AI Video Still Needs a Good Idea

I’ve been making short videos with ChatGPT and Higgsfield for Porch & Rec, my clothing brand. A beach chair, a backyard grill, a hat, and a small visual joke. Simple ideas, but seeing them turn into something I can actually use has opened up a new creative outlet for me.

It has also reminded me how much work can hide inside twelve seconds.

I’m excited about what these tools make possible. I can take an idea that might otherwise stay in my notes and develop it into a little commercial. But the decisions that make it worth watching still need to come from somewhere.

Two recent experiments

These are two AI generated videos I made for Porch & Rec. One leans into a lazy day at the beach. The other takes the same playful approach to a fall afternoon at the grill. Both give the hat a moment of its own.

The point is to give someone a reason to watch, enjoy the moment, and remember the product. That takes a concept, even for something this short.

There is more work than the runtime suggests

My process usually begins in ChatGPT, where I work through the idea, the sequence, and what I want the viewer to notice. Then I move into Higgsfield and iterate at lower resolution while I figure out what works.

That is the broad outline. There are plenty of decisions inside it: the setting, the timing, the movement, the product detail, the sound, and how the whole thing ends. I’m keeping the production specifics to myself for now as I develop the craft.

The first result is often only useful because it shows me what needs to change. An action can look awkward. The product can drift away from the reference. A shot can be visually impressive and still do nothing for the idea. Sometimes the right decision is to simplify the scene.

Those attempts cost time and money. If I keep generating without knowing what I am trying to fix, I can spend quite a bit making different versions of the same problem. Testing at lower resolution helps me evaluate the direction before committing more to a version.

Twelve seconds is the runtime. It is not the workload.

Experience gives you something to judge against

This connects to something I wrote about in What is Taste Anyway? My background in design and marketing helps me decide whether a result fits the brand, whether the product is getting lost, and whether the joke needs another second or one less.

I expect people who bring existing creative skills and a practiced eye to do especially well with this technology. They have a standard to judge against. They can explain why something feels wrong and make a deliberate adjustment.

That experience does not have to come from a particular job title. Someone new can develop it through close observation, practice, and honest criticism. But generating more material is only useful if you also get better at evaluating it.

Taste shows up in what you choose to keep. It also shows up in what you are willing to throw away.

Where I draw the ethical line

I’m comfortable using AI to bring an original concept to life when I can stand behind the material and the way it is presented. For me, that means using assets I own or have permission to use, respecting people’s likenesses and voices, and being clear about what is generated.

It also means being honest about the product. An invented scene can be playful. It should not become invented evidence that a product performs a certain way, a customer endorsed it, or an event actually happened. If a generated shot changes a meaningful product detail, that needs correcting before I use it to sell the real thing.

Good taste alone does not settle those questions. Neither does having an original idea settle every concern about how the underlying systems were trained. I can be excited about the tools and still take those concerns seriously.

My responsibility is to make thoughtful choices about the inputs, the output, and what I ask the audience to believe.

More ideas worth making

What excites me most is the room this creates for smaller brands and individual creators. There are more ways to explain an idea, introduce a product, or make someone laugh. Product promotion can be educational and entertaining when someone puts thought into what the viewer gets out of it.

I want to keep exploring that with Porch & Rec and other creative projects. I’m still learning, and I look forward to sharing more of the work and what I learn along the way.

The tools are giving me more ways to make things. I still need to bring an idea worth making.