I build guitars for a living.

Twelve years of it. I know where a neck wants to sit. I know how a top responds when you take another thirty thousandths off it, and why one piece of maple works when the board next to it in the same billet does not. That came slow and it came through my hands. There was no shortcut and I never went looking for one.

So when I say the barrier to making film came down, I am not saying craft stopped mattering.

I am saying I got in without paying the toll. For most of my life, making a moving image that looked like anything meant a crew, a budget, a decade inside software, or all three. I had none of that. I had a day job with sawdust in it.

Now I make video. Not tests. Not storyboards. Finished pieces that people watch to the end.

No crew. No budget. No film school. A guy who builds guitars for a living, and a browser tab.

Runway is a partner of mine and I have a code with them. Glenn25 takes 25% off any plan at app.runwayml.com. I would be paying for it either way, which is the only reason I am comfortable putting it at the top.

The models got good. That part everyone knows.

The part that matters more is that they got assembled. A year ago, making a finished piece meant five subscriptions, four export steps, and a lot of dragging files between tools that had no idea the others existed.

A lot of my work lives in Runway now. A chain that used to run across five tabs runs across one, and when the pipeline stops leaking you stop losing the idea somewhere in the middle of it.

There are good single-purpose tools for every step of this, and a few of them beat Runway at their one thing. What Runway gives you is the whole thing in one place, which is what most creators actually need when they are trying to move up a level.

The thing that changed is not technical. It is attention.

A short accounting, because the scale of it is easy to miss from outside.

Every model, one login. Gen-4.5 is the flagship video model. Sitting next to it: Seedance 2.0, Kling 3.0, Nano Banana Pro, Seedream 5.0, FLUX.2 max, Gemini Omni Flash. Six labs, one credit balance, one asset library.

Editing that respects the shot. Aleph 2.0 and Edit Studio arrived in May. You change one thing in a clip and everything else stays put. Thirty seconds at 1080p. Swap a color, change the time of day, remove the thing you did not notice was in frame.

An agent that builds the whole video. Describe what you want, argue with the concept it proposes, and it produces the scenes, the voiceover, the music, assembled and handed back to you.

Node-based workflows. Chain models together, save it as a template, run it forever. My character consistency chain took an afternoon to build and has run a few hundred times since.

Audio and a timeline. Speech, sound design, and score from a prompt. Performance capture onto a character image. A real timeline to cut it all in.

I am not going to pretend I use all of it. Nobody does. But the floor moved, and the floor is what stops people.

This started as one still image and a sentence about what should happen next.

You will be bad at it for a while. I was. That is the same deal I got with wood and it turned out fine.

Two things shortened the curve for me more than anything else.

Learn the prompt structure. For text to video: camera shot of subject performing action in environment, then supporting detail. You need both a visual description and a motion description. Leave one out and the model invents it for you, and it will not invent what you had in mind.

For image to video, describe motion only. The image already carries the composition, the lighting, and the style. Your words have exactly one job: what happens next. Describing what is already visible in the frame is wasted text.

Use the free training. Runway Academy is free, no plan required, and almost nobody I talk to has opened it. Courses on advertising, visual effects, games, character performance, and custom workflows. There is a prompting guide in there that is better than most things people charge for.

Two techniques worth stealing on your first day. Sequential prompting, where you give the model an order of events with rough timestamps, so it stops rushing everything into the first second. And last frame extraction, where you pull the final frame of a finished clip and feed it back in as the start of the next one. That is how you get sequences longer than the model's limit.

Start free. It is 125 credits and it does not expire, which is enough to find out whether you like the feeling of it.

Watermarks come off at Standard. That is the floor for anything you intend to publish.

If you are making things weekly, Pro is the honest starting point. Standard runs out faster than you expect once you count the generations you throw away, and you will throw away a lot of them. That is not a flaw in the tool. That is the work.

One detail that decides the plan for people whose creative time comes in bursts: Standard and Pro credits reset every cycle and do not roll over. Max carries one month forward.

Pro drops from $28 to $21. Max drops from $76 to $57. Standard drops from $12 to $9.

The code is Glenn25 and it works at app.runwayml.com on any plan.

I am telling you plainly that this is a promotion and that I am a Runway partner. I am also telling you that I ran the piece below after a full shift at the bench, on the same subscription you would be buying.

Made after a full shift at the bench.

Here is the part I actually want you to take away.

I did not become a filmmaker. I am still a luthier. I still spend my days with sanders and wood glue and a dust collector that is too loud.

What changed is that the thing in my head stopped being out of reach. It turned into a skill. I practiced it badly, then less badly, then well. Same as everything else I have learned.

There is no gate on this anymore. There is just whether you sit down and start.

Glenn25. 25% off any plan. app.runwayml.com.

Glenn Williams is a luthier at PRS Guitars, an Adobe Firefly Ambassador, and a creative partner with Topaz Labs and Runway ML. He writes The Render, a weekly newsletter on AI image and video generation, and produces the AI-animated folktale series Stor-AI Time. Find him at @GlennHasABeard on X, Instagram, and Threads, and at glennwilliams.net.

Video in this article was made with Runway. Platform details verified against runway.com and academy.runwayml.com in August 2026.