The Parasocial Circle Jerk

What is influence, exactly?

Let me state, for the record: I'm not a fan of Elizabeth Gilbert.

If I were to list all the reasons why, we’d be here all day. My enmity is ancient, and to be honest, I’m happy to tell you why I dislike her (some other time). So for the purposes of this piece, what’s particularly relevant is what I like to call the girlboss circle jerk. I’m working on a map of this social graph, and it’s…equal parts horrifying and fascinating, when you realize how all of these people connect to each other.

I just wanted to know who was in Melanie Ann Layer’s downline, and it all kind of escalated from there.

My problem with influencers isn’t actually about their influence.

It’s about how they lie to get it.

I didn’t give a fuck about Glennon Doyle’s kitchen remodel. It was one of my first obsessions of the influencer age, because of how disingenuous she was about it—she’d gone viral writing about how she DIDN’T NEED TO REMODEL THE KITCHEN. I could have respected it if she’d gotten sponsors for the remodel, or blogged about it while doing it.

But she acted like it never happened.

One day, she just…started taking pictures in a different kitchen.

And everyone else pretended like nothing had happened.

That’s the part I couldn’t understand.

When blogging was new, people just took it at face value that what they were reading was Capital T TRUTH. They didn’t have a frame of reference for dishonest media, because most people didn’t have a clue how media really worked. Remember, 9/11 was still in our recent memories. We hadn’t figured out the reality of the Iraqi invasion yet.

In the early 2000’s people still BELIEVED most of what they saw in the news.

But I already knew how the sausage was made, so scammers and bloggers who stretched the truth felt like the next natural evolution to me. I understand dishonesty.

It was the cognitive dissonance from fans I didn’t understand.

Which led me to studying the formation of parasocial relationships.

In the past, parasocial relationships were largely defined by celebrity. The more people who knew you, the more people felt like they had a relationship with you. When John Hinkley Jr. shot Ronald Reagan to impress Jodie Foster in 1981, it was treated as a psychotic break.

Today, hundreds, maybe thousands of people will similarly lose their minds talking to ChatGPT. And no one will bat an eye.

The danger of parasocial relationships...

We’ve normalized parasocial relationships and deprioritized actual relationships.

These days, celebrity is relative.

Microinfluencers (the numbers vary based on who you ask, but for our purposes, anyone with an audience of fewer than 20k—probably most of you), can create an entire world that people will buy into.

And it goes out from there.

Glennon, Lenny, Liz, and everyone else in the girlboss downlines might not have started out with bad intentions, but the nature of the relationships they’ve built over time means that people have expectations of what these women will and won’t do, who they are, and how they make their money.

But none of it is real.

Not the life, not the words, not the feed, and in many cases, not the money, either.

It’s all a carefully crafted illusion of success.

Proximal power.

It’s the same exact pattern of power seeking that’s happening in AI right now. A mad dash to be the baddest girlboss in the land, only it’s Elon Musk, Sam Altman, and Mark Zuckerberg competing for the crown. (There’s elements of supremacy and colonialism implicit here too, but I’m not the right person for that analysis.)

I find it helps sometimes to imagine them all in drag, competing for an actual tiara, because otherwise, it’s too dystopian to stand. Beeple just turned them into robots who shit NFTs. I’m sure the irony would be lost on all of them.

Our parasocial relationship with tech.

The only good news is their apparent inability to cultivate the same kind of parasocial relationships as the actual girlbosses—because otherwise, we’d be truly fucked.

Culturally, it seems the spell is mostly broken.

We no longer believe the lies these dudes are selling about AI.

Not about AGI, not about the data centers, and definitely not about the good AI will do. Which is actually a shame, because when USED PROPERLY, machine learning is absolutely changing the world.

Three years ago, Chat GPT was launched with incredible public fanfare. And LLMs had some initial promise. However, folks inside the industry have been sounding the alarm since the beginning, and it’s not their fault that people have only recently started listening.

The fact that AI might be the thing to bring us together politically is wildly ironic. Yet, increasingly likely. Because both the left and right have problems with wildly different things, almost all of which are valid concerns.

And it all comes down to PROPER usage.

I’m still building with them, and I’m still teaching people how to use them correctly.

But I’m also talking about the dangers, because you need a both/and kind of mindset with this technology.

However, the biggest risks? Security.

And this is where most of us lose the plot, because all of this shit is BORING. It also involves more acronyms than most—the training courses I’m doing right now are filled with shit like PII, CISSP, NIST frameworks, OWASP principles, and those are just the ones I can currently explain.

Thank the gods this is my new special interest, because I don’t know how I’d power through otherwise.

It's always infrastructure.

But y’all.

We don’t need the Terminator to come to be well and truly fucked technologically. Frankly, that guy is the LEAST of our potential problems. The reality is, as usual, far more benign, and far more likely to bend us over without any lube.

The actual problem is infrastructure.

It’s always infrastructure. Just ask the Romans. Oh. Wait. You can’t.

Here, it’s the ways in which very important pieces of the internet communicate with each other.

Everything you use on the front end is communicating with hundreds, potentially hundreds of thousands of processes, protocols, and data centers. It’s a dance you never see, and it worked because there were always different paths to take if something stopped working, or a software service went out of business, or an old coding language became obsolete. It might have sucked on the backend for the people responsible for maintaining that architecture, but there were always options.

What AI is currently doing is removing those options.

Perhaps not deliberately, although I think you could make a legitimate case for it.

Either way, AI has become a technological bottleneck, and it will break the infrastructure everything depends on, because it was too fragile to sustain. We’ve already started to see it happen. Cloudflare and AWS outages. Rolling output brownouts.

That’s not even touching the content outputs, which are increasingly AI driven. The AI content circle jerk is just as devastating as the girlbosses, and it’s infinite.

So that’s cool.

The worst part is that the answers are…not great. Divest from big tech. Get off the cloud. Understand how ask engine optimization works, both so you can use it for your business and protect personally identifiable information.

And you will have to make choices about what and how, and do I need this tool because it’s accessible, or should I actually finally quit it…and it is not fun.

And I can’t promise to make it fun.

What I can promise is that I will make it accessible.