Rendered at 20:57:02 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
CoolestBeans 20 hours ago [-]
> The act of reverse engineering and porting decades old games to modern environments requires a complex, deeply involved set of tasks from multiple disciplines... It's the quintessential deep, skilled knowledge labor profile out there that people understood as a career protected by depth of skill and knowledge acquisition.
Game development, especially this kind of porting, is an extremely technically demanding task. But it is far more constrained and precise in its requirements than the sort of software development that happens outside of the game world. Most developers deal with vague and (importantly for AI) unverifiable requirements that change continuously as a project develops. AI is not as good of a tool for this type of programming because unlike the tasks demonstrated in the article, there is no verification function an AI can cheaply use to validate if it is approaching completion of the task.
In the example in the article, the bots can look at the entire memory state of the game, interact with it to see how the state changes, and then repeat the process on their working copy of the port. So long as the overall behavior of the port is getting closer to the original, the AI can validate its getting closer to its goals. Extremely impressive stuff. But this verification loop exists in a closed world. No equivalent exists for most programming disciplines. No equivalent can exist because shifting the requirements to fulfill the vague business goals is an essential job of a software developer.
So while I am extremely impressed about what AI can accomplish every day, I also have a loud voice in the back of my mind skeptical of the actual economic impact. Because the market keeps demonstrating that these technically simpler but vague requirement tasks have a lot more value than something like game dev, at least monetarily.
visarga 17 hours ago [-]
> But this verification loop exists in a closed world. No equivalent exists for most programming disciplines.
You can do N-version programming with coding agents and you get tests for free. It is easier to implement something than to test the same thing. This approach turns implementing into automated oracle for testing, takes less time (can implement in parallel), and every divergence between versions is either a bug fix or a requirement clarification. In order to make the N-versions more diverse we can use different model providers, programming language or libraries.
shoo 15 hours ago [-]
If the main problem is unclear & changing requirements, having 2 different implementations of a system that attempt to implement the vague / changing / missing requirements doesn't seem immediately helpful -- they'll almost certainly disagree, and then you can force one to match the other or so on, and have two implementations of some arbitrary thing.
But none of this "implementing stuff" is making progress to solving the issue of unclear / changing / vague / missing requirements.
em500 15 hours ago [-]
You don't try to force them to match. You show a few very different implementations that all satisfy the underspecified the requirements. Not in an antagonistic way, but to help clarify and sharpen the requirements. If any implementation would do, then there's no problem.
visarga 4 hours ago [-]
When they diverge you use the agent to classify bug one one of them or spec is unclear. You can both fix bugs and refine the spec iteratively.
halo 14 hours ago [-]
Agreed - I suspect recompilation and decompilation is an unusually good task for models for many reasons.
applfanboysbgon 20 hours ago [-]
> Because the market keeps demonstrating that these technically simpler but vague requirement tasks have a lot more value than something like game dev, at least monetarily.
Game industry generates hundreds of billions in revenue. There is plenty of value in it. The value happens to not be in decompiling 80s/90s games you have no legal right to sell, of course.
georgeecollins 19 hours ago [-]
There is more value in remastering old games than you might think. And there always has been. One of my first gigs in the game industry was producing an Infocom Clasics CD-ROM at Activision. That’s games from the seventies and eighties re-released in the nineties.
Toynbeeidea 6 hours ago [-]
And I was very happy to purchase BOTH volumes, as a long-time Infocom fan! I still have the discs in storage in a box of old cds in my closet. Of course, now I just have the z-machine files on my computer and run them through Gargoyle...
ndiddy 18 hours ago [-]
Was it an issue that Infocom had been shut down? For example for making a version of the Z-Machine for Windows 95 or whatever was current. Did Activision have an archive of all the Infocom development stuff, or were you able to go off of whatever documentation fans had come up with when they were making unofficial Z-Machine interpreters?
georgeecollins 18 hours ago [-]
Activision bought Infocom. We did have an archive of the Infocom servers which was kind of amazing. I remember notes for an adventure game set on the Titanic. There was a design document for Leather Goddesses of Phobos II.
ndiddy 8 hours ago [-]
That's interesting, thanks for sharing!
applfanboysbgon 19 hours ago [-]
If you have the rights to sell them, yes.
