My favorite tidbit of the series (but it's in Metal Gear Solid 3): in the torture scene where Volgin interrogates Snake in an attempt to discover who is the spy in his ranks, of all the persons present in the room (and if I got everything right from memory):

- Volgin is not an agent

- Snake is an agent (USA spy)

- The Boss is a double agent (Soviet/USA spy)

- Eva is a triple agent (USA/Soviet/Chinese spy)

- Ocelot is a quadruple agent (GRU/KGB/CIA/Philosophers spy)

And of course, after that Volgin then kills Granin thinking he was the spy, but he wasn't.

Try the classic analyses: https://www.deltaheadtranslation.com/MGS2/ and https://www.aumaan.org/form1/tus1/features/dreaming1.htm.

Kojima saw the writing on the wall so speak, and told us what the future held in a series of metaphors and dense monologues.

Mirror from the thread in case the Pixeldrain link does not work: https://gofile.io/d/Jf1tX8

very thanks!

Not much, just accurately predicted the next 30 years exactly

Dunkey MGS Explained https://www.youtube.com/watch?v=aaLiLRVeaZA

I need scissors, 61!

What did actually happen? I haven't played the game.

Trying to explain why a Minecraft Legacy leak is worth anything when we have decompiled Java takes too long.

check for console headers. those aren't that easy to get out of LLMs

Including the cardboard boxes: https://taskandpurpose.com/news/marines-ai-paul-scharre/

Surveillance states and mass media manipulation are a staple of sci-fi. Those topics were particularly common around that time too. E.g., The Matrix.

I've only played the original PlayStation 2 release, but wasn't the PS2 release of Metal Gear Solid 2: Substance the version with the most content? I'd take a leak of that one over the other versions.

There is no way you could recreate a convincing enough 90s era codebase of a japanese videogame + its associated tools + scripts and commented out codepaths with current ai tools.

It’s the real code there is code for known removed content (tanker escape scene and the 9/11 removed cutscene). Also AI can’t do what you’re theorizing yet.

>and ask them to decompile the game and structure and format as if it was the original source code. Asking them to ensure it compiles works fine

lot of people claiming this the end result is the AI downloading an emulator and rom

Whoa, since when is there a Jagged Alliance 3? Is it any good? JA2 is one of my favorite games of all time

It's (probably) a real leak. There are original comments in Japanese describing cut content and game logic that was scrapped in the final release.

How do you verify that everything is correct?

I wonder how they must have felt to have coded a 9/11 scenario of Arsenal Gear crashing into the Twin Towers, and then watched it actually happen a few months later.

This was cut from the game but is still in the source code leak, just commented out! https://www.youtube.com/watch?v=DNSPx-oRV60

Also note that the source code was leaked on April 30th :p

Everyone focuses on that part, but what the hell was with the girl that could repel bullets but it turns out that it was magnets but actually maybe she does have powers? And what about the weird thing where you start sword fighting your father/president/clone?

Yeah I know, sure, fake information on the internet. It's so prescient I guess, but the actual story is incomprehensible.

That’s great, now do kingdom hearts

>Also AI can’t do what you’re theorizing yet.

Did you try the above links? I haven’t shared the full source but all game mechanics listed in the ja3 guide including code snippets where helpful.

> It’s the real code there is code for known removed content (tanker escape scene and the 9/11 removed cutscene). Also AI can’t do what you’re theorizing yet.

There are lots of decompilation community efforts for N64 games, etc.

Someone should train a model on this. Giving the decompiled symbols good names, etc.

De-minification and de-obfuscation while we're at it.

It should be easy to generate a ton of "synthetic" (actually real) training data for this by simply compiling sources and using that as (input, output) pairs.

It’s playable but nowhere near as good and a lot of criticism is warranted. A true ‘70%’ game.

