Hacker News Reader

Hacker News

Comment by echelon | Hacker News Reader

Parent Full thread

echelon

No, we explicitly do not use reverse engineering. We do not decompile binaries, we do everything 100% clean room:

https://github.com/storytold/photocraft (inspired by Photoshop)

https://github.com/storytold/wordcraft (inspired by Word)

https://github.com/storytold/pdfcraft (one of the more mature apps)

https://github.com/storytold/vectorcraft (another app close to 1:1 parity)

(etc.)

Using REA for something as high profile as what we're doing is likely to result in lawsuits. We're doing everything we can by the books.

We cannot look at Adobe sources. Use of Ghidra is disallowed.

REA is probably great for personal apps and for abandonware, but I think if you publish the results and it's found to have decompiled the original proprietary sources in discovery, you might be in for a bad time.

View on HN

Using an LLM is not a clean room, imo. Its just IP laundering. Which is fine I guess if everyone is doing it, including the companies you're stealing from. I just dont know what the implications will be for progress.

Licensing/copyrighting encouraged people to think up of new things, and new ways of doing something. Now we're just all copying eachother.

This is novel Rust/egui code that I imagine looks nothing like Adobe code. I've never seen their code, but it must be a mess of old C++, right?

It's considerably faster than their apps (at startup) too.

Well I hope they trust their LLMs completely, because I'm sure Adobe is currently going over their source code with a very fine comb to build a case. And they have LLMs to help with comparing too.

If the code is in a completely different language, what can Adobe do it's not gonna be copyright infringement? Derived work if the algorithm is the same? But at that point it's math.

I guess they can always sue over infringed patents?

I agree. The vast amount of data ingested and internalized by LLMs has effectively been "laundered". But they are so powerful and evolving so fast that no one can be spared of their impact. We have to to learn to live with it. Traditional proprietary software being "laundered" is just one part of the broader story...

all human culture and creation is a form of "ip laundering" i guess

That's an interesting perspective. Indeed human creativity is often a process of reinterpretation, reinvention, and renewal. We've seen sprial-like patterns of similar ideas thoughout history, especially in the arts, intellectual trends, and lifestyles(but maybe not tech). Hopefully, LLMs can help human accelerate these processes in ways that benefit us

Doesn't clean room typically apply to cases where the "dirty" team has legitimate access to copyrighted code, like the IBM PC BIOS which was published in the technical reference manual, and uses this access to write a functional specification for the "clean" team?

I'm not sure how clean room would apply to commercial applications distributed in binary form, as there's no way to look at even disassembled code without violating the license agreement and therefore being in breach of contract and subject to potential copyright infringement claims for copying or even continuing to use the software, let alone cloning it, and surely you're not going to be subject to a copyright claim based on familiarity with the application from merely using it.

The issue is that LLMs very likely have access to the source code of these products as part of their training data, and are then being used to generate clones. There's no barrier in the middle to ensure copyright violations don't leak

The reason why 'reverse engineering' has gotten so good is because what we're actually seeing is fully automated luxury plagiarism

It is very unlikely that LLM had access to original source code of Photoshop.

If adobe use any AI tools to help develop photoshop, its all been loaded in. Or if its ever ended up anywhere on the internet, even in private

The first part (re: AI tools) is ostensibly not true. It all depends on the degree to which the data that they "don't train on" gets laundered and anonymized to the point where they are able to justify training on it without it being "yours" any more.

And I don't think that is publicly known at the current time. (If anyone has any tangible info on this, the please let me know...)

It could have access for code of an older version with how there has been leaks of sources of various commercial applications over the years.

They have had access to plenty of A/GPL software. They are all highly problematic from an ethical and legal point of view.

> It is very unlikely that LLM had access to original source code of Photoshop.

It doesn’t matter. It’s likely had access to the knowledge of somebody who worked on that source code.

I have worked at companies that have done clean-room implementations. I was not permitted to look at their source or interact with those teams, because I had been exposed to the source of what they were re-implementing in a prior job.

As a human with a mushy brain I was never going to remember the source, but the fact I had been proximate to it was enough to lock me out. You have no idea what the LLM trained on, so you can’t prove a LLM derived clone is a “clean-room” implementation.

Some questionable idea that the LLM "had access to the knowledge of somebody who worked on Photoshop source code" seems like a poor basis to allege it's not a clean room implementation.

Wouldn't the burden of proof here be on Adobe if they wanted to make such an argument?

But look at the bright side in your reality: this technology can losslessly compress any source code at several orders of magnitude better than any prior algorithms!

Isn't that an incredible breakthrough in your world?

The practical uses are more applicable to laundering open source code covered by copyleft licensing like GNU.

People made new and better things just like companies without copyright too. In recent times companies made things worse too.

What intellectual property is being laundered here if it's not even the same ecosystem (C++ vs Rust)? If it's so easy, and you just need to rebatch something existing, why has nobody done this over the last 20 years?

Adobe almost certainly has patents covering aspects of Photoshops design and tools?

Microsoft famously has patents covering aspects of the ribbon interface in Microsoft office, which of course this suite must implement. https://en.wikipedia.org/wiki/Ribbon_(user_interface)#Patent...

