REA Reverse – Engineer Anything

> It really depends on your perspective of what the point of the hobby is. Were decomps/recomps a hobby because people wanted to make sure their old games could run on modern hardware and be extended with mods etc., or were they a hobby because people enjoyed the actual process of using Ghidra and reverse engineering by hand, and the decompiled game was just a side effect?

I think your question would make sense if these were the only two reasons. Someone who just wants to make sure their old games can run would see these tools as a good thing. People who just enjoy reverse engineering by hand wouldn't care this exists, or maybe even, a helpful tutor.

So perhaps then, the sentiment being as negative as it is, is revealing something about the hobby: that many participants, considered the metagame of being the first to do it, to be "smart enough" to have figured it out, the "street cred" if you were, is what they are lamenting, but they just either lack the words, or aren't being honest with themselves.

What's a VST?

> Were decomps/recomps a hobby because people wanted to make sure their old games could run on modern hardware and be extended with mods etc., or were they a hobby because people enjoyed the actual process of using Ghidra and reverse engineering by hand, and the decompiled game was just a side effect?

It’s interesting to dwell on what ‘we’ appreciated before versus what we appreciate now. Since LLMs lowered the bar, it feels like raw human effort has come to be held in higher esteem.

Are we going to re-evaluate what we were impressed with in the past based on that new metric? For example, will accomplishments by experienced engineers be devalued in comparison to accomplishments by outsiders? Will the software engineering equivalent of ‘folk’ or ‘outsider’ art become more valued? There will be a natural instinct to distrust outsider contributions because, well, how did they get so good?

Are some metrics of human effort going to be de rigueur going forward?

I think you misunderstand the OP, the issue isn't the work if AI can do it we should still pool together and not waste tokens on the same problem 10-100x times because 100 people are doing the same thing in parallel.

Since it's not educational as it used to be, if you built your own version of n64 emulator number 10143 you got something out of it you learnt some interesting stuff, learnt software development, hardware etc.

Now we are literally all just taking a magic machine pouring money on one end getting half baked output the other but no progress is being made on money is being wasted. Since even if 1 person had built it everyone would have had access already.

This is the same wave of slop coding/vibemaxxing we had a few months ago where everyone was building product X for the 1 million-th time without any actual progress being made.

Math has a similar problem now, before if 1000 people tried to solve a problem they all grew from that experience, the LLM is not growing/learning anything from this effort.

All we are doing is burning resources for our pleasure, I say that as someone who has now written his own compiler, vcs and browser stack, with vibe coding, though largely I would claim I have used and tried to read the code, clean it up, to learn from it where different things break.

But even this feels absurd waste of resource imagine all of us collectively burning 200$ per month plans which assuming a at cost pricing we are all burning 200$ of electricity each month to do what exactly? Some folks are on their 10th subscription that means most people are burning 1000-2000$ of ultra cheap energy generally sourced from a mix of gas/oil turbines burning away at big data centers these days.

We could all solve it by making a site.. cough github.. and not duplicating work.. cough..

But somehow because it's easy no one seems to care.

I am hardly against software progress, and if AI gets better we will have better software (because most people write horrible software most of the time) so I am happy about it, to an extent that I can be, with my profession feeling somewhat challenged.

But given I have always worked as a fixer of broken stuff it's very much also feels fine to me. Either way I am very much against these 200-500 dollar plans that allow you to burn hundreds of dollars of energy for little productive gain duplicating the same work 100th time.

I am sure software developer using 100-200$ plans just for work can get more out of it. And it's probably making up for the effort they would have put in their work otherwise.

It's a bit more than a need to stand out. It's a recognition that the definition of value has moved.

If a thing is no longer difficult to perform or to produce, then I might still employ or consume it if it's useful, but I will no longer value it. No matter how useful a potato is, it's not worth any money, time, interest, effort, emotional investment, etc. Even though in a vacuum it's almost like a magic ultra utility food, most useful and valuable thing ever.

Same for countless mass produced utilitarian goods like plastic bowls and basic computers.

I don't want be an injection molded cup or a potato not just because they aren't rock stars, but because they aren't valuable enough to be a person's identity or livlihood at all, even a basic unassuming workman's.

> perhaps it is going to be more along the lines of models, variants, self hosting, etc.

Unfortunately those are also better discussed with LLMs.

I recently had to refresh on some algorithms and some manual coding for an interview where using LLM is forbidden. It was super fun, I forgot how fun solving software puzzles can be!

It’s unfortunate our industry has become so boring. Because our society values capital above all else, quitting my job to do something fun would relegate me to economical loser status. We really need a shift to the left in politics and a way to redistribute the productivity gains this technology allows.

> AI is just a tool, if your ability to form relationships was easily displaced by a tool, then I'd say it says a lot about the ability rather than the tool.

In abstract, sure, but that is a much shallower discussion that many people are strongly disinterested in. It's generic, where a human decompile may not be (as a cloner I haven't looked into those projects). I don't like the idea of a group of fans of art frames saying don't bother looking at what's in the art frame, our conversation is equal.

