowsky, arguably the most influential rationalist, deserves a Nobel Prize for motivating him and others to pursue AI development. On the commercial “innovation” front, one of the most spectacular frauds of the 21st Century, FTX, was motivated explicitly by its founder’s commitment to Effective Altruist principles (hey, I did not say it was all positive!) Rationalism holds a dominant position in discussions about AI safety, often implicitly. Its proponents have been exploring this topic for decades and this longstanding focus has influenced real-world policy, as many adherents now occupy influential policy-making roles⁶ and play a significant part in shaping legislation, such as the recently proposed SB 1047 Bill in California.

There’s more power in autism than in numbers

---

In his article, Ross Douthat anticipates some objections to his points and argues that the non-conformist subcultures the The Internet creates, vivid and rich as they may be, are too small, too niche, too cloistered to have a real-world impact. His skepticism is not without merits, if one judges from first principles: after all we should expect a niche movement of online weirdos to remain confined to The Internet. But rationalism clearly defied this prediction. The question is: why?

The traits that make autistics better at questioning the Monoculture are obvious: an explicit premium placed on transcending common cognitive biases (indeed, one of the more influential rationalist blogs is called Overcoming Bias), a keen desire to systematise information across a broad range of topics and a commitment to engaging with points regardless of their provenance. But these are not the traits that make a movement successful: if anything, they seem designed to scare away normal people and drive any movement into the ground. The drivers of success must be sought in other places.

An old picture of Robin Hanson and Eliezer Yudkowsky, who used to blog together at Overcoming Bias, one of the most influential rationalist blogs. As much as I vibe with them, these are not the aesthetics of a movement that will take over the world.

Autistics care. Disaffected ones generate ideas. Successful ones implement them.

---

In his “Why is everything liberal” and “Liberals read, Conservatives watch TV” essays, Richard Hanania argues that Culture belongs to those who care and engage with it. This is the answer he has to the conundrum that despite being roughly equal in numbers to Conservatives, Liberals dominate Culture. While the median Conservative likes to grill and watch relatively low quality TV, Liberals care enough about their ideology that they engage with higher quality, written liberal media and are willing to make personal and financial sacrifices to stand for their beliefs — including taking low paid high status jobs like journalism, which propagate their ideology. Taken together, these realities lead to the emergence of a thriving intellectual and cultural Liberal ecosystem and a cranky, shady and much less influential Conservative one. These insights can be generalised beyond Conservatives versus Liberals: ideologies are powerful not merely by virtue of the numbers of converts, but how much these converts care and how much they are willing to engage with the specific medium through which culture is spread.

There is a group of people who care even more than Liberals: autistics; for them caring often turns into obsession. Thirteen years ago, Tyler Cowen framed these qualitative observations in economic terms: he argues autistics should be expected to have lower price elasticities and to specialise in production to a greater degree. That is, what they enjoy they enjoy much more than the average person and what they dislike they also dislike more, which makes them less likely to respond to differences in pay when choosing a job and place more value on their enjoyment of it — sometimes to the point that enjoyment is the only thing that matters. This explains why, unpaid and unrecognised, anonymous writers on forums like LessWrong have written thousands of words of often highly technical posts on topics ranging from Economics to AI safety. As an example, this is a completely serious, 24 min reading time (!) post made in response to Eliezer Yudkowsky’s — random speculative tweets about how we risk being killed by AI via diamondoid bacteria (something that does not exist):

What is funny is that “diamondoid bacteria” is a concept that does not actually exist and is invented by Eliezer himself. However, the concept is treated with utmost seriosity by this poster, who patiently speculates what Eliezer *might* mean.

There's a saying that for any unconventional question related to economics, social science or AI/AI safety, you're likely to find a discussion about it on a forum related to the rationalist community. This multitude of posts has given rise to a culture that is surprisingly rich and diverse, especially considering the relatively recent emergence of this movement. Of course, this commitment with the topic of their interest does not stop at posting online: when those influenced by rationalism have the chance to act in the real world, they end up embodying those principles — see effective altruist Helen Toner and the OpenAI “coup.”

But Caring is not enough: there is another crucial dynamic tied to autism that led to rationalism having actual, real-world impact. Namely, that autistics are high variance in outcomes and some of them do actually achieve extreme tangible success:

In any case, if you take these cognitive abilities and disabilities and stick them into a rapidly evolving market economy, you get some people who achieve relatively high status and other people - many others - who end up with much lower status. That’s my basic view of autism as a social phenomenon, namely that for reasons rooted in perception and cognition there is a very high variance of outcomes across individuals

(Tyler Cowen - The Age of the Infovore)

At its core, rationalism has to be understood as a movement of autistics disaffected with existing institutions who care a lot and generate content. Their visions are funded and implemented by autistics who have managed to harness their specific strengths towards tangible, real world success. It’s not random that the epicenter of rationalism is San Francisco, a place where people with autistic traits are much more likely to succeed and achieve wealth in the tech or tech-adjacent industries and, feeling an affinity for the movement, donate to it. It would be a mistake to consider very rich autistics as the only real world catalysts of rationalist success: it’s also the countless software engineers who have thoroughly absorbed the vernacular of the ideology and have imbued arguably the most influential innovative and important American industry with its spirit. Some of these software engineers, raised with the spirit and concerns of rationalism, go on to become important players: as mentioned before, Dario Amodei has credited AI safety fears with motivating him to start Anthropic after having worked for OpenAI.

