Most of my career has been in creative software.

I’ve spent years watching designers, editors and creative teams work. And no matter how good the tools get, an enormous amount of our time is still spent operating them.

Moving files. Renaming layers. Updating layouts. Cleaning presentations. Relinking assets. Exporting versions.

None of this is difficult.

But it takes time.

And when someone spends two hours doing something a computer could do in two minutes, that’s bad for them and bad for the business. It keeps people from focusing on the idea and the creative work.

We can describe what we want in ordinary language and have a machine understand us. Then we open the software where we actually work and go back to clicking through menus, panels and dialog boxes.

One idea is that AI will replace all of this software.

I’m not so sure.

People have spent years learning these tools. Companies have spent decades building workflows around them. And much of the software is already very good.

InDesign is great at publishing. Excel is great at spreadsheets. Cinema 4D is great at 3D. Keynote is great at presentations.

They give us precision, control and editability. They work with the files, templates, plug-ins and systems we already depend on.

Maybe we don’t need to replace them.

The other idea is to put AI inside every application.

That’s already happening.

But real work doesn’t happen inside one application.

A project might start in Finder, move through Photoshop and InDesign, get reviewed in Keynote, tracked in Excel and delivered through a browser.

The application changes. The project doesn’t.

So why should the intelligence start over every time?

What if there were one layer that understood the work and could operate across all of them?

That’s why I started building Parasite.

Keep the tools. Lose some of the work.

Press ⌘/ and tell it what you want.

Parasite figures out how to do it.

Instead of telling the computer every step, you tell it what you want done.

Parasite works across the tools you already use and learns how to operate them.

And when it figures something out, it remembers.

A task that required reasoning the first time shouldn’t require the same reasoning the hundredth time. Parasite should get faster, cheaper and more reliable the more it works.

Eventually, it should just know.

That’s the bigger idea behind Parasite:

Can software learn how software works?

If it can, we may not need to replace all the software we use.

We might just need a better way to use it.

Philip CronerudFounder, Parasite