I haven’t written in so long that I’ve accumulated enough thoughts for a volume of stories. The last publication about bananas was only two months ago, yet it feels like they’ve been with us for years. I won’t even mention SD—what kind of analog tool is that?)
It’s hard to keep up with updates in AI if you want to do something else. New releases appear several times a day, and this has been the case for the last 1.5−2 years. There’s a feeling that Pandora’s box of AI surprises still has plenty inside, and the pace is only accelerating. The frightening part is that it no longer surprises us.
Startups are maturing alongside their pipelines, allowing them to replace entire production studios in one go. Without experience, it’s unlikely you can pull off such a trick, but those with enough experience can work wonders. I note that we have received a plethora of buttons for decomposed "beauty," sometimes not very smart, but it works! Look at the ballerina from the render. Can this be done without AI? The answer: easily, but it’s expensive and inefficient: green screen, actors, filming, post-production—in the realities of the developer’s conveyor, this solution is impractical. But with a banana—it takes a few hours. If you have a clear vision of the result, it all comes down to the quality of prompts and typing speed.
Another issue is the scatter of tools across platforms. You always have choices from APIs to aggregators like Higg’s. Ruthless competition forces developers to implement exclusive tools: which is good, but each such platform costs at least $ 20−40 per month. And there are no universal solutions.
Next, we will talk about automation—a thing that has been stealing my sleep for the third week now.