You can draw a line from Xenowhirl's disassembly of Sonic 1 and Simon Wai's research of the Sonic 2 prototype to the creation of Sonic Mania and the blossoming fangame / ROM hacking community, and from there the careers of a lot of indie devs. It's not that these people aren't capable of talking to each other, it's that they get to know each other in a personal and professional context by bonding over a project they have a shared passion over. If these disassemblies just popped up on GitHub one day, and if the fan games and the ROM hacks were created by solo devs with an AI model... well, we end up with quite a few neat projects. But we don't end up with the rich, vibrant communities that get created in the inherent process of collaborating on a project.

Of course, in a post-AI world people will still want to work together on things. But these sorts of tools do disincentivize it. That's the point I'm getting at here.

...and today even adults are served explicitly with all these LEGO editionsn of whatever building or classic car,for above500 USD+ :-D

You might have noticed throwing billions around has diminishing returns

It's harder in the exact way that AI is good at, though. You need to slog through very large amounts of boilerplate until you find what you're looking for.

It's not really more complex once you're practiced at it, just tedious.

> that will be abandoned in a week, because the whole point was to get that 1M YouTube views.

Longbets 1 week, haha.

We're working our asses off on this.

Most of the team are artists who use these tools actively and we want the replacements for ourselves. I'm a filmmaker, so you can imagine my frustration of being bitten by the "unsubscribe fee".

Not really. Someone using Adobe Suite of products cannot easily switch to Davinci Resolve or Dark Table.

These are NOT the same thing whereas this one particular suite of clone products is exactly menu by menu, keyboard shortcut by keyboard shortcut and feature by feature a pure clone of corresponding products to that extent that they even have a separate dashboard just to track feature parity: https://github.com/storytold/craft-repo-status-app

It is NOT there yet of course but it won't be there in a year? That's an absurd claim to make. I have checked the repos multiple times and every new release keeps fixing something that wasn't working before.

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

/s

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.

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

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?

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

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.

Local inference solves that.

We used to pay for phone calls not long ago. Now a video call is free.

use open models and abliterate them

This kills the talent pipeline, and it'll create a spam problem for the other folks because now people will try to github PR spam their way to getting on a CVE.

It's worth talking about the fact that you can't even talk about DMCA to a model trained on the Library of Congress unless you're one of the approved people. And that's before reverse engineering something or writing code.

So in this future, it sucks to be you if you're someone trying to make your small app more secure, someone trying to upskill, a tinkerer trying to bypass corporate lockdowns for a device they own (a recognized DMCA exception, btw), a teenager trying to learn about security...

It locks away much of the richness that produced hacker culture behind glass. You can look at their press announcements and PR pieces, but you can't touch.

And as they're lobbying the government for "sensible regulation," this inevitably leads to a future where computing is controlled.

It's the direction their existing reports are taking. They recently released one in September that talked about how they stopped "bioweapons." What were said bioweapons efforts? Oh, it was scientists using Claude for grant writing, paperwork and grammar. At national labs.

These people are basically proud of impeding real research to make better painkillers and study a neglected tropical disease, https://news.ycombinator.com/item?id=49651727

And this is being used to lobby against "dangerous" open-weight models because gasp a scientist might use them to write a grant! To make better antidepressants.

At what point do they start reporting someone taking apart an iPhone and trying to DIY a repair with a schematic as a thwarted "cyber security incident?"

I have CVP with the new program (including mythos access) and still get constant denials for silly situations. Most recently I fed a URL to my agent from a security blog and asked if to add it to my obsidian vault with appropriate tags-- cyber flagged. You're not missing much. OpenAI and/or most Chinese models are much more lax in their restrictions.

Vibe-reverse-engineering

I mean thats not why I have Claude decompile things. I have it so it because my software should work exactly how I want it to.

Good point, I'll add "sync the jellyfin playback stats back to Spotify" to the prompt. Maybe with a multiplier for artists I want to support extra.

Yes! APKMirror is a good choice. AndroZoo works too, but it's academic and you need to apply for API access.

You should be able to get free APKs straight from Google via Aurora Store.

Yeah installation has always been such a security issue. So many programs are just random links that download a file. You have to trust that the host has not been compromised all packages that were used to build it were not compromised etc.

With ai models getting better we may be able to do analysis on the actual underlying bytes of the files we download to properly scan them for malicious code patterns and build systems which sandbox programs and watch inbound and outbound traffic/ system level actions from them and flag suspicious requests for further analysis by smarter models.

REA shows that ai are very good at understanding low level code and reverse engineering it so this could potentially be applied to application level security aswell.

A big difference between the safety of "next > next > next > finish" and "curl | bash" is one of them is dynamically loaded from an external source that could change between runs, and the other can be fully downloaded and vetted in a single check, and then once it's safe, it's probably safe 10 years from now.

>installs capstone

hello fellow Claude user

Why wait months for Nvidia to fix their compiler bugs when its far more faster to do it in the open? Libraries that invoke nvcc don't want to wait either.

So really closed-source compilers are a hindrance and you're subject to unspecified time-frames to even get to fix your issue rather than doing it yourself.

Intel and AMD have already open up theirs without question. Nvidia is the one who claims to be supporting openness in AI but can't even open up nvcc let alone CUDA.