aizk

If you don't mind me asking -- what do you do for a living where you're just off buying an new car and a house in Japan?

Anime fans permeating throughout on the internet? Impossible. Never heard of such a thing.

Serious question -- how?

See here https://news.ycombinator.com/item?id=50135410

Caveat: I get a budget of free ai tokens at work, as I work for an open source AI inference provider.

My goal is to reverse engineer an entire gamecube game -- I'm pushing an existing community decomp that was stalled to the finish line, solo, with lots of AI help.

This work was directly inspired by https://decomp.dev/ -- which has just shut down a few days ago because of, ironically... people like me who enjoy decompiling things with AI, so I'm still weighing / grappling with the ethics of how to release my work, it's murky waters at best.

I use Opus 5.5 as the orchestrator, as well pi harness because it's somewhat lightweight. But for the really hard functions I don't use deepseek I use opus managing other opus agents.

I'm working to release something soon as proof that I can do a 100% decompilation job, but the main idea is that when decompiling a gamecube game your sources of truths are individual functions or files, so you can build off that primitive.

Each ai gets a worktree, and a shared giant context file that's basically a record of every experiment that has happened. They're assigned tasks to essentially do as much as they possible can for each individual funciton. After about an hour claude uses a script to merge the work that's actually correct (verified by tooling that checks if the c goes to asm) and then there are usually 3 rounds of reviews by independent ais, because there will always be some slight mistakes. The techniques used as well as the mistakes the models make are recorded in the giant ledger. There's a bunch of random one off scripts such as stuff to clean up, and tools to make sure the agents don't accidentally google stuff / delete stuff / use too much space.

Claude steers the ship via ssh on my strong linux tower at home so it keeps my laptop nice and cool and at this point I just check in from time to time, I am basically a manager, and Claude has it's own workers.

Last point -- I don't know any c besides the basics, I've only been "coding" python / ts for a couple of years, but even then with the extreme strength of models the future is managing agentic swarms, so I'm leaning in as much as I can. I really see this as the future of large scale projects. For some people it's a mockery of decompilation, but, my goal is the finish line, not the journey.

As the zoomers say these days "Unstoppable scroll vs unskippable title"

I use Opus 5.5 and swarms of about 50 deepseek models to reverse engineer old videos games. The scale is roughly 600k loc generated from pure asm, simply by knowing what the old compiler is. Truly staggering stuff.

Exactly what the NSA would say.

Awesome find greg, thanks for sharing :)

I think the audio is missing a lot of the hiccups, delays, and quirks that come with humans in meetings but the rest is pretty good!

This may sound funny but I feel games would lose a lot of fun if they were all written in rust and had classes of bugs just not available to them. For better or for worse quirks and bugs in games have shaped how people approach games, and also have given games charm for decades.

Mindboggling that people used to do this by hand.

No that's the theoretical token value cost based on chaining multiple 20x subs.

I think the future of this is -- get your AI agent to investigate this port for anything actually worthwhile, comparing it to the go version, and then take those good ideas and actually maybe bring them into the real compiler.

Maybe. Who knows, it's all changing so fast!

Yes it's (hopefully) not distilled from every single major American provider.

I think this is just the price of admission to being a dev in 2026.

At least for me, the enumerated svg-like icons give it away, but that's merely a prompt fix.

I think the AI writing disclaimer was a decent touch.

I don't work for Pangram, but I do a lot of writing at work. API docs, blogs, code, tweets, linkedin, all that.

If you've tried AI detectors a couple years ago, they're basically in the same position that coding agents were a few years ago, where everyone was skeptical at first, but the tech has gotten a lot better. Give it a shot, it's quite good. They are slightly tuned a bit towards classifying things as AI, but I imagine that's deliberate.

The only thing is that their models are pricey, but, very useful.

I don't necessarily know if I see it this way.

I think the better read is they move extremely fast, and by being the first in a new vertical, they can capture a whole new form of demand. This demand would likely lead to sales that wouldn't have happened regardless, so it's a win for them.

I was chatting about something similar to this with coworkers. When the doordash cli got announced I thought "Wow, that's amazing!" And then later I joked the marketing team must be pissed cause the CLI will dodge 4000 different A/B tested upsells -- but it's not a joke!

Did you just happen to have all those comments saved? Quite the timing.

Tokens all the way down

Well, how do we know there aren't errors in their construction within the lean code? Does it just "not compile" or something, or is it deeper / more fundemental than that.

People had joked a couple years ago "Well if they solve a Millenium problem it's AGI"... Well here we are.

No no, if you've ever spent serious time editing, it's a big deal -- if they pull this off correctly, it will be a game changer. Editing has so much grunt work, where you just have to spending time organizing, collecting files, renaming things, finding missing files... Just tedious project management stuff that is excellent for an AI to do.

I wonder how they'd ever cash out.

I partially owe my career to benchmarking AI with Minecraft, so, I'm going to disagree with the author on this one. Games are a great way to test a model, it's not that deep.

You missed a critical detail. They couldn't defend themselves with closed source American models so they had to investigate with Chinese models that had less restrictions. Great optics for hosting open source models!

Is there an equivalent but for 4090s?