I was thinking about making content and asked myself whether I should just do an AI avatar. What stopped me was noticing two different things happening around AI right now.
One is people who've built a business out of "this is going to die," fearmongering about the next model purely for hooks and views. It works for them because that's literally what they exist to talk about, the way a news channel exists to cover the next headline. That's not what I'm doing. I want to show up consistently as the actual thing, not as commentary on the thing.
Someone like MKBHD, the big creators who don't lean on AI generation, mostly don't need to. They already have the capability and the resources. AI sits in between: genuinely great when you already have direction and just need to get there faster. An actual designer with real skill, instead of spending ten or twenty hours on one thing, can get there in an hour, sometimes ten minutes. That's the win. It's not the win for someone who doesn't have the eye for it in the first place. For people starting out with a small budget it still gets a ton of work done, same as it does for people with big budgets, just at a different scale.
The organic food comparison
This is the clearest way I've found to say it. Normal food isn't unhealthy just because it isn't organic. Organic sells at a premium because people know it took more time and care to get there, and that association compounds. Content works the same way. AI-written or AI-spoken content isn't automatically bad, but the reason showing up as genuinely me, in my own words, matters is the same reason organic sells: it signals that an actual opinion, an actual fight, went into it.
The moment you use AI by default, without giving it your take, without fighting back on what it hands you, everything starts sounding exactly the same. I noticed this watching ten different short-form videos on ten different topics from ten different people. The wording, the rhythm, all of it felt identical, because they'd clearly said "here's my topic, give me a script" and read it straight out. I get why people do it. It's fast. But it flattens everything.
It becomes slop the second it's asked to supply the opinion itself instead of just executing one.
That's the actual reason I keep a personal knowledge vault, not a side effect of it. Everything I put into it, everything I speak, however roughly, is going in so that eventually an AI can learn how I speak and how I process things, not just what I've concluded. I might not be the best writer or speaker yet, but I want to get there, and I'm using AI to get a better version of my own thinking out, not to replace the thinking. The vault is my taste and how I handle things, handed to a system so that when it produces something, the wording and structure might be cleaner than mine, but the thinking underneath is still mine, not generic.
Where the compression actually helps
The reason big-budget work is good is that every artist involved has their own taste and standard. Give those same people the tools, and a smaller team can produce the same quality in a fraction of the time. Ten days of work on one minute of footage, compressed way down, without losing the craft, because the craft was never the bottleneck. The hours were.
The win is speed, not substitution. Someone without the eye doesn't get the same result faster.
Where the ethics actually shift
Budget matters. If a story is human-written and AI is used to get the best visuals on a small budget, an indie film, a short, people don't mind, because the constraint is real and honest. They're using AI to reach a stage they couldn't otherwise afford. If a company is expecting a billion-dollar outcome and puts in ten dollars of actual craft, that doesn't work, the same way you can't get more energy out of a system than you put in, unless you're already famous enough that people forgive it, which is really its own separate thing.
For someone building a story to get to the next stage, AI is a legitimate way to close a resource gap. For a studio that can already afford real 3D modelers, real designers, people with genuine creative vision, using AI to skip having those people isn't the same move. That's the version of AI in creative work I actually believe in: it compresses execution time for people who already have vision. It doesn't replace needing the vision in the first place.
I tried using AI to edit a video directly, once. It wasn't good. It didn't tell the story or carry the flow the way it needed to. Execution without a fought-for opinion behind it comes out generic, even in editing, not just in scripts or avatars.