Gunplay is weak. Accuracy drops off waaaay too fast based on maximum range of the gun and burst fire has arbitrary damage reduction per bullet. So short range guns almost always missed (mechanics documented from source in the above guide) and if they hit they did little damage. It means the only viable weapons are long range weapons. Rifles and assault rifles. A submachine gun is worse than a sniper rifle even at close range.

The plot has a key gameplay changing moment that triggers waaay to early meaning you have to work to see much of the game content. Everyone tries to avoid the trigger on the second playthrough which is a silly thing to do game design wise. A desire to teleport across the map was the original motivation to the above from my point of view.

Enemies are bullet sponges in the late game too. A lame way o balance weak ai and gunplay.

It could have been as good as ja2 but they just didn’t refine the above enough.

Raw assets are probably the better tell

I’m sure the builds from doing what i’ve been doing won’t generate identical bytecode but it’s fun for the sake of messing with the game or understanding it (eg. The checksum logic for newstowers save game logic was cooy pastable as was the whole save game structure formatting itself and clearly matches the game - it works!). Likewise with all the JA3 mechanics documented in that linked guide.

What? He completely misses Metal Gear Rising: Revengeance. Luckily that's pretty straightforward: https://www.youtube.com/watch?v=41v3L0zCkNY

Must have been having a stroke...

Kojima knew that everyone had played MGS1 and were highly anticipating the sequel, so he deliberately trolled them all by baiting their expectations then subverting every single one by doing the opposite of what people wanted him to do. And he also spawned a lot of AI memes along the way.

I wouldn't be too sure about that. The original decompilations of Mario 64 and Ocarina of Time were done mostly by hand because LLMs weren't really around yet, but these kinds of projects seem perfectly suited for handing the gritty work off to AI: There is a clear output (exact binary recreation) and a straightforward path to get there (look at this assembly code and produce some C code from it). The decompilation of Twilight Princess jumped from very little to basically 100% of core code in the past year alone: https://github.com/zeldaret/tp

I have no doubt that this would be possible for MGS2 as well.

Absolutely. This is just some delusions of a vibe coder at best. Not with just current generation of AI tools but essentially never. The conversion from C, C++, Rust or whatever, through post-processing (macros etc), through IR generation, through compile time optimizations, through link time optimizations, to the generated machine code is a one way street for low level languages. You can get a pretty close higher level approximation that matches the flow/logic/structure - but the code will never be anywhere near close to the original source code. I could write the same C++ program in 3 different ways and get identical assembly, how do you go back to the exact source? The answer is that you don't.

Here's the same simple program, written in 3 different ways, producing identical binary compatible code: https://godbolt.org/z/qWrc8fEnn

How does the AI know whether it should produce back the snippet #1, #2 or #3? It does not. It cannot.

That's pre-2026 thinking. At this point, with the ability to lash IDA or similar tools to an agentic harness, there is no longer any such thing as a closed-source binary.

At least that was cut from the game. I remember that episode from The Lone Gunmen, the X-Files spin-off series, where they stopped the (inside job) hijacking of a commercial airplane from crashing into the twin towers, which aired just a few months before the attack.

I only watched it a few years later, and I had to double check the airing date.

> but the actual story is incomprehensible.

You're almost there. Double the convolutions in your post and you'll finally understand Kojima.

And yes, that chick really does have to be topless, its not sexism. See, she breathes through her skin. Thats why she's such a good sniper

>Yeah I know, sure, fake information on the internet. It's so prescient I guess, but the actual story is incomprehensible.

apply this logic to john carpenter movies and they fall apart, too.

Escape from LA is just about as absurd, just less paranormal.

Nuclear powered personal submarines, surfing typhoon waves from California plate shifts, transforming bullet-proof stealth helicopters, emp superweapons and engineered DNA-tailored superbugs ( or atleast the claim of them ) , VR replacing reality, Cuba taking over the democratic world after a theocratic US rescension.

it's just entertainment campiness. When I play a Kojima game I just sort of imagine it internally as a video game equivalent to a John Carpenter movie.

