Why was I Banned?
I was banned on Twitter a few weeks ago, the reason was for violent speech. Samswara wrote this tweet. In response I posted a gif of Jamie Lannister—a guy from Game of Thrones—saying “burn them all.” So before I go any further let me be absolutely and totally clear with you reader: I do not think we should burn the homeless. Actually, Jamie Lannister himself never wanted to burn anyone; he was quoting the Mad King when recounting a story.
Was I thinking about all this context when I posted that tweet? Yes? Sort of? Samswara wrote that the ‘Right wing answer is to kill them all, which reminded me of that scene from Game of Thrones. I thought it was funny in a sardonic ‘South Park’ sort of way. Kind of juxtaposing the insane right wing claim to kill them all with the Mad King from Game of Thrones. It was a wry cultural reference. My apparently incorrect understanding of Twitter was that meme gifs and cultural references weren’t interpreted as calls to violence.
Look, I’m not saying it was hilarious, but it was just one of many, many random gifs I post daily. I wasn’t inciting violence, again, we definitely should not burn the homeless. Within my friend group on Twitter though, it was interpreted as a silly joke. Unfortunately, I suppose other people who don’t know me interpreted it differently.
My Twitter
I first began using Twitter late January of 2020 when I was losing my mind as there was this clearly serious virus taking off, and no one was paying attention. Or worse, they were saying things like “the science says travel restrictions don’t work” or “you’re being racist against the Chinese.” I had never used Twitter before, but I was active in the greater Slate Star Codex world, and it seemed a handful of them were on Twitter.
Where it made sense, I’d chip in my ideas. Eventually Eigenrobot made it onto my feed, and that was my entry point into this whole new social world.
I had blogged for a long time, but no one ever read anything I wrote. For the first time it felt like I had some traction. In some sense I never cared about getting more followers. Or more specifically, I didn’t go out of my way to do things to maximize engagement. I wrote what I thought was interesting. I wrote about my thoughts on the philosophy of science, on learning to code, on data science, and other nerdy subjects.
I specifically avoided culture war, hatred, quote wars, flame tweets; these things get followers and engagement, but it’s low quality, and in any event it’s just not my style.
While I never optimized to get followers, it was a source of pride to have such high quality followers. Whenever I looked through my follower list I saw lots of engineers, scientists, and other impressively smart people. Not a lot of people who followed me spent their time and energy being hateful, that wasn’t what my simcluster was about.
Initially, amid the excitement and mania of losing an account I’d invested so much time in, I thought it might be fun to make a new account. Post new banger threads to get my followers back… But honestly after a few weeks, it’s tough to want to invest time into quality content again. I’ve had years of thoughtful, long form threads, removed from general access and taken away from me. Will they just be taken from me again?
I’ve had people message me that they were sad that they lost bookmarks to things I wrote, which is both a great feeling to know things I wrote had that effect on people, and also really, really sucks to lose all that investment of my time and intellectual energy. I invested years into writing on my old account, and it made me new friends, I was quoted on MarginalRevolution, I had interesting people messaging me on Twitter to meet up, I even had friends from real life surprised when they found out they already followed me on Twitter.
More than anything, creating a platform is hard, it’s a grind and it takes time. I was a fantastic poster for years, had an excellent reputation as a reasonable guy, paid for Twitter, was unusually kind, and losing all that investment of time, money, and energy, with no ability to appeal or regain access to my old writings is heartbreaking.
This is not Elon’s Fault
Contrary to some of my joke tweets, this is not Elon’s fault and it obviously has nothing to do with him. My tweet looked like a simple call to violence, and whoever reviewed it took it at face value, and that was the end. Doing this sort of stuff at scale is hard.
For example, I think there is a good chance people were fired at my previous job because of arbitrary hyper-parameters I set on a model. Should the growth rate constraint on the parameter be .6 or .65? When you’re predicting at scale, that might be the difference between needing 100 vs. 140 contractors. You simply can’t take a look at every edge-case, it’s not possible.
We’re at a unique point in time where we can’t scale truly high-dimensional exception handling. Are we on the cusp of it? If I could talk to a twitter LLM agent, would they unban me? I’m not sure. I’d like to think that they would, because against the context of the rest of my account, which has been sincere and thoughtful, I think I deserve the benefit of the doubt here. Whereas if my account was just constantly retweeting Groyper1488 memes, probably I wouldn’t deserve the benefit of the doubt.
In any event, Twitter doesn’t have an LLM agent yet. Just humans. So if you’re a human at Twitter, or you know one well enough to share this with them without it being awkward, I’d definitely appreciate another chance. I’ll post fewer gifs.
(What should we actually do about the homeless?)
This is maybe a strange thing to write about on this post, but I think it’s worth actually answering the question (I have lots of threads on my old account, which I can’t link to anymore).
Essentially I believe it’s our collective responsibility to not permit people to slowly kill themselves on the street, and that fundamentally does require coercion. I also don’t think prison is the right place for these people.
The painful truth is many of the visible and suffering homeless, often due to childhood trauma or an unfair genetic disposition to mental illness, will never be capable of taking care of themselves. There is no “just get a house and some services and then you can get a job.” It’s that they need lifelong care. Our care is so often either being strapped to a chair and force-fed Seroquel, a prison cell, or dying of a fentanyl overdose on the street.
My solution would be to create communities outside of major cities, with different levels of care, creature comforts, and also make-work for those who want the opportunity to cultivate or build whatever it is that gives some small amount of meaning to their lives. These would be used as alternatives to prison, but mandated by a court.
This isn’t my idea, Orwell wrote about it in Down and Out in Paris and London. He temporarily lived as what they called then a tramp. The life of a tramp then was different than on the West coast of the US, as they were not permitted to sleep in public. Instead they were shuffled from boarding house to boarding house, and if they could not afford one, they were locked in a cell. Orwell wrote that what the tramp needed was to be “depauperized and this can only be done by finding him work—not work for the sake of working, but work of which he can enjoy the benefit
What we need to provide them with—those who are capable—is work for their own benefit. Of course in Orwell’s time there was no fentanyl. It’s an open secret that most opioid addicts cannot quit; less than half of those who enter treatment make it more than a year sober. There is no temporary housing, followed by a minimum wage job for most of our visible homeless. They are deeply traumatized and mentally ill persons with an addiction that kills most of its users.
What we can offer them is a life where we protect them from themselves, and from the criminal elements that would sell them the substances which rot them to death, with the ability to find some. meaning through work. While it’s maybe cliched that the pastoral can cure all social ills, Orwell wrote: “Yet there is a fairly obvious way of making them useful, namely this: Each workhouse could run a small farm, or at least a kitchen garden,” as an example of how to create that meaning.
Curiously, what has been your thought process about your choices on Twitter? I'm especially thinking of pseudonymity, grinding for highbie status, and saying things that might get one in trouble (either one's account, or one's IRL self in case of doxxing). I've chosen to avoid pseudonymity, which has the downstream effects that I minimize usage during work hours and that I avoid saying things that could (due to misinterpretation or not) get me in trouble. With regards to the latter, my two criteria are, (1) while I provide enough evidence of my libertarian leanings that an HR person or journalist might _want_ to cancel me, I try not to provide evidence that would _enable_ them to cancel me, and (2) while I say things my spouse might disagree with, I won't post something that, if it were to get me in trouble, I would be embarrassed to explain to her (either as overly objectionable or as needlessly reckless). These restrictions on myself do mean that I have a lot of thoughts that come to me that I don't post, and I do occasionally have FOMO about it.
Sorry to hear you went through that. I was one of your followers (I left Twitter in the last few months) and I always enjoyed your threads.