One ChatGPT session, one finished video

I have lost entire evenings to video editing

July 23, 2026

In partnership with

Timeline scrubbing. Export settings. Re-rendering the whole thing because one frame was a hair off. If you have ever made a short video, you know the specific kind of tired I mean.

So when I hit a post from someone who took an idea to a finished one-minute video without ever leaving ChatGPT, I stopped scrolling on the spot. No editing app. No jumping between six tabs. One session, start to finish, with the files landing right on their computer.

The stack: Skills, plus GPT 5.6, plus GPT Image 2, plus a video model, plus Remotion. That combination is the whole trick, and I want to walk you through how the pieces snap together.

Win AI Search Without a Big Team

92% of VCs use AI to find companies. 58% of buyers start there, too. If you're not showing up in AI answers, you're invisible before the conversation even starts. Join HubSpot for Startups, Anthropic, and Marketing Against the Grain on July 30 (11 am ET) for a live AEO teardown. Real startup. Real recs. Register and unlock the free Startup Visibility Bundle.

*Ad

The five moves, in order

Here is the workflow, with the reasoning behind each step:

Build the skills first. They made two custom skills up front, a storyboard skill and a Remotion skill. This is the foundation. A skill teaches ChatGPT how to run a repeatable task your way, so you are not re-explaining your process every time you open a fresh chat.

Describe the big idea. Not a shot list. Not a script. Just the concept. The skills handle turning that concept into structure, which is the entire reason you built them first.

Generate the storyboards and images. The storyboard skill maps out the scenes, and GPT Image 2 draws the visuals for each one. You get a full visual plan before a single frame of video exists, which is how professional video has always worked.

Bring in an outside video model over MCP. This is where the still images start moving. MCP, the Model Context Protocol, is the bridge that lets ChatGPT call an external video tool without you ever leaving the chat window.

Assemble in Remotion. Because Remotion builds video from code, the AI can write and adjust that code directly. No dragging clips around a timeline.

They made a sharp point here: if OpenAI still shipped its own video model, that fourth step would vanish. No MCP, no outside API, just one continuous system. That gap is the only seam in an otherwise closed loop.

Why the experiment matters more than the video

The goal was never really the video. They said it plainly: the point was to push the new ChatGPT to its limits, use case by use case, especially the jobs most people currently hand to Claude.

Five days of testing went into it. That is not a hot take posted an hour after launch. That is someone genuinely stress-testing a tool before forming an opinion, and I trust that a lot more.

Here is how they mapped the two systems against each other for knowledge and content work:

ChatGPT Skills work the same way Claude Skills do. The concept carries over cleanly.

ChatGPT Work plus Codex lands roughly where Claude Cowork plus Code lands. Comparable pairings for comparable jobs.

GPT 5.6 sits around Fable 5 or Opus 4.8 in raw capability, in their read.

That is a closer race than I expected before I read it.

How Jennifer Aniston’s LolaVie brand grew sales 40% with CTV ads

The DTC beauty category is crowded. To break through, Jennifer Aniston’s brand LolaVie, worked with Roku Ads Manager to easily set up, test, and optimize CTV ad creatives. The campaign helped drive a big lift in sales and customer growth, helping LolaVie break through in the crowded beauty category.

*Ad

Where it still comes up short

They were honest about the gaps, which is what makes the rest believable:

There is no equivalent of Claude Design yet. That capability simply is not there.

GPT 5.6 misses details. In their tests, it kept dropping the small stuff.

It runs slower. The same content task in the ChatGPT app took noticeably longer than on Claude.

So the verdict held: for content and knowledge work, Claude stays the easy call.

The part that is bigger than any tool

Here is what actually got me, and it has nothing to do with which app wins. Whatever you have built, your skills, your workflows, your prompts, moves to another system.

That is a real shift. It means the hours you sink into good AI workflows are not chained to one vendor. You are building portable assets, not renting a room in someone else's house.

Try it on your own week

If you are already paying for ChatGPT, their advice was blunt: push Work and Codex until they break. Here is where I would start:

Pick one repeatable task you do every week. Not your hardest one. Your most frequent one.

Build a skill around it instead of re-prompting from scratch each time. Write down the steps, the format you want, the rules you always apply.

Run that same task on both Claude and ChatGPT with identical inputs, and compare the output side by side.

Note where each one cracks. Speed, detail retention, format consistency. That is your decision data, not someone else's benchmark.

Once the basics hold, chain in an outside tool over MCP. That is the jump from a chat assistant to an actual production pipeline.

The thread running through the whole experiment: the skills did the heavy lifting, not the model. They built the storyboard and Remotion skills first, and everything downstream got easier for it. Front-load the structure and the model becomes almost interchangeable.

Tonight, open a chat and type one sentence: here is the outcome I want, ask me the questions you need to build the skill. Do not build anything yet. Just see what it asks you. That single question is the tell for whether you have picked the right first task.

Stop Paying for 6 Tools. One AI Does It All.

Most e-commerce sellers juggle 6–8 tools and pay hundreds monthly to keep operations running. StoreClaw replaces the stack with one autonomous AI engine that monitors competitors, optimizes listings, automates marketing, and tracks profit 24/7. Connect your store and let AI handle the work — no prompts, no complex setup, no credit card required.

*Ad