A lot of people focus on the whole dynamic of corleone family and their position in American society in the Godfather. But seriously did you see what vito was eating in the restaurant?? Makes no sense

Kojima always has some paranormal activity in his games. And he does like to subvert expectations about what is technology and what is paranormal

During the mainline of the game you are an anti terrorist specialist sent to rescue the president from a bunch of stereotypical mad bombers

Well, basically, the leader of the terrorists is the 43rd US president, who is a clone of a super soldier despite being ostensibly anti cloning in office. (Yes, he's Dubya). He's planning to detonate a nuclear weapon above Manhattan so an EMP field knocks out all communications. The reason being that the government has been taken over by a conspiracy called the Patriots who are using a sinister AI to filter all internet traffic

However Dubya has been double crossed by an old Russian guy who was infatuated with the super soldier Dubya was cloned from. He claims to be representing the Patriots but suddenly his arm comes to life, claiming to be Dubya's brother. Oh, by the way, he lost his arm and got a transplant from the body of the other clone, Dubya's brother. It's implied in this game to be a supernatural occurrence.

Ghost Dubya 2 (in the body of elderly Spetznaz guy) goes on a rampage and tries to kill everyone. Dubya and the anti terrorist specialist end up fighting on top of Federal Hall, which if you've ever been to Manhattan is just next to the NYSE and TJ Maxx.

When you're there the Patriots phone you and they claim to be dead. Well, not dead exactly. They claim that in the same way life emerged from organic chemicals, life has emerged from the neural net of ideology and content published on the internet. Kind of like if ChatGPT became a bit uppity. They claim that Western Civilization is too corrupt and contradictory to be left in the hands of humans. You have to kill Dubya who is somewhat of an anti hero and it's kind of sad

Eventually another brother of Dubya who is kind of the hero does some digging, and his bromance partner discovers that the Patriots were real people but they all died 100 years ago. Also they were funding the real heroes all along!

It makes a lot more sense in MGS4

My take was more along the lines of: it wouldn't be convincing enough, if anything it would be too clean and perfect.

Does the TP decomp use AI to achieve their speed?

What is the state of the art of compilers here? What size of project are we speaking here?

What is the experience faulty decompilation, and the existence of bugs in the binary?

Could one decompile a binary to a more modern language than C?

I’m interested in how LLMs handle obfuscated code. Throw LLM with IDA MCP at EasyAntiCheat_EOS.sys or the like (as the most common examples of heavily obfuscated software) and see how far they can get.

Here's another interesting one. I think they cut a scene from a movie too.

https://youtu.be/Ozz8uxW733Q?t=81

To be clear, I enjoy the games, I've played through MGS2 all the way through twice. It's a good game.

I even enjoy the story, but I view the story as something like a soap opera; a story that is extremely compelling and addictive as you're watching it, but upon turning it off you find yourself questioning what it is you just watched.

Man, now I have the music of The Duke arriving in my head.

Thanks!

edit: the soundtrack sure was a banger at the time

https://youtube.com/watch?v=eX_t63eZqyk

> but what the hell was with the girl that could repel bullets but it turns out that it was magnets but actually maybe she does have powers?

MGS always rides the line between "it's actually just tecnology" and "no, it was really supernatural." This is true of Psycho Mantis in MGS1 as well, and The Sorrow in MGS3 is literally a ghost. I assume that Kojima has read up on the Stargate Project and other psychic soldier projects and likes to think about "what if one of those bore fruit?" Anyway, in the context of Fortune, it's a dramatic take on that: her powers were actually a magnetic device planted on her by The Patriots, but ACTUALLY actually she really does have some psychic ability which allows her to fight Ocelot in her final moments.

> And what about the weird thing where you start sword fighting your father/president/clone?