None of these strengths would have mattered if disaffected autists wouldn’t have been driven away from the Real Life to the Internet.

The Real World Autism Glass Ceiling & Escape to The Internet

---

Give me your tired, your poor, Your huddled masses yearning to breathe free, The wretched refuse of your teeming shore.

The well known verses inscribed on The Statue of Liberty could well be applied to The Internet. It’s a place where prior achievements and one’s exact identity become less important, so those who find solace and an escape here are often the refuses of Real Life. If not outright losers, they are lacking in some important way when it comes to their offline life, even when they might seem functional from the outside. Much like an America of the 17th century, The Internet is full of conmen, of persons of dubious character, of those who for good reasons have been isolated by society. Yet it’s also a land of opportunity for some honest people, a prime example being the underachieving, or at the very least, frustrated autist.

Most workplaces are structured by and for non-autistic individuals, often making it challenging for those on the autism spectrum to leverage their unique abilities in professional environments. While autistic individuals may excel in academic settings or in entry-level positions that prioritize analytical skills, their career advancement frequently stalls. This stagnation largely results from the increasing emphasis on conventional social skills necessary for climbing the professional ladder, particularly in roles that focus on client relations at higher levels. In many professional settings, autistics very often face a real “Glass Ceiling.” It is this relative lack of success and inadequacy that drives many to the Internet. Eliezer Yudkowsky himself is a high school dropout and seems to never have pursued any “normal job” or education: a failure by traditional metrics. Of course, this has recently changed, with TIME Magazine having named him one of the most influential people in AI in 2023.

Considering this, it’s probably no wonder that rationalism comes with a certain skepticism of established institutions — with the good and bad that this entails. The good is hopefully by this time obvious. The bad includes a tendency to dismiss accumulated knowledge if it comes from traditional sources and to rely on first-order reasoning at the expense of empiricism, often to absurd degrees. If one operates outside universities with labs and big funding, one has to rely almost entirely on thought experiments, and in time that can warp the mind. A consequence of this “anti-empiricist” attitude is that I consider rationalists to be at their worst when they have to reason about the world of atoms: from coming up with fantasist scenarios of how AI will kill us via “molecular machines” to seemingly disregarding why we do need clinical trials for assessing drug efficacy. If future movements are to influence progress beyond the world of bits, I think they need to stay away from this attitude.

Autism & The Internet will defeat The Monoculture

---

The rationalist example provides a broad blueprint for any autistic driven online movement. I think we are just at the beginning when it comes to such ideologies — already, a backlash to rationalism, or at least its AI risk arm, has started to emerge. Again, on the Internet⁷. That is the e/acc movement and associated anti-AI regulation think-thanks and organizations. I know Robin Hanson thinks the Internet (or anything for that matter) won’t save us from our low fertility promoting Monoculture. However, I’d like to remind him that his worries about low birth rates have been propagated mostly online and would have probably been next to impossible to voice otherwise. I believe there is a decent chance such concerns will gain more traction and will percolate into the mainstream — with signs of this already happening.

I anticipate that cultural evolution will accelerate on the internet because academia, traditionally a sanctuary for autistic individuals, has become increasingly inhospitable. As academic environments adopt more corporate-like, administration-centric approaches, they become less accommodating to them. This goes to back to an argument I made in an earlier piece of mine: The Flight of the Weird Nerd from academia, where I used the term “Weird Nerd” to substitute for autistic. If our salvation from The Monoculture is to come from anywhere, it won’t be from “real world”, established institutions — the sclerotic nature of which many critics of the Internet themselves often complain about. It’s going to come from the Internet and most likely, from people with autistic-like thinking styles⁸.

1. I’ll remind everyone that Europe was very Christian for centuries…

2. This post is not an endorsement of rationalism per se — I have in fact criticized some aspects of it extensively in the past. Rather, it’s intended as an analysis of the factors that enabled what is undeniably a success story.

3. Rationalism and Effective Altruism are not exactly the same thing, and there could be an entire post discussing the differences between the two. However, for the purposes of this essay, they are intertwined enough that I’ll discuss them together.

4. Some might doubt that rationalism has been truly that influential — this is not the place to convince them and the quotes and examples I have presented are not really intended to provide an exhaustive argument for it. Clues are scattered across thousands of twitter posts, group chats and niche forum posts, which makes collating the evidence particularly challenging. Arguably, the person who has advanced the most in this endeavour is NYT journalist Cade Metz, so those interested could check out his book — keeping in mind that he’s not a very impartial critic of the movement.

5. This is quite ironic, since one of the overarching focuses of rationalists has been risk from AI. Nevertheless, it seems like having talked so much about the “Machine God”, even in a negative way, has ended up inspiring quite a lot of people to try and create it.

6. Most notably, Paul Christiano has been named as U.S. AI Safety Institute’s (AISI) Head of AI Safety. But there are many, many others.

7. There is also a more mainstream opposition to rationalism coming from existing institutions, but I do not expect that to create anything particularly interesting or innnovative.

8. Something I did not address in this post is other influential movements that have emerged from the Internet in the last decades. Arguably, the so-called “Great Awokening” was catalysed on Tumblr and a lot of far-right extremism was also revived on the Internet, on forums like 4chan. These topics are addressed by Angela Nagle in her book Kill All Normies. While these movements have been influential, I do not consider them entirely “novel”, in the way rationalism has been. Rather, I think what happened is that existing ideologies were transformed into something more extreme on the Internet.