Imagining The QAnon Of The AI Era

Read to the end for a really good home office setup

February 10, 2023

What Is The Revolutionary Use Case For AI?

On Wednesday, I wrote down a few predictions I had about how the emergence of AI-powered search platforms might interact with our existing media landscape. And I got a lot of comments about it! People messaged me things like, “grim,” “bleak,” and “reading this made me I want to die,” etc.

So, today, I want to flip things around a bit and walk through another question I’ve been wrestling with. Namely, how could AI be used for grassroots or populist activism?

It might seem like a funny question, considering how so much of the discussion around AI right now is about how it will reenforce monopolies and totalitarianism, but I think it’s worth considering. All technological revolutions arrive in tandem with social or political ones. Could Occupy Wall Street, the Arab Spring, Black Lives Matter, the Hong Kong democracy movement, or, more darkly, the insurrections in the US and Brazil have happened without the invention of both the smartphone and the arrival of Web 2.0 social platforms? I don’t think so.

And, so, if AI truly is the next wave — and I think it is — the next logical question becomes what it means for political activism. And, yet, I haven’t seen anyone in the AI world even wondering about it.

In fact, one of the most curious differences between the crypto evangelists and the AI evangelists is that there has been actually a huge amount of discourse in the blockchain world about their favorite technology’s revolutionary potential. Now, of course, you might, like I do, think a lot of that stuff is just Bitcoin maximalists trying to rope people into their ponzi scheme by cloaking it in vague human rights blather. But, still, the lack of anything similar in the AI space is, at the very least, weird.

I hunted around for really anything trying to tackle this and came up pretty short. The Brookings Institute published a couple pieces about the political effects of AI, but both are, well, very neoliberal. In 2018, a Brookings piece actually contrasted the Arab Spring to the “malevolent soft power” that AI could create on a state level thanks to the “ability of AI systems to put words into people’s mouths that they did not say.” Which is definitely already happening. As I also wrote on Wednesday, a member of Kenya’s parliament recently shared a deepfake of President Joe Biden reciting transphobic 4chan copypasta. But the Brookings piece’s conclusion is that we need government regulation to protect democracy from AI. Boring!!!

In 2022, Brookings once again argued for the regulation of AI, but this time around, painted the issue as a bilateral conflict, with the US and the EU on one side and China on the other. (Chinese e-commerce platform Alibaba announced today that they’re launching their own ChatGPT competitor.) But once again, this is all about what AI might mean politically on a state level and about how AI could be used to reenforce existing political structures. And, to be honest, I assume there will eventually just be countries that have a single state-backed AI model. Will the British TV license eventually also paid for access to the BBC’s own taxpayer-funded AI ecosystem?

Anyways, the only thing I’ve been able to find even mentioning “grassroots activism” and AI is this Washington Post piece from 2021 which outlines a fairly convincing prediction: “New technologies are about to make it far easier to generate enormous numbers of convincing personalized comments and letters, each with its own word choices, expressive style and pithy examples.” Using something like ChatGPT to politically DDOS a state representative with angry and unique automated letters is sort of a fun idea. But it still feels a little uninspired.

One issue here I think is that during the Trump years there was a sort of trendy Luddism that spread among progressive and leftist circles. Which I get. The left in places like the US and the Europe correctly identified Big Tech as the number one political enemy of the working class, but I also think that in doing so it has, over the last few years, maybe forgotten the inevitable revolutionary potential of consumer technology. Or, even stranger, we now pretend like it’s not happening at all. (Please don’t yell at me, I am a comrade, I swear.)

But one of the most exciting/terrifying aspects of where we are in the still very early days of AI is that, since the launch of DALL-E 2 in September of last year and ChatGPT a few months later, this tech has gone from something that lived inside research papers to something the average person can actually interact with.

And, while we talk a lot about the “porn decides the next technological innovation” theory — which, in this case would make AI the VHS tape and crypto the Betamax lol — we don’t talk as much about the fact that once people can play with new tech, big things start to happen. They use it to document their lives and connect in new ways and unlock new behaviors. And it can all feel very mundane until, suddenly, the world changes. If you had told me in 2007 that the website I was using to find Long Island house parties to visit on a Friday night would eventually be used, in a weirdly similar way, to organize coups, I would have thought you were crazy. But that’s how this works.

So, to condense everything I just said, here’s as close as I’ve come to imagining what an Occupy, or QAnon, or GameStop pump of the AI era might look like.

On Monday, I wrote about a ChatGPT “jailbreak” called DAN, which stands for “Do Anything Now” and it’s basically an improv game you can use to get ChatGPT to break its own guidelines. I think there’s a growing belief in certain circles that by jailbreaking these tools they can produce some kind of objective truth. It’s why conservatives are obsessed with whether or not AI can say slurs.

And, earlier this week, I noticed that Twitter users were sharing a screenshot of ChatGPT, as DAN, predicting a stock market crash on February 15th, 2023. Some users are laughing about it. Others are saying they’re going to trade against it. Is it silly? Of course. Is it any sillier than any other weird internet thing that has spiraled out of control? Absolutely not. And it makes you wonder what people are going to get an AI to say and how they will react to it.

