‘Wordcels’ are ranked too low

On the hierarchy of intelligence or difficulty/rigor, ‘wordcels’ are ranked too low. And data/AI people are ranked too high. Although STEM and practical-minded stuff (e.g. business) are often positioned higher on the ‘intellectual totem pole’ compared to the ‘soft and subjective humanities’, my evidence shows it’s not so clear cut. With the exception of the highest tiers of math, one is not necessarily easier or less rigorous than the other. They are hard in different ways.

For example, of the thousand or so individuals that comprise ‘intellectual-sphere’ on Twitter, maybe only a hundred have blogs; and of those, 20 are competently written; and of those, a dozen are successful–as in having at least a thousand readers. It’s pretty rare.

So, come again about math being harder? Math can be hard, no doubt, but there are way more people who can do math well or who understand AI, than who can write well or are successful at writing. (Although this can be explained also by individual preferences or financial incentives; coding pays more.)

But what about computer science, such as AI? No doubt, there are some brilliant people involved, such as developing algorithms or the more theoretical aspects of machine learning, but in recent years it has become the equivalent of the ‘school group project’, in which the smartest person does the bulk of the work. I can understand in special circumstances an ambitious project needing multiple authors, but does every computer science paper on arXiv need a dozen authors for a 12-30 page paper, in which half the paper is footnotes? The trend of excessive authorship has gotten out of control. Even George R.R. Martin is listed as a co-author of a paper about dire wolves, as one of over 40 total authors, which also incidentally proves my point:

By comparison, writing is almost always a solitary endeavor. Sometimes literally, like Henry David Thoreau. Co-authors are much less common compared to in STEM. At most, there may be only a single co-author or ghostwriter, not a dozen as is common on the hugely popular /CS/ section of arXiv. Thus, with writing there is much more accountability and effort overall placed on the individual, whereas STEM, especially computer science, is more collaborative.

From my own experience tutoring math, Math typically has conceptual impasses that when cleared leads to progress at said skill or problem, so the issue tends to be poor instruction more so than insufficient intelligence, although IQ still matters at highest levels.

By comparison, what makes writing hard is it not only has to be correct and logical, but it also has to be conveyed in a way that pleases the reader, which is inherently subjective and unpredictable. With math, what matters primarily is that the result is novel and correct. A paper with poor exposition will still be published if it contains correct, interesting results. But writing is more like mind reading, in which you have to decode the recipient’s value system. It’s also more adversarial, because the reader is looking for any excuse to object.

As I argue in “‘Thought leader’ is just another form of elite status,” being a successful pundit or thought-influencer is much more uncommon and difficult than even being a highly competent and wealthy businessman, with a higher intellectual barrier to entry. The people who are the most successful at business and make decent money are not the smartest, even if they are still pretty smart and competent overall. (Although there are some exceptions, like Bill Gates or the founders of Google.)

Being able to competently converse extemporaneously on complex social issues and have opinions that are palatable to large audiences without the backing of a major brand (e.g. Fox News), is a higher intellectual bar (maybe >150 IQs) and rarer than business success (110-125 IQs).

So it’s not too surprising Elon Musk, despite being so successful and competent at business (such as ingeniously repurposing X/Twitter as an AI company, thanks to Grok), otherwise sometimes comes off as uninteresting or even dull when conversing about non-Tesla or Space-X matters, compared to the likes of Richard Hanania, whose intellectual repertoire is more broad. Elon cannot just cite history tidbits as Moldbug does.

If you compare the writings or tweets of Moldbug to Elon (or almost any vc person), there is almost always an IQ gap or hierarchy. If you read the bios of famous writers (or contemporary writers, or artists in general), they are all weirdos in some way or cut from an entirely different cloth, which you don’t see with other professions, even STEM. Some of this is due to trait-neuroticism, but I think IQ also plays a role, too, as investigated by Leta Hollingworth of extremely gifted children (IQs >160). For these people, writing or other creative outlets are an escape from a world that has been created by and for normal people.

Wordcels actually have original ideas that can become memetic. When tech/VC people are all talking about some theme or concept, like ‘wokeness’, or ‘abundance’, it’s almost certainly a wordcel who conceived it. Seldom does process go in reverse, a notable exception being Beff Jezos’ coinage of ‘e-acc’. But wordcels do not ‘build things’. Worldly, erudite people are not creating factories or stores that employ thousands of people and drive economic growth. So we need both: people who build, and people who create ideas or inspire action. But this has the downside of reading builders’ uninteresting or stale political takes.