Will video editing actually save a software dev from the AI wave
I've been thinking about diversifying into something more "sensory" or high-bandwidth, like video editing. The logic is that while AI can generate a clip or stitch scenes together, the nuance of human pacing, emotional resonance, and high-level storytelling still feels like a massive gap in current models. It's a hedge—moving from pure logic (which AI excels at) to something that requires a mix of technical precision and human aesthetic judgment.
But is that actually a smart move, or just trading one volatile skill for another?
The pivot toward "Human-in-the-loop" skills
If we look at the current AI workflow, the most stable positions aren't the people doing the grunt work; they are the people orchestrating the output. This suggests a few different directions for anyone feeling the heat:
- The Orchestrator (AI Agent Management): Instead of writing the code, you become the person who designs the architecture and manages a fleet of specialized LLM agents. This requires deep systems thinking rather than just syntax memorization.
- The Creative Director (Multi-modal mastery): This is where the video editing idea comes in. You use AI to generate the assets, but your "skill" is the taste and the ability to curate a cohesive vision.
- The Physical Bridge: Skills that require interacting with the messy, non-digital world. Robotics, hardware hacking, or even specialized field engineering. AI is great at digital sandboxes but struggles with the entropy of the real world.
Why video editing is a tricky hedge
I've been looking into the tools coming out for editors, and it's a double-edged sword. On one hand, the barrier to entry is dropping. You can use generative tools to remove objects, extend frames, or even generate B-roll from text. On the other hand, this means the "middle class" of video editors—the ones doing basic assembly—might face the same displacement that junior devs are seeing now.
To actually make video editing a viable hedge, you can't just learn how to cut clips. You'd have to go deep into color grading, complex sound design, or narrative theory. You have to move up the value chain into the "taste" territory.
Rethinking the "Developer" identity
Maybe the mistake is thinking we need to leave tech entirely. Instead of learning a completely unrelated craft, the real hedge might be learning how to become a "Full-Stack AI Engineer." This isn't just about using ChatGPT; it's about understanding deployment, fine-tuning local models, and building custom RAG (Retrieval-Augmented Generation) pipelines.
If you can build the systems that others use to automate their jobs, you aren't the one being replaced—you're the one selling the shovels during the gold rush. It's a high-stakes game of staying one step ahead of the automation curve, and honestly, it's a bit exhausting to think about every single day.