Notes on AI Option Paralysis

A pep talk to my overloaded brain

So we're at a point in AI where option paralysis is much more the enemy than any obstacle you could actually come up with in the process itself.

I hit maximum brain overload with AI movie making processes tonight (which, honestly, has been building for weeks), and I’m unpacking that because I think this is going to be a key for those who get ahead with AI movies.

The core problem is: there are so many good video generators right now. SO many. A bunch of them can do amazing things. They all have slightly different angles and features. They all have different price points that make them more viable or less viable, which you have to weigh against quality and availability. Like—how easy is it to automate with X model? How costly is it to use at scale, and how much can you depend on it for a general workhorse vs a specialized use case?

Four weeks ago, I had an entirely different process plan for making movies than I have today. Today, I cancelled one of the main workhorse generators of that plan, because it’s no longer the most viable, and that viability shifted in the last few days.

Which is crazy.

I asked Gemini to personify what it feels like to be wrangling AI video tech right now. Gemini delivered, ha.

Last week, I spent around 9 feverish hours and ~$200 in credits using the Lovable AI app-building app to build out the bones of a chapter to videos app for my future movie studio.

My shiny new app I made is AMAZING. I love it. And also, things have happened in the last few days that make it less viable cost-wise as a solution than a different plan, or at least something I need to adapt around again.

(Screenshot because I have to show off my shiny thing I made, huzzah!)

All of which adds up to: in an industry landscape that is literally moving at light speed, how do you lock down a process enough to make all of your hard work on a project matter?

A few things:

1. The tech is moving fast, but video tech that’s excellent now will still likely hold up for a while. Most video flagship models now have a base quality level that will hold up, and even if it doesn’t for whatever reason, there will be more ways shortly to “re-skin” a movie with video to video generation. (Luma can do this now, though the quality is still getting there.)

2. You can’t keep up. It is physically impossible. I am tying myself into pretzels trying to, and, that anxious feeling of my gears spinning while the world passes me by is not only not true, but not a stable state for grounded creativity. The reality is, absolutely everyone using this tech right now feels the same way. We’re all excited and frantic and so full of FOMO we’re leaking it out our eyeballs. (More or less.)

3. The one thing about this field that won’t change soon is just how quickly it’s moving. So it makes no sense at all to invest a ton of time in a single workflow. Now, that said—the nine hours I spent making my app are absolutely productive hours. I’m going to put that app through its paces, learn more about the process, get better prompts, and be able to do that in a concentrated manner that I might not be able to synthesize as easily if I tried to do this across a bunch of different generators manually. I’ll revise my app as I go. It’s already helped the way I’m thinking about things a ton.

tl;dr: Shit’s moving crazy fast, you can’t catch up, grab a surfboard and have fun.

We are living in a state of technology that we’re not wired for. And part of riding that wave is rewiring so that you can surf on it without totally breaking yourself and your own processes.

For me, that’s letting go of my need for perfection, for “the right way” (a myth of myths), for even having the most optimal cost-effective plan, or the most optimal time-effective plan.

If you can get within 50-60% of your “right way,” you’re golden. Right now, you’re probably still golden if you can manage 30-40%. The field is so wide open it’s not even funny.

It’s so new that your process will change as you go, even if you do lock down a process for a single project (which is a good idea, I say without having done that yet).

Your absolute cutting edge innovation is not worth sacrificing your story or your sanity. Or putting off your vision.

I spend a lot of time on LinkedIn these days because that’s where so much of the news around AI is happening, and there are people showing off innovations in every single direction you could think of. It is SO good. And SO bad for my brain when I’m already scrambling with FOMO.

And I have to dial back and think sometimes—everyone posting at the cutting edge of the tools has only a small slice of the picture. Almost everything happening right now with videos are “hell yeah!” videos showing off the tools. And that is great! That’s amazing. Some of them are short films, some of them are lead-ins to longer projects. Most are cool feels and vibes things.

But, beyond a very very few studios or creators who’ve posted longer form projects, most of the flashy bits are just that.

As authors or screenwriters or storytellers or creators, whatever you’re calling yourself in your head, we already think in longer strides. We already know the bigger picture, and we have to be thinking in those terms in the long term, not ultimately chasing after the quick and shiny.

If you build to trends, trends pass. If you build IPs, IPs last.

All of which is my pep talk to myself to pipe down, you don’t have to shift all your plans the moment a new shiny thing comes out, if you’re not hopping on the latest thing…it doesn’t matter. It’s ALL good.

Keep shiny, but keep sane out there, y’all. I want to watch your movies.

-Novae