Meanwhile, Wikipedia has a policy that says screenshots of applications have to be "as small a version as possible" in order to meet the fair use exception for presentation of copyright work. I can only imagine an exact clone of the interface could be similarly ruled to infringe on Adobe's intellectual property.

Copyright is so flawed... This would mean an hommage is a breach of intellectual property in general, even if you paint it completely differently (say watercolor vs oil)

Patents don't forbid you from replicating the technology for yourself.

Your brain is an IP laundering machine. And it has been doing that even before we had any IP laws.

>Licensing/copyrighting encouraged people to think up of new things

I don't think it did... think about it for a second. If you have to recreate something from scratch every time vs. being able to/being allowed to take something existing and improve it, which regime results in more innovation? Probably the second, because progress builds on each other. I might have a better idea to implement more features, you might be able to optimise it, someone else might make it more polished, someone else might add a nifty UX trick, someone else might refine the art, etc. etc.

Traditionally, this was only done on open source projects and even there, the barriers to contribute have been historically high. (Have to open merge request, get through the gatekeeping, document your stuff, etc.)

Now it's much easier, point agent into code (or binary), change stuff, have fun;)

How is this stealing?

> Licensing/copyrighting encouraged people to think up of new things, and new ways of doing something.

I don't think this has been true since the copyright acts extended copyright from a max of 28 years to a minimum of 70 years + life of the author (or 95 years for corporations). I think this is kind of the inevitable conclusion of such absurd length copyright.

Copyright is no longer used as a grace period and reward for innovation, it's a tool for corporations to hoard and restrict IP. A core aspect of human creativity is building off other things and things you grew up with, but now people have zero chance of seeing ANYTHING created in their lifetime entering the public domain.

This is an inevitable rebellion to the situation these laws put us in.

Licensing/copyrighting encouraged people to defend their licenses and copyrights instead of improving their ip. Information should be free.

Please don't take offence by this, but:

If AI models can generate designs faster and produce work that is "good enough", what is the actual future of design tools and the design profession in your opinion?

I've tested this myself with some frontier models and the results are kinda impressive enough that it raised the question if design skills are already obsolete. If that is the case, what use are these tools now?

Good enough is not really good enough.

I've seen a lot of "really cool unique designs" that are obviously just a digested regurgitation of amalgamated corporate slop. Anyone wanting an actual unique design language needs to hire actual designers.

Those designers may use AI tools but the tooling isn't to where they can be completely replaced. I challenge anyone to show me an e2e LLM design toolkit that can actually replace a designer, not just one shotted "wow that looks so cool" character designs.

In the future everything will be some form of command line interface. Don’t worry about design.

The death of IP in the west. Good thing? Bad thing? Who knows. Definitely the start of something big though.

Mightn't be so bad if all the value wasn't being captured by trillion dollar corps, despite them benefiting off the labor and creativity of countless non consenting individuals

Those trillion dollar corps also employ millions of people in very high paying jobs. And the extraordinary stock returns have made their engineers (and countless public investors) wealthy across the past half century.

People wishing for the death of the mega corps, are directly wishing for the obliteration of millions of great jobs.

I’m not sure people are “wishing for the death” in the literal sense.

We just need to apply the antitrust law and break these trillion-dollar corps into smaller competing corps.

There’s no reason for a single company to provide your email, browser, search, operating system, entertainment, AI, etc.

How many of those jobs do you expect them to retain as they build increasingly capable models?

How much of the profit will be redistributed to all the artists, programmers, scientists, etc. whose works are the foundation of their product?

The poorest 50% of global citizens hold the same combined wealth as the top 10 individuals.

How are massive and increasingly concentrated corporate profits going to help them? (Or anyone else for that matter).

Cry me a river

>People wishing for the death of the mega corps, are directly wishing for the obliteration of millions of great jobs.

Yeah, I'll take the obliteration of a few million overpaid American jobs over the destruction of hundreds of millions worldwide, the social upheaval that goes with it, the destroyed families and the direct and indirect deaths. Fuck your promo packet and fuck your cozy desk job.

well they won't give me one of those great jobs

IP is not dying. It’s being consolidated to just a few mega corps. Good thing? Bad thing? Definitely the start of something big though.

Imaginary Property was always an illusion, one which is now crumbling quickly despite a lot of desperate attempts at maintaining it.

Everything was, is, and always will be a derivative work.

How do you know llms you are using were not trained on decompiled apps?

why would that matter? That sounds like a problem between Adobe and the AI companies.

> We do not decompile binaries, we do everything 100% clean room

AFAIK clean room does not forbid decompilation, I meant that's how it's been done for ages. What's important is that the observer does not write the code, to avoid being impressed by the decompiled one, risking creating a copy close enough to deem it copyright infringing on the original.

Because the copyright applies to the actual code.

Would love to see an InDesign clone that supports writing to and from their proprietary format.

Wow. Maybe AI can win artists over after all.

great work, congratz.

The AI labs already figured it out:

1) reverse engineer the code 2) train a model on the code 3) use the model to write the clean code

Step 2 is the key “cleaning” process

So maybe a good strategy would be to use something like REA, put it on GitHub, wait for the LLMs to train on it, then just use the frontier models

I prefer the word *laundering

LLM's have already trained on similar enough code

Even better, 1 and 2 done, just move to 3 and profit

Reply on news.ycombinator.com