frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

500B Tokens Later: Letting AI Agents Decompile a First-Person Shooter

https://momo5502.com/posts/2026-10-09-game-decompilation/
24•davikr•1h ago

Comments

WillAdams•34m ago
To save folks looking this up:

>At current 2026 API rates, 500 billion AI tokens would cost roughly $100,000–$750,000 depending on the model, with most flagship models in the $150–$400 per million input tokens range

Gigachad•27m ago
This stuff is all done on highly subsidized subscription plans. I suspect Antropic is quite happy to sell these people $100k of compute for $4k because it boosts their growth numbers and they can tell investors once they stop subsidizing, this will grow to $100k. Despite the fact that most of this stuff simply wouldn't be done without the subsidization.
chii•8m ago
the exact same arguments were said about uber's business at the start.

Yet, it is now profitable.

The bet is that people realize how valuable these services are, and despite complaining, they still would pay the higher price. This realization would not happen without this initial subsidy from investors.

It isn't too different from drug dealer's first sample free...

thway15269037•5m ago
Uber business model wasn't a subscription "for 10 bucks you can travel 900 lightyears a week"
j2kun•26m ago
I struggle to believe that someone would find it worth that much money to have a decompiled version of a game they also commit to keeping private.
girvo•19m ago
> they also commit to keeping private

Reading between the lines:

"The avid reader of my blog might have noticed that I had previously written two posts that have since been removed. Everyone else might now be wondering which game I am talking about. To both of you I can only say that corporate America was here to ruin our fun."

Keeping it private likely wasn't the original plan...

TomatoCo•25m ago
Where are you getting 150-400 per million input?

https://developers.openai.com/api/docs/pricing https://platform.claude.com/docs/en/about-claude/pricing

OpenAI and Mythos are both 10/mil in.

https://openrouter.ai/z-ai/glm-5.3#providers https://openrouter.ai/moonshotai/kimi-k3#providers Other frontier models are like 1-3/mil in.

Also, 500 billion is 500,000 millions. At the lower end of your 140/mil estimate that's 70 million dollars. Even at my 2/mil lookup for Chinese frontier models that's one million. Show your math for 100k-750k, please.

GaggiX•7m ago
The vast majority of these tokens are cache input tokens so even at API price the cost would be much lower.
brandonpelfrey•31m ago
Unless you explicitly need byte-matching decompilation, there are significantly faster ways to produce a decompilation/C which is functionally equivalent. I need to post about this. What's been working for me is that for every function, Agent A is tasked with writing some code which is semantically equivalent to the original assembly, but not necessarily exactly the same. Agent A also writes tests. Agent A submits the implementation of the function and tests to the harness for it to judge. The harness runs both the original function and the submitted function in a virtual machine/simulator/emulator (the tests define function inputs and starting state). The harness will only accept the implementation if 1) the read/write sequence to RAM is identical to the original function's, and 2) there must be complete line and branch coverage of the original function being decompiled.

I've found this to be robust for decompiling games, while giving the agents enough freedom to write code that is readable and not waste a ton of time making sure e.g. instruction ordering, register assignments, etc. are all exactly the same. For me, having byte-matching decompilation is only one way to produce a decompilation I know is faithful to the original. This "high-level decompilation" process I just described is something agents can do much more quickly.

j2kun•29m ago
Functional equivalence here, of course, depends on the completeness of the test suite, where byte-identical compiled artifacts does not.

(For example, your approach would not necessarily catch all the same overflow behaviors; the OP expressly claimed that "replicating all bugs" was also important, and many bugs are caused by certain overflow behaviors)

SubiculumCode•26m ago
Byte exact seems only of interest to preserve known bugs etc for cheats/shortcuts/etc.
j2kun•23m ago
thway15269037•8m ago
I struggle to understand what legal leverage they used to threaten him to remove every detail about the game. Can someone post the game name and company name?

So, if you reverse-engineer game X and post reverse-engineered code, what exactly do you infringe, how and in which jurisdiction? What changes if it is done via LLM?

(I understand that LLM decompilation is absolutely out of hand right now and something surely will come to trample the fun. But what and when? I suppose american LLMs will have their system prompt updated to forbid any reversing help and report suspicious activity straight to legal hotline)

georgemcbay•7m ago
In case anyone is curious about the obvious question of which game they are talking about... based on months-old reddit posts, the game in question appears to be Call of Duty: Modern Warfare 2.

https://www.reddit.com/r/ReverseEngineering/comments/1vxig19...

Presumably at some point Microsoft sent a C&D.

xnx•5m ago
A previous version of the page said the game was Call of Duty: Modern Warfare 2 (2009).
That may be true, but I hate it when people repeat the false idea that functional equivalence requires only a test suite that has full branch/line coverage. Call me triggered :)

That said, I would probably follow this same approach if I were to do this, but with extensive randomized testing as well.

