Face Value
Ringside Seats to 3d Chess
Various companies developing generative AI have highlighted, directly or indirectly, existential risks from AI systems. Executives have signed open letters, pontificated on podcasts, etc. One common reply to these statements is something like:
Obviously, these AI guys are cynically trying to pump investors. They’re bragging with a thin veneer of social consciousness. They don’t really believe AI is going to threaten humanity, but saying “out product could end the world” is an efficient way to say “our product is the biggest deal ever”
I’m sympathetic to replies like “these guys are crazy” or “this seems like a cult” or even “I don’t trust these dudes”. But the block quote represents something beyond any of that: it not only declines to take the relevant claims literally, but also supposes a specific disingenuous motivation for making them.
When I saw people archly “seeing through” existential risk worriers, I found myself exasperated. I wished the commentators would stop trying to infer which complex social games AI companies were playing, and would instead take their claims at face value. Why?
I’d been in some of the same spaces as some of the people ringing AI alarm bells, and they’d been making the same arguments - including publicly - for years before becoming prominent
It didn’t really pass the sniff test that “our product could end the world” was a great way to build hype, relative to other options
It felt like the commentators were trying to project a vibe of being world-weary and in the know, but they didn’t actually know very much about the specific people and claims involved
I don’t actually want to litigate this specific case. I chose it because it’s one where I feel like I knew about the relevant community, and it felt plain to me that the commentators didn’t know what they were talking about. It’s far from the only example, though: people resort to “let me analyze this 3d chess move” argumentation all the time.
I think we should do that less.
Examples
Here are some more examples of people choosing to take claims as evidence of something complicated, rather than engaging with them directly:
A person says “trans women are women”, and a conservative commentator claims that they’re trying to groom children and normalize pedophilia, rather than simply believing that trans women are women
A religious person says “life begins at conception”, and a progressive commentator claims that they’re just trying to control women, rather than simply believing that blastocysts have immortal souls
A libertarian argues that ephebophilia and pedophilia are meaningfully distinct, and everybody assumes they’re motivated by a personal desire to sleep with teens
A poor person argues that rich people are hard working and should pay less taxes, and people assume that they secretly believe they will one day be rich, and are trying to set things up well for their future self
On reflection, all these reactions are weird. A random person with a favorable view of trans rights probably just has a trans friend, or trans feelings of their own, or has watched some documentary about a trans person that they found emotionally compelling. Why assume they have an interest in harming children?
Similarly, I’ve long been confused by how quickly people go to “they just want to control women” in abortion debates. It’s true that conservative, religious cultures tend to strongly constrain female sexuality, and maybe there’s an abstract sense in which specific beliefs like “blastocysts have souls” fall out of that, but the church lady who wants to “save the babies” probably just literally thinks she’s saving babies, right?
Arguments trying to litigate the precise boundaries of a justified taboo will always garner a knee jerk reaction (that’s just what taboos are), so I’m not all that interested in defending people who talk about “ephebophilia” all the time. But at least some of them have girlfriends (or boyfriends) they don’t want to swap out for arbitrary 15-year-olds, and really do have an academic interest in the evolutionary psychological implications of various patterns of attraction.
And the poor person favoring the rich strikes me as the silliest case of all. How did it become conventional wisdom that such people see themselves as “temporarily embarrassed millionaires”? Perhaps it happens sometimes, but it strikes me as much more likely that people advocating for lower taxes on the rich… actually, straightforwardly lionize the rich.
So why interpret things so obliquely? Wouldn’t it be easier to just say (or better yet, think) “I disagree”, rather than leaping right past the factual claim and into the murky terrain of the speaker’s motives?
Why?
Now that I’ve established that I’m against ignoring arguments and assigning arbitrary motives, let’s try to assign some arbitrary motives to arbitrary motive-assigners.
First, I think leaping to motive is a good way to enforce local taboos. I suspect that most readers bristled, in addition to whichever example is the closest match to their own politics, at the ephebophilia example. Because thinking carefully about the boundaries of pedophilia is unpleasant and undesirable, it’s easier not to do so. If you leap directly to an interlocutor’s motive, you can denounce them without engaging with their ideas.
Second, it’s fun to believe in conspiracies about your enemies. I don’t really want to rant about partisan politics or whatever, or even to imply that all such conspiracies are automatically false. But human beings tend to get a kick out of thoughts like “the people who oppose my policy preferences hate children”.
And third, analyzing motive is a way to feel above it all. It lets someone feel like Dr. House, or an oldschool hardboiled detective who always sniffs out the underlying reality of the case. With the added bonus that it’s very hard, when making claims about someone’s secret motivation, to ever be proven wrong.
Why Not?
I don’t know how helpful it is to shout “take arguments at face value” into a megaphone. Maybe it isn’t actually amazing life advice for arbitrary people, or even arbitrary readers of this blog. But I notice I’ve been taking more and more stuff in the discourse at face value, and it feels really good for me.
If I don’t want to get into some taboo topic, I prefer to just think “I’d rather not get into this”, vs. ruminating on what pathologies motivate the taboo peddlers; even if I happen to get the pathologies exactly right, thinking about other people’s flaws isn’t all that healthy an activity. Also, I probably won’t get them right.
If I want to have an excuse to dislike some faction I disagree with… well, first of all, I’d rather do less of that, all things equal. But in general, there are perfectly ordinary object level reasons available, and I don’t need to resort to imagining sinister behind-the-scenes games they may or may not be playing.
Also… the world is complicated. It’s nice when you really do have an excellently honed sense for some domain or other, and you can figure out with ease what’s really going on. Probably the best stock traders in the world really can sniff out bullshit in a quarterly earnings announcement, for example. But nobody knows all that much about everything, so you’ll much more often be muddling through than masterfully reading between the lines. Reading what’s on the lines is hard enough.
Finally, it just frees up space to take people literally. Not always, of course - this post has no bearing on how much to take literally in one’s personal life, for example. But when it comes to discourse, to strangers declaring things in public… all else equal, I’d rather just take them at their word. And if I know them not to be trustworthy, or to habitually lie? Then I probably shouldn’t listen in the first place!