20 hours ago [-]
brookst 17 hours ago [-]
Tangentially, I have a very nice laser cutter with terrible software (xtool p3). Their electron app is a nightmare of poor usability and unwanted “value adds”, while basic stuff like defining order or paths to cut is somewhere between opaque and broken. And the process of exporting DXF, importing to their software, positioning, setting per-path power and speed… such a PITA.
It took claude an hour or two to reverse the gcode dialect and camera controls. Now I just tell it what to cut, load the material, and claude sorts out where to place the work, power, speed, variations for corners and stuff, and cuts. If there’s difficulty I pop the work on a scanner and claude diagnoses and adjusts.
It is going to be very hard to control anything that’s in consumer hands, whether for good or bad reasons.
Peanuts99 14 hours ago [-]
I built a pen plotter and my biggest challenge was converting and previewing the GCode for plotter images. Recently though I've build a huge toolbox of tools with Codex using spare usage and it's an absolute game changer. Good use of it for sure.
criemen 13 hours ago [-]
My kettle (Fellow EKG) has an annoying display bug. With a simple prompt, Astra went and downloaded the firmware, disassembled it, interpreted it and confirmed the bug. It also told me that there's a firmware signature check, so I have unfortunately no way of flashing a fixed firmware myself.
I'm not sure I'd be adventurous enough to do that anyways, as I don't know enough about ESP32 microcontrollers to verify what it's doing vs. the chance of bricking my device.
Yet,
I'm a programmer working with porting of video games.
This is absolutely mind-blowing. And at the same time definitely overblown.
This stuff is impressive, but reversing is specifically what these models seems best at and porting new games to weaker hardware is way tougher than this. Not to mention making new games.
gmerc 15 hours ago [-]
Author here: It had no problem creating gameboy and NES ROMs for me https://myroom.isfine.org (order the gameboy from the catalog) was an initial test, I’ve halfway through a demake of dragon strike after stumbling on the fact this was trivial
Agentlien 14 hours ago [-]
That demo is cute and has a lot of cool stuff in it but I cannot even find a catalogue. Assuming it's what it sounds like this seems like more of the same: porting old games to another emulator - this time via ROMs.
The real challenge with porting is, as I mentioned, porting modern games to constrained hardware. Specifically, porting demanding new PC games to current-gen consoles.
gmerc 14 hours ago [-]
Pretty sure porting a PC game into a gameboy ROM, even a 1989 PC game, is exactly that.
Trust me, operated by a skilled individual AI will have zero problem with that. I never make the case for amateurs doing these things.
Catalog is under “shop”.
Kuyawa 17 hours ago [-]
I ported Prince of Persia to Swift for MacOS [1] over the weekend using DeepSeek Harness, spent three hours in total, $3 in tokens, and got a playable and enjoyable game bringing back good ol' memories
It is a faithful port of Jordan Mechner's 1989 original game and the guiding idea was just one sentence: port the game exactly, rewrite the engine completely
I didn't write a single line of code or test, not even the readme. I don't know if it delivered swift, assembly or brainfvck as directed but the result is spectacular and I would never had done it in ten lifetimes
Coding is solved, human programming is dead, apps will cost close to zero and those who don't see the writing on the wall are hopeless
I'm trying to think of the number of projects that consisted of:
The work has already been done. Recreate this to be identical but just in a different programming language.
I'm going to say never.
stubish 11 hours ago [-]
Recreating the IBM PC bios, enabling clone PCs and changing the market completely?