Well, that's actually pretty straightforward. You fought Solidus Snake, who was the third clone of Big Boss, and the "main antagonist" of the game (scare quotes because arguably the Patriots are the real antagonists). At the end of the game, it is revealed that Raiden's girlfriend Rosemary is a) real, b) alive, c) pregnant and d) being held captive by the Patriots, and they task Raiden with killing Solidus if he ever wants to see her again. Why a sword fight? I dunno, some things we must simply chalk up to "because it's cool."

If you want to dive further into who Solidus was, yes, he was also a pseudo-father to Raiden, in the sense that Raiden was raised as a child soldier and Solidus was (at the time) his commander. Solidus was President during the events of MGS1 under the name George Sears, but by MGS2 he's leading the Sons of Liberty as part of his anti-Patriots plot. Unfortunately, Ocelot is still working as a Patriots double-agent at this point, and Raiden's involvement is in turn orchestrated by the Patriots, so it doesn't quite work and the whole operation is manipulated from the start.

That stuff is just silly mostly-self-aware camp to keep the setting entertaining, don't overthink it. Everybody already knows and acknowledges that the series is absurd.

I don't think it's impossible, but it would take a lot of time and a lot of money; likely more time than good enough models have been commercially available.

I have been working on an incremental decompilation-based reimplementation (basically how OpenRCT2 was done) of Worms Armageddon for the past 2 months with a lot of help from LLM tools; primarily Claude Code and Ghidra MCP. I've worked on it almost every day, reaching Claude Code Max 5x's 5 hour session limit multiple times every day. Suffice to say as a software rendered, sprite-based 90s PC game, Worms Armageddon is several orders of magnitude simpler than MGS2. Despite that, I think it will be 2-3 more months of work before I can compile a fully independent version of the game.

This is despite the game being an almost ideal candidate for automated RE, as it uses deterministic game logic with built-in checksum checks in replays and multiplayer. I've downloaded all the speedruns I could find for the game (as replay files) and I've retrofitted the replay system into a massively parallel test framework, which simulates over 600 games in about 30 seconds. So Claude can port all game logic independently without much need for manual testing; the replay tests can almost guarantee perfect correctness.

MGS2 doesn't have anything like that, so every ported function requires extensive manual testing. Even with LLM tools, an accurate decomp could take years (unless you're willing spend thousands of $currency per month on it).

Keep your eyes open for Sonic R too. Sadly a lot of the online Sonic community has been toxic to the dev for being transparent about using Claude for the majority of the disassembly. Even though he's a very talented developer with lots of credit to his name, and only took a few weeks compared to a year+ if fully manual.

Decompilation to C (and even C++!) has been done automatically for 2-3 decades at least. I am not sure what has changed in recent years other than people playing fast and loose with copyright (and GitHub allowing it, likely because their LLMs also stand to benefit). Introducing LLMs here is only going to introduce errors, delays and likely push you away from a reliable result.

The challenge here is readability. Reading the TP source leak you link I think it's even behind the current state of the art, as it's barely above assembly. This is where I suspect even the smallest of LLMs may help, since you don't care that much if it introduces errors.

Who cares? Who said anything about recreating the exact code? You will get usable, compilable, and surprisingly readable source code, in your language of choice, that yields the functional equivalent of the binary.

Barring obvious edge cases that could show up but don't usually, like intentional race conditions. Timing is the one area where things get iffy.

We were robbed of the end to that game. I loved it so much and will always wonder what closure I would have been robbed of if they had just completed it.

Dunkey Kingdom Hearts explained https://www.youtube.com/watch?v=8o1ieehttdA

Having followed his bsky during his announcement, he started off per-emptively dissing on his haters that... didn't even exist yet. Constantly posting memes about how everyone was dissing him and how AI was totally superior (and then posting his angry sessions with Claude when it got something wrong) when most other users were just "that's cool man". The thing that made him quit bsky was a (now-deleted) thread someone posted criticizing the weird crash-outs. I think he was more... normal about the whole thing, people would have received the project quite a bit more positively.

Veal doesn’t make sense?

