Incidently, I’ve been playing around with Gemini Veo’s video models recently, just to satisfy my own curiosity. The fact that a video is generated at all is impressive. But you’re not likely to get a good video, or even a mediocre video, out of the model by itself. Or at least, I haven’t yet. The best one the model produced still needed some work in DaVinci Resolve to fix the corny audio.

I’ve not posted the videos anywhere, but I’m thinking of the montages I see on YouTube featuring these videos that are obviously generated by AI. And the signs are not hard to spot: bad audio, objects that don’t look like anything, someone speaking into a handset of a phone that already had a handset. These are all things I saw with the videos I asked to be produced. Posting those as they are, without any form of basic editing or even review, with my name attached to it would be embarrasing. Those that do are, in my opinion, engaging in the production in slop. The bar is that frickin’ low.