Context
Let me be blunt: written AI prompts can’t possibly be the future of creation. Why? Because most people are just not good writers. I saw this firsthand during my days at Facebook, watching millions of people freeze up at the sight of a blank text box. If you want evidence, look no further than the rise of photo and video sharing—people love these mediums precisely because you don’t have to write. The blank text box is the scariest thing on the internet for most people.
So the idea that the next era of creativity will be built on typing out detailed prompts—whether for images, movies, or anything else—strikes me as fundamentally disconnected from reality. People hate writing, and the average person is not about to start typing out nuanced visions for their creative work.
Market Signal
Sure, English-language prompts are having a moment. Why? Three reasons:
Lowest Common Denominator Input: Even if you can’t make a movie, you can (maybe) describe a scene in basic English. So it demos well.
Easy to Copy-Paste: Everyone knows how to ctrl+c/ctrl+v, and written prompts are easy to tweak or remix.
Lots of Training Data: The internet is flooded with text (even if it’s mostly bad writing), which makes it easy for LLMs to learn from.
But if you’ve ever looked at the intricate, hyper-detailed prompts behind those viral AI images (lighting, focus, background, style, etc.) it’s obvious: this is not how most people want to create. Even as a confident, fast writer, I can’t imagine the average person preferring to write their vision out this way.
Takeaways
Writing is not the universal creative unlock people think it is.
The future of AI-powered creation will need to be visual, tactile, or otherwise non-textual for mass adoption.
Startups and investors should be skeptical of “prompt engineering” as a broad consumer behavior.