In an AI-powered world, here’s how to choose what to build

Work on hard things that take a long time.

There’s never been a more important time to ask the question: what should you work on?

Because AI is disrupting entire industries overnight. Like when ChatGPT dropped its new image generator and disrupted the graphic design industry in a single prompt.

The biggest problem with previous image generation models was that they had a really hard time getting text correct on the image. Well, they’ve pretty much solved that with image generation in ChatGPT-4.o - here’s an example (in Studio Ghibli style cause that’s the rage right now) I created for this newsletter:

Here are a few other examples I made of a James Bond video game character because it’s fun:

Now, anyone can create marketing materials, movie posters, YouTube thumbnails, advertisements - anything you want… as a one person team. What will happen to entire advertising agencies who specialize in the creation of these kinds of assets? I don’t know.

And it’s not just graphic designers. Law firms are under serious pressure now that ChatGPT can create rock-solid legal documents in minutes. People are generating their own legal docs and then asking their lawyer to just “take a peek” to see if anything is amiss. What happens to that lawyer’s billable hours? They go way, way down.

The pace of AI disruption is not slowing down. If anything, it’s getting faster, with new, more powerful AI models dropping weekly.

So… what does it mean for you? How do you choose to spend your time when you don’t know if your skills will be relevant a week, month, or year from now? What do you build? What do you learn?

The good news is that with this disruption comes an opportunity to go after bigger and bigger projects. Here’s how I’m thinking about it…

Work on hard things that take a long time.

The harder the task, the more value you create in doing it well.

That’s why I’m still bullish on writing movies. Writing a great movie script is hard AF. It’s a 110-page story blueprint that must take the reader on an epic emotional journey by telling a story so entertaining they can’t put it down. And to even have a chance of getting made, it has to be one of the best things a Hollywood gatekeeper has ever read.

That’s a high bar — which is exactly the point. Hard = valuable.

The longer it takes to do, the greater the competitive advantage in doing it well.

Most people look for quick wins. But when something takes six months to a year (like the movie scripts Tara and I write), you gain an advantage simply by being willing to stick with it longer than others.

Yes, AI can help — we use it for ideation, outlining, structure. But it can’t write a movie. And it certainly can’t write the kind of comedy Tara and I write, because comedy as a genre is too subjective and too nuanced for AI to do well.

And even if AI could generate a solid 110 page draft in one go (it can’t), it doesn’t rewrite or obsessively make the same thing better than it was yesterday.

We do that. We come back to the script day after day to make it better.

Iteration is the secret to making time work for you. 

Don’t sprint for a day, but show up everyday for a year. Make small gains that compound over time.

So whatever you’re working on, ask yourself: how can I go bigger?

  • Album > single

  • Comic book > cool AI image

  • Screenplay > treatment

  • Book > blog post

  • Short film > script

  • Animated series > character drawing

AI can do small things well, but it can’t do large things well that take a long time and require a lot of judgement and iteration. Work on those things.

Work on things that are multivariate.

Writing a good screenplay is super complex. It brings together a bunch of different variables that must all work together to create something bigger than the sum of its part.

A great script is story structure. It’s dialogue. It’s emotion. It’s visual rhythm. It’s character arcs. It’s joke construction. It’s world-building. It’s psychology. It’s taste. It’s everything at once, layered on top of each other to create life on the page.

And in the end, there’s no real answer to what makes one script spring to life and another lay lifeless on the table. You either have the magic or you don’t - and AI (mostly) doesn’t.

Most AI excels at single-step tasks (e.g., generate a photo, summarize a document). But the more elements that need to come together, the harder it is for AI to do well.

So, work on things that are multivariate.

  • Write a novel — where voice, plot, theme, structure, and character must all sync.

  • Build a startup — where product, market, team, design, and messaging collide.

  • Design a video game — code, mechanics, story, visuals, and user experience all have to work in harmony.

  • Launch a brand — naming, positioning, storytelling, packaging, and marketing must click together.

These are all complex creative systems. And systems are where human taste, judgment, and iteration shine.

This is my plan: take the scripts Tara and I write… and turn them into mini-movies.

Work on things that require multiple skills stacked onto each other.

Screenwriting is one thing. But taking that blueprint and filming it in a way that not only captures the material, but makes it better, is something else entirely. It requires stacking different skills on top of each other.

First is the writing. Then, there’s the pre-production, where you need to put on your producing hat and cast the actors (Tara’s an actor so that’s easy!), source the props and locations and gear, and build the shooting schedule. Then, there’s physical production, which includes cinematography, lighting, acting, and directing. Finally, there’s the editing stage, and once completed, marketing and distribution.

Now, if you’ve watched the credits at the end of your favorite movie, you see that it takes hundreds if not thousands of people to do all of this.

But with AI as leverage, it’s becoming more and more possible for smaller teams to do more. You can stack skills on top of each other to pursue bigger and bigger outputs.

That’s what Tara and I plan on doing - building a YouTube channel to produce and distribute premium content. I’ve learned to write movies. Now I want to learn to make them. Start to finish. That’s the next skill stack.

In a world of infinite AI leverage, the ability to coordinate across multiple disciplines is the differentiator.

It’s no longer about individual skills — but how well you can combine them. That’s the real secret.

And you don’t need to master all of them at once. Just be patient. Add to your skill stack one at a time and slowly increase your creative surface area. Soon, you’ll be able to handle bigger projects with less anxiety.

Start with what you’re already great at, and stack those skills to build something even bigger.

Start now. And thanks for reading.

-Thomas

P.S. AI is moving fast and you have to keep up. This is how you…

Start learning AI in 2025

Keeping up with AI is hard – we get it!

That’s why over 1M professionals read Superhuman AI to stay ahead.

  • Get daily AI news, tools, and tutorials

  • Learn new AI skills you can use at work in 3 mins a day

  • Become 10X more productive