Two Prompts in ChatGPT Replaced a Video Editor and a Motion Graphics Designer

Marcin dropped a raw, unedited selfie video into ChatGPT running GPT-5.6 Sol, and two casual prompts later he had a tightened cut with custom animated motion graphics, no editor or designer involved. He opens the video with a text overlay that reads "Motion Graphic Designers are cooked," then walks through the receipts on screen. His argument cuts both ways. Clients can stop paying for basic motion graphics work, and designers who adopt the tool can deliver more volume at the same rates.

The Setup: A Raw File and a Lazy Prompt

The workflow starts with a file called IMG_6558.MOV, a raw talking-head clip with false starts and mistakes left in. Marcin uploads it to ChatGPT and types a prompt he later calls terrible: "edit this video, take out all the mistakes, make it nice and tight." The frames show the ChatGPT interface parsing the request and describing its plan, which includes removing the false start, mistakes, and long pauses. There is no timeline, no editing software, and no instruction more specific than that one sentence.

The First Pass: A Clean Cut With Deliverables

ChatGPT comes back with a completed edit and a plain summary: tightened from 29.8 to 18.1 seconds, pauses removed, audio leveled and cleaned up. The response includes an embedded 18-second player plus two download options, a 1080p share-ready version at 33 MB and a full-quality 4K master at 146 MB. That last detail matters. The model is not just trimming a clip, it is producing client-style deliverables in multiple formats without being asked.

The Second Pass: Motion Graphics From Inspiration Images

Marcin then feeds it reference images and a second loose prompt, visible on screen: create motion graphics that are animated, make them feel dynamic, time them to the exact words being spoken, keep them short and sweet, add logos, use your own judgment, and "be a motion graphic master." The interface shows the model working for 7 minutes 19 seconds across a four-step process, building three custom graphics. This is the part he frames as the real threat to designers, since taste and timing were delegated to the model.

The Result: Synced Lower-Thirds in One Shot

The finished 1080 x 1920 video shows an animated lower-third status card reading "AI Agents: Automation, Ready on Demand" with a progress bar, appearing exactly when he says those words. The model's notes describe the other graphics it built, including a sequential Role and Tone card for "personality" and a filling uptime ring with a timed number for "24/7." Marcin stresses this was a one-shot output from a deliberately weak prompt, run as a test to see whether it actually works.

What This Means for Both Sides of the Trade

His closing advice is blunt: go play with this and never pay for motion graphics again. But the opening line of the video offers the flip side, telling motion graphic designers to start using the tool themselves and charge clients for the output. The economics he implies are simple. The floor for basic editing and lower-third work drops toward zero, so the value moves to whoever directs the tool with better references and better judgment.

Key Takeaways

GPT-5.6 Sol in ChatGPT edited a raw 29.8-second clip down to 18.1 seconds, cleaned the audio, and delivered 1080p and 4K files from a one-sentence prompt.
A second prompt with style reference images produced three custom animated motion graphics, timed to the spoken words, in about 7 minutes.
The output included polished touches nobody asked for, such as multiple export formats and word-synced lower-thirds.
Marcin ran this as a test with intentionally weak prompts, so the results represent a floor, not a ceiling.
The advice targets both sides: clients can skip paying for basic motion graphics, and designers can adopt the tool to raise their output.