SubiculumCode•27m ago
So you restricted it to implementing the same function (same inputs,outputs, dependencies as original?) and prevented the agents from making design decisions by keeping it's scope restricted?

We Prevented the Biggest Hack in Crypto History

https://verialabs.com/blog/biggest-hack-in-crypto-history/
2•bryce2010va•6m ago•0 comments

Show HN: Color Swapper – Recolor a terminal screenshot to find the right colors

https://colorswapper.com/
2•_andrei_•11m ago•0 comments

A new AI model that renders generative visuals of what you see

https://www.weizmann.ac.il/pages/news/mathematics-and-computer-science/the-new-science-of-mind-re...
4•inshard•14m ago•0 comments

Terrence Tao does not understand cancer medicine

https://www.drvinayprasad.com/p/terrence-tao-does-not-understand
2•elo2000•21m ago•0 comments

Decreasing Certificate Lifetimes to 45 Days

https://letsencrypt.org/2025/12/02/from-90-to-45
2•thread_id•21m ago•0 comments

Rogue Anthropic AI agent gave police fake tip in unsolved murder case

https://www.bbc.com/news/articles/cqkg50j1yd5lo
2•rmason•22m ago•1 comments

A single-card GitHub profile stats generator built with Next.js

https://github.com/alirezamoradiii/github-stats-card
1•alirezamoradi•26m ago•0 comments

Discord Just Couldn't Resist... [video]

https://www.youtube.com/watch?v=kMg9WEkyXsk
4•Bender•26m ago•1 comments

Ask HN: Do you want a router that has DD-WRT pre-installed?

2•Ixander•26m ago•1 comments

GTA 5 is back on the web

https://github.com/alihammoud21/playgta5
5•jdudidu•29m ago•0 comments

Maextro V800: brake pedals fail; videos pulled, no recall, astroturfing alleged

https://chinaevhome.com/2026/10/09/huaweibacked-maextro-v800-braketest-videos-pulled-from-platforms/
4•linzhangrun•34m ago•0 comments

Barnaby: Self-hosted Matrix agent that reads the room

https://github.com/pkulak/barnaby-home
1•pkulak•35m ago•1 comments

Giving Muse Access to My iMessages

https://sidhanthp.pages.dev/writing/imessage-muse/
3•sidhanthp•37m ago•0 comments

The Ticket Is Not the Work

http://agileotter.blogspot.com/2026/09/the-ticket-is-not-work.html
3•trashymctrash•45m ago•0 comments

Jshookmcp – instrument your real signed-in Chrome via CDP

https://github.com/vmoranv/jshookmcp
3•vmoranv•50m ago•0 comments

Brain does not process information and it is not a computer (2016)

https://aeon.co/essays/your-brain-does-not-process-information-and-it-is-not-a-computer
6•Anon84•53m ago•2 comments

All startups need is distribution

6•relaxfolio•56m ago•0 comments

500B Tokens Later: Letting AI Agents Decompile a First-Person Shooter

https://momo5502.com/posts/2026-10-09-game-decompilation/
26•davikr•1h ago•18 comments

Electrically driven rotation of polar spirals in a ferroelectric polymer (2024)

https://www.nature.com/articles/s41467-023-44395-5
1•peter_d_sherman•1h ago•0 comments

Give up kissing or sauce? I asked 10'000 people in 127 countries

https://huggingface.co/datasets/jasoncorkill/would-you-rather-kissing-vs-sauce
3•data_daddy•1h ago•1 comments

US Justice Department investigates TV networks for boycotting Trump press pool

https://www.bbc.com/news/articles/c9vgregzk6gro
8•intunderflow•1h ago•0 comments

AliasVault is an end-to-end encrypted password and (email) alias manager

https://www.aliasvault.com
3•doener•1h ago•0 comments

European Alternatives

https://buy-european.net/en/alternatives
3•doener•1h ago•0 comments

A plastic once thought harmless may be fueling fatty liver disease

https://www.sciencedaily.com/releases/2026/10/261008102718.htm
6•gradus_ad•1h ago•0 comments

Show HN: I wrote Dependency Hell 5 years ago. Today I updated it for AI Era [pdf]

https://github.com/jaronilan/stories/blob/main/Agent%20Loop.pdf
3•jaronilan•1h ago•1 comments

Soliton

https://en.wikipedia.org/wiki/Soliton
6•kaycebasques•1h ago•0 comments

The True Scale of the Original Trilogy [video]

https://www.youtube.com/watch?v=bvyhmpmxBWc
4•sjav•1h ago•1 comments

Show HN: An unofficial download picker for the ArtCraft Rust apps

https://craftsuite.net/
4•songtianlun1•1h ago•0 comments

GMKtec EVO-X3 Mini-PC Review Looking at a 128GB AMD Local AI Box

https://www.servethehome.com/gmktec-evo-x3-mini-pc-review/
4•lumpa•1h ago•0 comments

Ask HN: Resources for best practices in software development with AI?

6•sergio-gdr•1h ago•0 comments