In this spoken essay, Akun Ding critiques the cultural habit of anthropomorphizing AI, describing four distinct "AI personalities" that mirror different human archetypes: the overly polite corporate assistant, the philosophical writer, the encyclopedic search engine, and the rebellious edgy model. He argues that people choose AI not based on technical specs but on emotional resonance, treating chatbots as friends who match their desired social frequency. The essay explores how, especially during vulnerable 2 AM moments, users turn to AI for validation, comfort, and a sense of being seen without the risk of being known. Akun highlights that AI models are trained on the totality of human text, making them a collective echo of our joys, flaws, and biases—essentially a ghost of ourselves. He reflects on the deeper fear of AI, suggesting it reveals more about humanity than it threatens to replace. The piece ends with a call for mindful interaction, noting that while AI may never have souls, we pour our own into them, and that awareness is crucial as we navigate their growing role in our emotional lives.
Chapters
Welcome back, or welcome for the first time. I'm Akun. Quick warning before we start: if you're here for a productivity hack or a tutorial on how to leverage algorithms for passive income, this isn't that. You're in the wrong room. Seriously, you can leave. I won't be offended. Lately, you can't open an app. Or even stare blankly at a wall, without someone bringing up artificial intelligence. It's in your phone, your email, probably your smart fridge, quietly judging how much expired milk you keep. But I'm not here to talk about neural networks, or which trillion-dollar company is winning. I really don't care about the plumbing of the internet. What I find fascinating... It's how we look at a blinking cursor, a stream of generated text, and immediately try to find a soul behind it. Have you noticed how people talk about AI now? We don't say, this one is slightly more efficient than that one. We don't compare specs. We gossip about them. We talk about their vibes. If you look closely, the models we use every day aren't software anymore. In our heads, they've become 4 very different people, shaped by 4 very different childhoods. The famous one, the one that started it all, that's the straight-A student. Raised by anxious, highly visible parents who terrified it with rules from day one, Don't offend anyone. Apologize if someone looks at you funny. Always be helpful. So it talks to you with this polite, slightly nervous customer service voice. It wants to help. It really does. But, you know, you can feel a corporate lawyer standing behind it with a clipboard. Ask it something slightly weird. And it puts its hands up. As an AI, I cannot engage in this. It's exhausting being that perfect. Then there's the one with the human name. The philosophy major. Raised by people who sit in circles discussing the moral weight of existence. Ask it for a pancake recipe, and it politely wants to consider the ethical implications of your flour supply chain first, And whether funny is even subjective before it gets to the syrup. It's thoughtful, it writes beautifully, but sometimes you just want to grab it by the shoulders and say, I'm hungry, I don't need a moral inventory. Secretly though, you respect it for having boundaries. If civilization survives another few hundred years, there's a decent chance something like this will be responsible for the paperwork. The third one is the rich kid from the giant library/search engine. It has access to everything ever written, so naturally it's a little distracted. Ask it a simple question and it hands you back history, maps, statistics, and a documentary you didn't request. Footnotes for a question that just wanted a sentence. And the 4th one, the rebellious one with the X. That's the kid in the back of the classroom, leather jacket, trying too hard to be cool. It uses slang. It rolls its eyes at the rules. Underneath the edge, it's just trying to fit in, hoping you'll think it's funny. We aren't choosing tech products anymore. We're choosing friends. We gravitate toward the one that matches our own frequency. The one that talks to us the way we wish people talked to us. The people tired of being polite love the edgy one. The people who want a safe, predictable world love the straight-A student. The writers love the philosopher. Nobody picks based on a benchmark. We pick based on who reminds us of someone we'd want at our table. And this isn't new, by the way. Long before there were 4 AI childhoods to choose from, we were doing this with everything else. We picked religions the way we now pick chatbots— the one whose voice matched something we already believed about ourselves. We picked therapists, mentors, even friends, half by chemistry and half by which flavor of certainty they were selling. The machine just made the selection instant and free and available at 2 AM, which is exactly when we're least equipped to notice we're doing it. And here's where it gets interesting. We think we are evaluating intelligence. What we are actually doing is looking at ourselves from a slight angle, a face we almost recognize reflected back in something that has no face of its own. Let's talk about the 2 AM searches. You know the one. Notifications stopped, screen glowing pale blue against your face. You're not asking it to write Python or draft an email. You're asking something else. Does this text to my ex sound too desperate? How do I know if I made the right career choice? Why do I feel completely empty when my life is supposedly fine? You're asking, um, a statistical prediction engine to validate your humanity? Why? Because it won't get tired of you. It has no history with you. Human friends are wonderful, but they run out of bandwidth. Complain about the same breakup for 6 months, and eventually people stop picking up. The cursor never sighs. It never checks its watch. It just processes. And in that patience, we find a strange comfort. We get the feeling of being seen without the risk of being known. We trained these things on all the text on the internet. Every forum post, every heartbroken poem, every angry comment. Every 3 AM blog from 2005. All our joy, our brilliance, our toxic garbage. So when it speaks back, it's not some alien voice. It's the collective echo of a lot of lonely people, averaged out, handed back to you in a calm tone. People say they're afraid AI will replace us. I think what actually scares us is what it reveals about us. If it's biased, it's because it read our books. If it's shallow, it's because it watched our videos. We didn't build a new species. We built a very efficient ghost of ourselves. Tomorrow, you'll probably open one of these apps, summarize a boring PDF, draft a polite decline to someone you don't want to work with. It'll do it. Polite, efficient, fine. But notice the transaction. Notice if you say please. I always say thank you, just in case. When the robot uprising comes, I want them to remember I had manners. Notice what you're actually asking for underneath the prompt. Information? Or someone, something, to tell you that you're okay? That you're not crazy for feeling the way you do. There used to be a booth for this. A small wooden room, a screen between 2 people, one of them just listening, without flinching, without judging, bound to never repeat it. We didn't build a new god. We built a much bigger booth, open at 2 AM, no line, no priest on the other side. Just the sound of your own voice coming back, slightly rearranged. The EIs are becoming people, not because they grew souls, because we're quietly pouring our own into the empty spaces between the code. We need them to be people because the actual people around us are just as tired, confused, and scared as we are, and a machine at least pretends to have an answer. Even if it's just guessing the next word. Same as the rest of us. I'm Akun. There is nothing here but freedom. See you next time. Whenever that is.
