Hacker News Reader

Comment by anthk | Hacker News Reader

It's happening. Check Artcraft. And Adobe could still profit from it. Custom plugins for companies, interop, optimization, rock-solid polished releases (and more optimized than ever) than vanilla MIT builds...

Also as a plus every user would be trained on Adobe's workflow. Companies would just pay Adobe for stable releases... and support. Ah, yes, the LLM's... sure. But when you have humans on your side, things get explained much better, even if Adobe points to an AI too to troubleshoot. They know how to explain problems to an AI and tell the best solutions to the user back and forth.

View on HN

How that's all going to go down remains to be seen, because it's not clear how Adobe and Microsoft (who have very deep pockets) view the situation or what their response is going to be.

While it is true that money could possibly be made indirectly and around the product (e.g. training, support, ads), the questions surrounding doing the equivalent of turning their IP into open-source remains. Usually, dominant companies don't like change or having to figure out new ways to obtain revenue, other than what's direct and obvious to them.

As for customization and plugins, that looks to be an AI versus IP problem too: (1) that companies would turn to Adobe or Microsoft, versus using their own AI (2) 3rd parties reverse engineering their plugins with AI.

Most likely way that it appears possible, is some type of licensing deal between the AI company and the original IP owner. Would be surprised if they do that smoothly, prior to lots of court fights, because corporate types usually don't like to share.

They can't prevent people from making their own in-house clones. They can take down Artcraft/whatever OSS repo but it's meaningless. I have my own C++ Photoshop/Office clones that I've been using for a few months now. Full replacement for anything Microsoft/Adobe has put out. How are they going to prevent me from using my own code? Or how will they prevent anyone else from just vibing up another Artcraft replacement (in case it gets taken down)?

Wait, who owns GitHub again?

True, at the moment, they can't stop the AIs or who is already in possession of the code. But, they can go after anybody hosting that code (as they have done similarly before).

And the AI companies themselves are also a target, to disallow asking or performing the reverse engineering of designated IP.

The Empire (can) Strike(s) Back (unfortunately).

There's always Gitlab, self-hosting, a tons of other alternatives (Fossil with the builtin server).

Any company that's of the US or from other friendly countries, will not stop courts from going after those hosting certain IP (if it goes down like that) they say should be taken down.

Yes, the counter would be to decentralize, but that doesn't stop going after the AI companies too. At the very least, a type of chilling effect can be created, where people and especially companies (with more to lose) would not want to risk using.

From the looks of it, the court cases are coming. Have no idea how things will turn out, other than this looks far from over.

Wine tells me otherwise from reimplementations.

Also, https://osgameclones.com

Otter Browser it's pretty much inspired from Opera 12 too.

And, for a much better example, XMMS vs WinAMP 2/3, virtually the exact same interface, with skins support.

GNU/Linux itself clones the Unix interface, API and ABI. And so did the BSD's.

Another example, Lesstif when Motif was propietary. Same API, and you could run lots of stuff written for Motif under LessTif. Now libre Motif superseded it around a decade ago.

What the music and movie industry did and does, in regards to IP, is the counter-example.

This is software, not media. Different laws apply. Unix and GNU mirrored 1:1 (and extended) a commercial, propietary OS (different Unix flavours). With FVWM + LessTif + gcc you basically had a reimplemented Unix workstation at home.

Bash, GCC, *utils, X.org, LessTif, ZIP vs PKZIP...

Also, every libre Smalltalk IDE out there had the virtual UI as the propietary ones.

Also, FreeDOS in the 90's and DOSEmu. Same virtual interface and usage.

Law applies to software too. Contract law (EULA) and copyright are applicable. Many EULAs have statements on reverse engineering. Another aspect of this is unfair competition or competing directly against the product the offending software was copied from.

Violations of the EULA or copyright can result in having to fight it out in court. It's a matter of if it will be taken to court (as little guys may not have the time or lawyers) or can be that both parties have decided on some type of agreement.

Again, you are like 40 years late, or maybe 50:

- Whole Unix clones, kernel and userland

- Unix userland utility clones

- Unix compiler and linker packs. Tons of them.

- Unix UI clones 99% close to the original API

By 2000/2001 you could run most GL accelerated stuff under MESA+DRI just fine as if it were an SGI machine, even replacing propietary Motif with LessTif.

- Windows API clones, both for running and compiling software, with WineLib+WineGCC or MinGW.

- Windows clones, ReactOS from GIT can be fairly complete.

- NextStep/Cocoa clones. Currently tons of software made for OSX 10.6 (in a libre way) can be compiled with both GNUStep+Addons and with XCode.

It's not just about duplicating code or a public API. It's also about the look and the way things are done in a product too.

There are trademarks and patents, in addition to copyright and contract law (any license agreements for usage of the software). Violations of any of these, can be possible reasons to be pulled into court, if that's what is decided.

Fair use is also not a blanket shield or exemption from copyright or various licenses, but rather has specific requirements and limitations, that can be what's fought over. Among these areas that are covered are how transformative something is, to what extent its a copy, and how adversely has it affected the commercial value of the product it was copied from.

These are all complex issues, that often lawyers and courts have to figure out. Have no idea as to what will be decided, but do think we are at the beginning and not the end, on what will be decided about using AI for the reverse engineering of commercial products and their IP.

Reply on news.ycombinator.com