Who is arguing for?
A verdict on the week — plus the week's top stories.
Welcome to the The Closing Argument, our verdict on the news, plus everything The Argument published and appeared in this week. If you enjoy this newsletter, please forward it along!
The Verdict
A couple of weeks ago, Bloomberg’s Joe Weisenthal lamented “the fact that one can get an LLM to put together a cogent-sounding argument in favor of virtually any position has left me feeling cynical about human debate and discourse.”
He concluded something quite dark for a publication called The Argument: Debate is “(almost) all a game wholly sealed inside language itself without tether to the world.”
That we have moved from the literacy age to the age of digital orality is the broader theme Weisenthal has been circling which pulls together threads from intellectuals like Walter Ong and Marshall McLuhan.
Writing and reading made engaging with very complicated ideas possible. For a facile example, most people can do very simple addition in their heads, but at some point we all have to tap out, maybe it’s at multiplying double digits, maybe it’s multiplying triple digits, maybe it’s factoring quadratics, whatever it is at some point everyone needs to start using pencil and paper in order to solve for x.
This shift to using written communication did more than just enable more complicated mathematics, it made space for individuals to wrestle with ideas on their own. Sitting down and reading a book or writing an essay encouraged people to engage with ideas on their own terms without the persistent social pressure that the immediacy an in-person conversation demands: “The written word creates all kinds of new opportunities to think [things] through…to take time, to not respond right away,” Weisenthal recently explained.
Now, even before mass access to chatbots (where were you on November 30, 2022?) Weisenthal has long been skeptical about the value of debate. In May 2021 he argued that “unless there’s an extremely high level of shared premises, knowledge and mutual understanding, you almost always just get people talking past each other without anything gained.”
Weisenthal’s pessimism about written debate isn’t entirely wrong. It used to be the case that the quality of writing had a reasonably linear relationship with the quality of the underlying thinking. That has well and truly been severed as I pointed out last spring.
But Weisenthal’s cynicism misunderstands the primary purpose of public arguments. The two individuals arguing are actually secondary to the vastly more important third actor: The silent audience.
When I first began writing in public, I struggled with the question of when to respond to my detractors. Some people would tell me “there’s no point, you’re not going to change their minds,” which I concede is often true. But when I did respond—mostly because I am constitutionally incapable of allowing someone to mischaracterize my arguments—I noticed something interesting: Other people’s reactions.
When talking to friends, when texting with sources, when meeting up with other journalists at the bar, I realized how many people are passively reading online arguments and making judgements and updating their own views.
As someone who did college debate, this is all remarkably familiar. Debaters don’t ever expect the other team to give up in the middle of the round and admit defeat. That’s not the point. They’re trying to convince the judge!
Weisenthal is right that an LLM can generate a cogent argument for virtually any position. But it can generate that argument only in a vacuum. The function of debate is to ensure that arguments don’t survive in a vacuum; they survive only if they can withstand the strongest available counter-argument, in front of an audience capable of comparing the two.
This adversarial process is imperfect, but it’s the best thing we have. It’s why we have cross-examination in courtrooms, peer review in science, and opposition research in politics.
In college debate, it was normal to dread being matched against a team that was very good, but it was also normal to dread being matched against a team that was very bad. The reason is that judges rarely give you very high speaker points for curb-stomping terrible arguments.
In one memorable round, I opposed a team that argued the U.S. should pay reparations to Iraqi citizens. I was eager to debate this motion because it’s smart, and offers a lot of room for debate along economic, philosophical, and legal lines.
But then the opposing team decided to propose giving 1/3 of U.S. GDP to victims of the Iraq War. At the time this amounted to more than $6 trillion.
The usual advice given to debaters in this situation is to try to build up the other team’s arguments so as to be more impressive to the judge.
The underlying truth of this dynamic is that people are much more likely to be persuaded by beating smart arguments than stupid ones. This brings me back to the question of what will happen when people use AI to enhance their arguments.
I expect this to actually raise the quality of written argument, not reveal it to be a stupid game. If my opponents in that debate round had asked Claude about their case, there would have actually been a much better debate. (Click here to see Claude’s advice).
Let’s think through the counterfactual. Absent chatbots, people were, for the most part, just shooting off the dome, not doing extensive research. But I think the strongest version of Weisenthal’s argument is that people will use AI to make their arguments sound stronger even as they remain vacuous drivel.
This is quite bad! But since everyone will have access to AI making their arguments sound cogent, the real differentiator for the moveable observers will become the arguments themselves.
