A 15-second 3D rotation of a real product. AI doesn't know what the other sides look like. Unless you spend more time and budget on training a LoRA.

An alternative is making shorter shots with different close-ups of the product and animating them.

The same applies to human motion. It works perfectly for shorter shots, but longer ones tend to “break” parts like hands, feet, and faces. The wider the shot and the smaller the details, the more distortion you’ll see over time.

It depends on the intention. Even if you're clicking around, if you're trying to understand what each action does, you're learning.

Experimenting is absolutely necessary, since within any tool you can always find your own workflows to reach your goals.

I don’t think prompt engineering really exists anymore. Most tools already offer useful features that solve the “mystery” of prompt engineering, like adding image references, defining composition, or using custom GPTs.

Besides, most so-called “prompt engineering” tricks don’t do much at all. And the few that help only slightly improve results compared to simply applying your own knowledge of the subject.

Stop testing every tool. Stop the FOMO. Only try new tools when you see them do something you couldn’t do before.

It saves you a lot of time – you’re not learning a new tool “just because,” and you can focus on mastering the ones you already use.

It might come naturally to some, but not everyone. But everyone can change. A great starting point for me was reading “Psycho-Cybernetics” by Maxwell Maltz.

I think in order to sell something, you need to know what you’re selling.

In my case, I know the cost of things done without AI. That’s why I can price my work at 40, 50, 60, or even 80% less than that — and still charge 2x more than my competitors.

If something seems too good to be true (too cheap in this case) to be true, it usually is. If someone doesn’t know how to price, they probably don’t understand the business either: the client's requirements, expectations, and pain points.

Humans have needs, AI doesn’t.

If we understand what creativity is, then yes, AI can be creative like a human. But the key difference is that AI has no need to communicate, express, or create anything unless we ask it to.