Useful?
If someone can rebuild your product in a weekend, what are people actually paying you for?
Would Adobe even win if they sued?
That's the question I keep coming back to this week. A developer used AI to rebuild rough versions of Adobe's apps and started giving them away for free. From what people who've tried them are saying, they're buggy and nowhere near finished.
Adobe hasn't done anything about it yet. I've been going back and forth on what they should do, because every option I play out costs them something.
The originalYears of decisions
The cloneA weekend, rough edges
So what if Adobe sues?
I'm not a lawyer, so take this as me thinking out loud. But "Adobe wins" means very different things depending on what they actually sue over.
The name and the logo. If the clones borrowed "Photoshop", the icons or the look, that's trademark, and it's probably Adobe's easiest win. The catch is that winning just means the clones pick a new name. That's about it.
The code. Copyright protects Adobe's actual code, not the idea of a photo editor. If the AI reproduced chunks of it, that's a real problem. If it rebuilt what the apps do without copying how they're written, it's a much harder case. That kind of rebuild has been legal for a long time. Clone PC makers rebuilt IBM's BIOS this way in the 80s, and in 2021 the US Supreme Court sided with Google over its reimplementation of Java's interfaces.
Patents. This is the one that would actually hurt, and I think it's the one people underrate. Adobe holds more than 6,000 granted patents worldwide and adds hundreds more every year, across image editing, video, PDFs and design tools. A patent doesn't care whether a person or an AI wrote the code, or whether anything was copied. If the clone does the patented thing, it's in trouble.
The licence. If the rebuild meant taking Adobe's software apart in ways its terms forbid, that's a separate contract fight.
Trademark
Names, logos, icons, the look
If Adobe wins · The clones rename
Copyright
The actual code, not the idea
If Adobe wins · Only if code was copied
Patents
Specific inventions, however they're built
0+granted Adobe patents
If Adobe wins · Feature by feature
Licence
The terms of using Adobe's software
If Adobe wins · Depends how it was built
So if Adobe wins, the likely version is narrow: a rename, a takedown, maybe a patent claim against specific features. The dramatic version, a court saying AI-built software infringes because the model learned from Adobe's work, would be brand new law. It would also be awkward for Adobe, which ships its own AI.
If Adobe loses, I don't think much changes. Rebuilding features was already allowed.
And if Adobe never sues at all, that might say the most. Suing one person who gave away free, buggy clones risks making them famous. Quiet takedowns are cheaper.
If Adobe wins
Probably narrow: a rename, takedowns, a patent claim or two.
If Adobe loses
The rules barely move. Rebuilding features was already allowed.
Most likely
Neither. Quiet takedowns are cheaper than making one person famous. And either way, what changed for good is the cost of copying.
The law barely moved. The cost did.
That's where I keep landing. Nothing about the rules changed this year. What changed is how cheap it is to copy something. It used to take a team and a year. Now one annoyed developer can get partway there in a weekend.
I've felt that myself, for much more boring reasons.
Our website audit used to take me three to four hours of crawling, spreadsheets and screenshots. I rebuilt it as an AI agent loop over a weekend. It now takes 47 minutes, and the findings are deeper, not shallower.
Our website audit
0minutes, start to finish
Same job. Deeper findings, because the agent doesn't get tired on page seventeen.
VectorizerKit started the same way. The tools I had for turning a logo into a clean vector file either uploaded the image to someone's server or handed back a file with thousands of messy points. So I built my own engine for that one job.
Neither of those was hard because of the code. The code was the fast part.
What got cheap, and what didn't
Building a feature got cheap. Very cheap, and it's getting cheaper every month.
Knowing which feature to build didn't. Neither did knowing who it's for, what they'll actually do with it, and what they'll hate about it on day three.
Got cheap
- Writing the code
- Copying an interface
- Shipping a feature
Didn't
- Knowing which feature to build
- Knowing who it's for
- Knowing what they'll hate on day three
It makes me think about my own work too. I've spent ten years designing things for clients, and a lot of it lives online where anyone, or any model, can see it. Some days I wonder how much of it is already in there, and I'm not comfortable pretending that question is settled. Plenty of people's work went into these models without anyone asking them, and the courts are still split on it.
But I still believe the part people actually pay me for is the conversation before anything gets built. Every week someone comes in asking for a feature: a chatbot, a dashboard, an app. Most of the time it's a symptom, and the real problem sits one step behind the request. A clone can copy what I shipped. It can't copy the conversation that told me what to ship.
The test I'd run on any business
If a competitor could copy your website, your booking flow and your quote tool overnight, for free, would your customers still pick you?
Here's a ten-minute way to find out. Write down the last three things a customer said when they recommended you to someone. Not what you think they'd say. What they actually said.
Try it · Takes a minute
The last three things a customer said when they recommended you
- 1
- 2
- 3
Mark each one as a feature or as you. (0/3)
Nothing you type here is saved or sent anywhere.
If they're about a feature ("the online booking is so easy"), that's worth protecting, because it's the part that's about to get copied.
If they're about you (how fast you answered, what you talked them out of, how well you got what they were going for), you're in a much better spot than you think. That part doesn't clone.
Where I land, for now
For the businesses I work with, the practical answer is the same whatever happens in court. If your product is the moat, the moat is shrinking. If the moat is how well you understand the person on the other side, AI just made you faster at the part they were already paying for.
The software is getting cheaper to copy. Knowing what to build isn't.
So, two questions. Should Adobe sue, yes or no? And what would your three be? I'd genuinely like to know. Tell me on LinkedIn.
Was this worth reading?
More from the notes.
- 01NOTE·1 Min Read
Watched a Client Click the "Regenerate" Button Six Times in Thirty Seconds Today, Getting Six Slightly Different Answers.
None better than the first. The model wasn't the problem. The interface was teaching them that more clicks = better output. We added a one-line subhead under the button: "If the first answer feels…
- 02BRIEF·1 Min Read
What They Asked for: A Chatbot. What They Actually Needed: A Better Intake Form.
A services company came in wanting "an AI chatbot for the website." They had three quotes already, all in the $40–80k range. They wanted to know what we'd charge.
- 03BUILD·1 Min Read
Rebuilt Our Website Audit. It Now Takes 47 Minutes.
It used to take three to four hours. Manual crawl, spreadsheet of issues, screenshots, a write-up. Mostly busywork — the kind a senior person does badly because they've already done it ten thousand…
Let's Build
Let's make it reeaaalll.
Bring it in rough. Half the work is figuring out what it actually needs to be.
Tell Echo what you're trying to make happen…Hiring? brenden@thinkpurple.ai →