A low-tech solution

Thanks so much to everyone who offered help and support as this blog’s comment section endured the weirdest, most motivated and sophisticated troll attack in its 17-year history. For a week, a parade of self-assured commenters showed up to demand that I explain and defend my personal hygiene, private thoughts, sexual preferences, and behavior around female students (and, absurdly, to cajole me into taking my family on a specific Disney cruise ship). In many cases, the troll or trolls appropriated the names and email addresses of real academics, imitating them so convincingly that those academics’ closest colleagues told me they were confident it was really them. And when some trolls finally “outed” themselves, I had no way to know whether that was just another chapter in the trolling campaign. It was enough to precipitate an epistemic crisis, where one actively doubts the authenticity of just about every piece of text.

The irony isn’t lost on me that I’ve endured this just as I’m starting my year-long gig at OpenAI, to think, among other things, about the potential avenues for misuse of Large Language Models like GPT-3, and what theoretical computer science could contribute to mitigating them. To say this episode has given me a more vivid understanding of the risks would be an understatement.

But why didn’t I just block and ignore the trolls immediately? Why did I bother engaging?

At least a hundred people asked some variant of this question, and the answer is this. For most of my professional life, this blog has been my forum, where anyone in the world could show up to raise any issue they wanted, as if we were tunic-wearing philosophers in the Athenian agora. I prided myself on my refusal to take the coward’s way out and ignore anything—even, especially, severe personal criticism. I’d witnessed how Jon Stewart, let’s say, would night after night completely eviscerate George W. Bush, his policies and worldview and way of speaking and justifications and lies, and then Bush would just continue the next day, totally oblivious, never deigning to rebut any of it. And it became a core part of my identity that I’d never be like that. If anyone on earth had a narrative of me where I was an arrogant bigot, a clueless idiot, etc., I’d confront that narrative head-on and refute it—or if I couldn’t, I’d reinvent my whole life. What I’d never do is suffer anyone’s monstrous caricature of me to strut around the Internet unchallenged, as if conceding that only my academic prestige or tenure or power, rather than a reasoned rebuttal, could protect me from the harsh truths that the caricature revealed.

Over the years, of course, I carved out some exceptions: P=NP provers and quantum mechanics deniers enraged that I’d dismissed their world-changing insights. Raving antisemites. Their caricatures of me had no legs in any community I cared about. But if an attack carried the implied backing of the whole modern social-justice movement, of thousands of angry grad students on Twitter, of Slate and Salon and New York Times writers and Wikipedia editors and university DEI offices, then the coward’s way out was closed. The monstrous caricature then loomed directly over me; I could either parry his attacks or die.

With this stance, you might say, the astounding part is not that this blog’s “agora” model eventually broke down, but rather that it survived for so long! I started blogging in October 2005. It took until July 2022 for me to endure a full-scale “social/emotional denial of service attack” (not counting the comment-171 affair). Now that I have, though, it’s obvious even to me that the old way is no longer tenable.

So what’s the solution? Some of you liked the idea of requiring registration with real email addresses—but alas, when I tried to implement that, I found that WordPress’s registration system is a mess and I couldn’t see how to make it work. Others liked the idea of moving to Substack, but others actively hated it, and in any case, even if I moved, I’d still have to figure out a comment policy! Still others liked the idea of an army of volunteer moderators. At least ten people volunteered themselves.

On reflection, the following strikes me as most directly addressing the actual problem. I’m hereby establishing the Shtetl-Optimized Committee of Guardians, or SOCG (same acronym as the computational geometry conference ). If you’re interested in joining, shoot me an email, or leave a comment on this post with your (real!) email address. I’ll accept members only if I know them in real life, personally or by reputation, or if they have an honorable history on this blog.

For now, the SOCG’s only job is this: whenever I get a comment that gives me a feeling of unease—because, e.g., it seems trollish or nasty or insincere, it asks a too-personal question, or it challenges me to rebut a hostile caricature of myself—I’ll email the comment to the SOCG and ask what to do. I precommit to respecting the verdict of those SOCG members who respond, whenever a clear verdict exists. The verdict could be, e.g., “this seems fine,” “if you won’t be able to resist responding then don’t let this appear,” or “email the commenter first to confirm their identity.” And if I simply need reassurance that the commenter’s view of me is false, I’ll seek it from the SOCG before I seek it from the whole world.

Here’s what SOCG members can expect in return: I continue pouring my heart into this subscription-free, ad-free blog, and I credit you for making it possible—publicly if you’re comfortable with your name being listed, privately if not. I buy you a fancy lunch or dinner if we’re ever in the same town.

Eventually, we might move to a model where the SOCG members can log in to WordPress and directly moderate comments themselves. But let’s try it this way first and see if it works.