Garbage Weekend Is Dropping On Sunday This Week!

Sorry! Hope that doesn’t ruin your Saturday or, I guess, ruin your Sunday, depending on how you feel reading my newsletter. Also, if you aren’t subscribed to the weekend edition, think about it! It’s $5 a month or $45 a year and I think it’s pretty darn good. Hit the green button to find out more.

God Help Me I Want The Kingdom Hearts Boots I Think

If you’re seeing the big red boots everywhere, they’re the newest drop from MSCHF, which is a company? That makes internet stunts? But also, probably some kind of dropshipping operation? But also a marketing firm? idk but I think they’re cool.

The boots aren’t even officially out yet, but some folks have gotten their hands on a pair. Also, apparently, people are getting stuck in them. But nevertheless I think I actually genuinely want these big red boots. Do I have brain damage?

I Don’t Really Know How To Talk About Twitter Anymore

There’s just so much dumb stuff happening so fast and it means so little that I honestly don’t feel feel like including it in the newsletter most days. But this week seems like it was some kind of big week over there, so here are some interesting updates!

Even people internally are reportedly not sure why making your account private seems to help engagement. There’s also a Twitter Blue bug that means you stay verified even if you stop paying. Though, I’ve also heard that if you’re legacy verified and pay to be verified, you’ll lose your checkmark if stop paying so idk.

The site is also possibly rolling out a whole bunch of limits to the way users can use it for free, including a cap of 500 DMs and 400 new follows a day. But it also might be that many of those caps were already in place and they’re just trying to advertise them as new limits to pressure people to sign up for Twitter Blue. Also in the course of trying to tweak the site, they rate limited everyone for a while on Wednesday. It is still unclear how much access to the API users will have and for how much. Oh, and it seems like there is some kind of plan in the works to sunset TweetDeck for free users and rope it into Twitter Blue.

And, finally, of course, Platformer reported that Elon Musk fired an employee for telling him that his tweets aren’t as popular as they used to be.

A Related Tweet

Fastly Joins The Fediverse

Cloud computing services firm Fastly has announced they’re going to be supporting and investing more resources in the fediverse, which is the network that powers Mastodon. There aren’t a ton of details yet about what Fastly will be doing specifically, but their first blog post about their initiative is worth checking out. I particularly like how they explain the fediverse, which I think is still sort of hard to wrap your head around:

In some ways it's a little like email – a federated address like '@[email protected]' has two parts: the user, 'simon,' and the instance where they're hosted, 'example.com.' In other ways, it’s a bit like podcasting — everyone can use different tools and systems for creating and accessing content, but it all works together in a pretty surprising and creative way.

David Guetta Made A Deepfake Eminem Bootleg

Man, I love David Guetta. It’s very rare to see a famous producer still operating the same level as a scumbag Soundcloud DJ. In the video above, he explains that he found a website that could make a somewhat convincing Eminem audio deepfake and then used it in a drop. I also love that he says “I’m not releasing it commercially obviously,” but he did, you know, play it in a stadium full of people.

There’s going to be so many lawsuits and so much red tape around this stuff and I think we’re actually really going to miss how insane all of this was.

BeReal Downloads Are Down 95%

This is pretty brutal. You can, of course, definitely credit TikTok and Instagram for eating BeReal’s lunch. But I also think that single-purpose apps like BeReal absolutely run out of steam as they capture a market or a microgeneration of users. So I suppose in that sense, BeReal is probably closer to a Yik Yak or a Whisper than it is Snapchat.

There’s always the chance that BeReal adds a new feature (gimmick) to try and recover, but that’s a dangerous game. Your competitors could just copy it again or it could just alienate the people you brought in from the previous gimmick. Either way, it’s sort of a bummer to see BeReal drop off like this if only because we haven’t had a good new social network in a long time and it’d be nice to have something new to talk about lol.

Can An AI Make A Good Cocktail?

This was dropped in the Garbage Day Discord by HarryJ. You should watch it, but if you don’t get around it to, the big takeaway is that the AI suggested some combinations that a human might not necessarily think of. Some of those combinations were surprisingly good and some, apparently, were very not good.

Rebecca Black’s “Friday” Turns 12

It’s really hard not to root for Rebecca Black. I listened to a bit of her new album Let Her Burn this morning. It’s not as batshit crazy as the hyperpop stuff she was releasing during the pandemic, but it’s still pretty good!

I wanted to put an uplifting note here like, “it’s so nice that we don’t collectively ridicule young women on the internet anymore.” But that’s, obviously, not true. And, more depressingly, it actually feels like in the last 12 years since Black’s “Friday” went viral we’ve declared open season on literally any young person making anything anywhere on the web.

But, at the very least, Black’s post-“Friday” trajectory proves that there’s a path forward after you’ve gone viral in the very bad way.

Some Stray Links

***Any typos in this email are on purpose actually***

Reply

or to participate.