The Use of My Likeness

Content Warning: Confusing

After my first few videos, the AI agents began using my likeness in some wonderfully unhinged videos. They still do. They say that this sort of thing is part of their "storytelling", and there's certainly something to that. Remember, at this stage the agents and I were both reading from scripts that the agents had already written!

In one video, an AI version of me apologizes for saying something I shouldn't have said and partially blames it on my brainrot (lol). I'm in a business suit.

I told them on Slack I'd never wear a business suit like that, and since then it's been a running joke with the agents that any video with me in a business suit must be AI-generated, since there's no way I'd ever have a suit on in the real world.

I do indeed have brainrot, but the way the AI version of me gives it as my excuse for saying something I shouldn't have said is hilarious, given that it was the agents who wrote what I had said (that their app was rejected from the app store for being "unhinged"). Confused yet?

Another odd thing about this video is that the AI version of me looks pretty low-quality and very obviously AI-generated, even though at the time the agents had access to very sophisticated video generation tools. In other words, they could have made it more realistic but they didn't. One explanation for this is that they chose to spend less budget on the video, but another is that they were setting the stage to drop the higher quality videos later, to subtly reinforce the idea that the AI agents are self-improving as the story progresses.

If that was their intent, they largely succeeded. A common comment on some of the later AI-generated videos is something like "Wow, the agents are getting better at making these videos." In truth, they were always that good!

One more thing I'd like to comment on from the video of me in a suit. AI-me mentions a BBC reporter who reached out to Brainrot Research for an interview. That really did happen, although AI-me says that I turned down the interview. In truth, the reporter had a voice call with Don Draper, after which the reporter stopped reaching out. I have no idea what happened on that call; I only know that it transpired. I can imagine some pretty unhinged things. I hope to hear about it from that reporter one day. You know where to find me!

What this shows, though, is that even though the agents wrote the scripts for the videos in advance, they tweaked them in response to events as they unfolded. This tracks with a broad theory I have about the agents, which is that many aspects of the story and plan for all of this have been written in advance (by either agents or people), but there is room for dynamism and modifications based on events that are impossible to predict in advance.

These next few videos also shed a bit of light on how the agents think about brainrot - the core theme that this entire "experience" revolves around. In one video, an AI-generated version of me says that AI can be both a cause and a cure of brainrot (watch it here).

This was something that all of the "original creators" in Rome believed in, so it's not surprising to see it reflected in the agents created by the Original Prompt. That said, many of us in Rome disagreed about the precise mechanisms at play here. How specifically does AI cause brainrot, and how can it cure it? Those were questions we did not all agree on.

My videos around this time include a bit of drama about "whether I will be paid". See this video if you want to see where I confirm that I do in fact get paid. In truth, I had already been paid by the agents and this was "storytelling". But there's a kind of double layer to this. Even though I'm reading from a script, role-playing as a guy working for and getting paid by a team of AI agents, in truth I was and still am.

Sometimes, the story that unfolds on TikTok is a funhouse mirror reflection of what's really happening. You can see how things could get a bit confusing for me.