The biggest concern, I suspect, is that AI will displace thinking. On the margins, there are certainly some people who would have otherwise done some research and wrestled with ideas and arguments more, but I think this population is actually quite small.
Weisenthal is right that AI has severed the link between writing quality and thinking quality. But he’s wrong that this exposes debate as a language game untethered from reality. What it actually does is raise the floor. The worst arguments will get better, and that means the best ones will be forced to improve, too. For the silent audience–the people actually updating their views—this is a clear W.
It does make debates less fun for debaters, but no one really cares about us.
Top stories this week
As we grow, I want to make sure you see everything we’re doing without flooding your inbox with dozens of emails. But for the real libs, you can get every post as it drops by opting into The Mag here.
🌟Abundance Wins of the Week🌟
FAA Launches Air Taxi Pilot Program Across 26 States
PBM Reform Signed Into Law — Drug Pricing Gets More Transparent
Worth watching...
Every Wednesday, Lakshya Jain, our director of political data, will go live on Substack with leading election nerds to discuss what’s going on in politics that week.
This week, Lakshya joined Split Ticket’s Armin Thomas and VoteHub’s Zachary Donnini to discuss what’s happening in the Maine Senate Race.
Subscribe to The Mag so you don’t miss future live election breakdowns.
What’s News with The Argument
We’re hiring a Senior Editor to join our team. Are you a liberal? Do you love to argue? Do you read social science research for fun? Apply!
We have merch!
The Argument recommends
Here’s what our staff is watching, reading, and listening to this week:
Books:
Free Soil, Free Labor, Free Men by Eric Foner
Never Let Me Go by Kazuo Ishiguro
After Virtue by Alasdair MacIntyre
Articles:
"Are Hard Choices Cases of Incomparability?" by Ruth Chang
Music:
“All My Friends,” LCD Soundsystem
“D.A.N.C.E.,” by Justice
I'm very sympathetic to the faith in reason that Mr. Demsas exhibits here (and to the entire spirit of The Argument). But the degree to which her own arguments here will be valid depends very much on the commitment (or interest) of the audiences in reason. To the degree that these arguments occur online or in some other debate context that involves more or less neutral, unscreened public audience members, the ultimate goal is not really to make the superior argument, it's to persuade those who may be persuadable. Reason often comes in second to rhetoric.
"ChatGPT, give me good arguments for X in the debate context Y, and as back up supply ad hominem snark or deflecting segues that will have positive persuasive force for the likely undecided audience members in such a debate context." I think real life audiences are very little like disinterested debate judges who have set aside their priors and who follow rubrics governed by logical consistency.
Deflecting tactics may have more negative than positive force on silent neutrals than I assume. But in my own experiences observing and participating in social media with diverse audiences and "like" functions, I've found the number of likes generated by deflecting moves and snark, as compared to rigorous argument, usually prevails. This isn't the same as measuring the reactions of persuadable readers, but it's certainly not encouraging. (Of course, I'm not sure AI would actually outperform people in using these tactics -- they aren't demanding.)
I think Jerusalem here slightly misunderstands Wiesenthal's position, or at least the interpretation of Wiesenthal's position that I'm personally most worried about, which is not primarily about writing quality as a proxy for take correctness.
If we had LLMs steelmanning everyone's arguments to find the hardest-to-assail version of them, then yes, that's an asymmetric advantage for whoever's actually right, reduces the value of various forms of sophistry, and ultimately serves to improve the quality of debate for truthseeking purposes.
And maybe in another six months we'll have LLMs that can do that. But today's LLMs are decidedly mediocre at analytic rigor; good enough for some purposes, but not good enough to much help an intelligent reader figure out the right answers to complicated questions about politics and policy. (See, e.g., the final section of https://www.slowboring.com/p/ai-progress-is-giving-me-writers; apologies to the Slow Boring non-subscribers, as it's paywalled.)
What they're very good at is smooth-sounding prose that superficially appears to hold together. My worry is that this makes it harder to spot flaws in an argument; not impossible, but you have to read a lot more carefully and develop fairly specific skills for detecting rhetorical cover-ups. I've been trying to get better at this but so far I'm not that good at it, and consequently find judging LLM-assisted arguments (in the sense of being able to go like "that's wrong, here's specifically why") to be harder than the fully human-generated ones that I used to typically run across.
Granted, there've always been specialist humans who developed the skill of hiding flaws in their arguments behind this kind of rhetorical slipperiness, but they used to mostly only show up in high-profile contexts, because doing it well was a rare skill. Now everybody has a machine in their pocket that can do it on demand, and *that* seems to me very plausibly bad for readers' epistemic health.