Ed Zitron Is Not an A.I. Skeptic ⇥ theargumentmag.com

Kelsey Piper, the Argument:

At some point, pretending that how people use AI is a complete mystery is just lying to your audience. And at some point, [Ed] Zitron’s “layers of skepticism” attitude — where he is skeptical that AI is a thing at all, that it has any uses, that those uses provide any economic value, that the revenue numbers are real, that adoption is a fad, that training costs are a meaningful R&D expense, that the capital build-out is going to happen at all, that the market could sustain the capital build-out if it happened — leaves one buried in too many impossibility assertions to actually sort them by plausibility.

It is radical skepticism, ultimately arriving at “perhaps nothing we see is real,” rather than principled skepticism about the relatively weakest links in the companies’ case for investment.

My main problem with this piece is that it acknowledge Zitron’s own framing as an A.I. skeptic when he is not one. A skeptic is someone who asks good-faith questions and uses the answers to build an evidence-based view of something. They can separate reasoning from a narrative, while understanding the role it plays. For example, I do not regularly read the Argument because I think it a pretty mediocre website with an Atlantic-lite viewpoint, but that does not mean this article is itself poor or making an unfair case. I think Piper’s frustration with Zitron is entirely earned. However, it is a mistake to think of any of this in terms of skepticism when Zitron’s understanding is, especially now, much closer to conspiracy thinking.

Each of Zitron’s articles is an impenetrable wall of text often reaching into the tens of thousands of words. This volume of material feels substantial — it can be substantial — to the extent Zitron explicitly markets his newsletter on the basis of its word count. Weird.¹ He explores basically two major themes: A.I.’s economic case, and its usefulness. Zitron is zealously opposed to the possibility of either in real terms; while he will occasionally gesture at people or businesses doing something with A.I., his default position is more-or-less that it has little use.

There are real criticisms of what generative A.I. does: problems with its output, like its repetition of stereotypes or bugs in code. There are criticisms for what it does to our world, like its energy and water consumption, and what it does to society, like how easy it is to generate junk articles and videos. The societal pushback is also notable, and its unethical foundations continue to be a ripe source of pain. But, as Piper writes in the second footnote, Zitron’s articles are “a superficiality of analysis” despite the voluminous output. Like a lot of conspiracy thinking, they are rooted in fact but have a stricter adherence to supporting an existing narrative.