This is really cool! Your process is compelling, and your choice of game is excellent. I'd like to read a long blog post about your entire journey from the beginning to a working binary once you get there.

For those wondering, there is a public Git repository at https://github.com/paavohuhtala/OpenWA.

>Decompilation to C (and even C++!) has been done automatically for 2-3 decades at least.

Only in a very rudimentary sense and definitely not in a working compilation (much less binary equivalent) sense. LLMs have turned this from a gimmick for static analysis into something that actually works pretty well for recompilation projects.

That is quite incredible if that is true. Need to read a bit into that. Can you point towards relevant literature/examples? Also: please see my questions in the comment to your other reply

I am pretty sure I am going to file this away for some copypastas when I want to confuse someone. I’ll just have ChatGPT or whatever swap out some words.

Thank you, my friend.

Holy shit

As it happens I do have the habit of writing very long blog posts - though none on OpenWA so far. The OpenWA readme file serves as a bit of an introduction, though it's already a month old.

Mandated last generation console support gimped the game's scope too... It pains me to imagine what could have been.

...what?

I watched the Dunkey video linked elsewhere and thought he was making it up, he wasn't? The whole story is just random stuff happening?

> Who said anything about recreating the exact code?

The person I'm replying to? Who said you will get the same code as if it were the original source?

I've only completed the first three, but at least in those, the stories get progressively off the rails and weird as they go on, especially MGS2.

It's one of those weird things, because the stories are kind of incomprehensible, but they're compelling in the same way that a soap opera is; while you're watching and playing, the story is very addictive and it doesn't feel that confusing as it's going on. There are even evil twins!

The moment I turn the games off, I kind of start thinking "....what the hell actually happened there?".

They do try and have a moral associated with it (though they do one of my pet peeves and just outright explain the themes they're going for at the end), so it's not completely trivial.

The games are a lot of fun, and they are certainly worth playing through in an emulator or something. I've played through the first MGS three times, and the second one twice, and I thoroughly enjoyed them both times, even rewatching the story unfold.

Just play the games. They're good fun.

> The whole story is just random stuff happening?

Welcome to Metal Gear Solid. Half the fun is creating a coherent narrative out of the sheer number of random postmodern happenings. You should see MGSV. Its entire existence is an attempt to connect the series back to the original excuse plots of the original Metal Gear games from the 80s.

This blog post is a great starting point: https://blog.chrislewis.au/the-unexpected-effectiveness-of-o...

It's not really "random", there is a coherent story (and a moral to that story) underneath all the weirdness and you'll sometimes forget how crazy it's gotten because your current objective tends to make sense (ex: "defuse the bombs").

But all MGS games are basically playable Kojima manifestos with sci-fi trappings. The actual minute-to-minute gameplay for them is really, really good so even if you don't dig the story they're worth playing.

The previous poster is right that MGS4 does clear a lot of this up. But MGS4 also slams more weirdness and callbacks to the previous games on top of it.

> Only in a very rudimentary sense and definitely not in a working compilation (much less binary equivalent) sense.

Working is the easy part; the hard part is getting something that classifies as readable C. LLMs do not really help reach the "working compilation" part but benefit from it.

We are way past "working compilation" when it comes to LLMs. They are already really good at writing readable, compliable code. The big problem with LLMs is making sure the output binary actually does what you wanted it to do. But if you define the goal not merely as instructions in a vague, unspecific human language and rather as recreating a given set of binary instructions after compilation, this big drawback goes away. So in a sense they are better suited for recompilation projects than for developing new applications.

Hmm, interesting, thanks. I played MGS 1 when it came out, multiple times, as I liked it a lot. I don't remember it being so weird, but maybe I didn't pay enough attention to the story.

The first one gets a bit goofy with the evil twin stuff, and the unintentionally funny dialog like "Do you think love can bloom on the battlefield???", but certainly doesn't go nearly as off the rails as the second one.