One of the points the article makes is that copying is now essentially free. So if you spend money building something, someone else can copy it and use all the money they didn't spend marketing their version. And why license software anyway, when you can just clone it and even tweak it better towards your business cases? Will you be paying your Adobe subscription or buying the knockoff version of Photoshop?
happymellon 10 hours ago [-]
So your argument is that cloning is such a low bar, that LLMs have solved coding by making innovation or new products not worth anyone's time?
That's not really what AI Boosters mean by "solving coding".
__alexs 13 hours ago [-]
This was literally the first coding job I ever got paid to do.
I would decompile the Google browser extension to extract the hash algorithm they used to sign page rank lookup requests, rewrite it in PHP and ship it to SEOs on weird forums.
They updated the hash every few months and I made enough money to pay for some summer activities as a teen.
Not me, AI was fully aware and used it as reference, the readme explains in detail how the original code, SDLPop and PrinceJS were used to build the fork and exactly in what measure
spider-mario 7 hours ago [-]
That raises the question of why do it anyway, and also diminishes the claim that it ported “the game” (either the abstract idea of it or the original closed-source implementation) and not those codebasese.
wonnage 17 hours ago [-]
Maybe you should try reading the readme it wrote you.
thin_carapace 16 hours ago [-]
human programming sure is dead for translating content A to content B, this is the most appropriate use case for our current modelling of generative ai. translation is one field of many, did something in particular give you confidence that being good at translation implies universal applicability?
Kuyawa 8 hours ago [-]
This is just an example related to the post itself about porting old games. AI exceeds all expectations. Personally I've been using AI for the last six months in many different scenarios and in all it just blows me away. The more I use it the less code I write
Do you have an example of what you think AI is not capable of in programming? I can give it a try
thin_carapace 6 hours ago [-]
regarding translation I totally agree with you, ai excels at complex translative skills eg. art, reverse engineering. programming is another good example in isolation, like a web site or simple application. although, from my experience, excessive generative ai usage dulls the imagination, and distances programmers from their code (eg your admission to not knowing the readme of your project).
i do not have an answer for your question, I may only extend from it tangentially. can you think of a way that generative ai could be offered such that end users ultimately experience a greater (rather than lesser) capacity to understanding their programming output? the loss of this understanding is one of many reasons I don't like to use generative ai.
bobajeff 20 hours ago [-]
Is not just games. I've already seen the beginnings of proprietary software such as Wolfram Engine and Microsoft Word vibed up into alternatives. How long till all of MS Office, Creative Cloud, Houdini get the same treatment? Truly exciting times.
brookst 17 hours ago [-]
Yep. And these companies know it. I’ve been in the room with execs candidly saying their billion-dollar cash cows will be easily replicated by a teenager on a weekend.
The sane ones recognize that sales, service, and support have value. Is that teenager going to ensure 24/7 operation for a multinational?
The less sane ones think “if one teenager can build MS office, our thousands of SWEs can make thousands of, uh, help me out here…”
exolymph 3 hours ago [-]
While it is certainly now quicker and cheaper to do this, it is not clear that maintaining a copycat app as an ongoing concern has much more business value than it ever has. Take Photoshop. Pixelmator existed (acquired by Apple eventually). Photopea is still independent. What else... Sketch? Figma and Canva for the "unicorn startup" category. Procreate serving a different chunk of users. Is there room for many more new entrants in this category to gain significant marketshare? Is the ability to dupe Photoshop the actual blocker?
gmerc 15 hours ago [-]
Word 1:0 for windows makes an excellent high speed wasm layout engine …
le-mark 19 hours ago [-]
Yes then expand to popular saas like jira etc, then photoshop. The potential of this sort of disruption was priced into stocks as “saaspocalypss” about may 2026.
Kuyawa 8 hours ago [-]
Shhh, I have slack, jira and trello in my crosshair
devinprater 14 hours ago [-]
AI has allowed blind tinkerers to add accessibility mods to games from Final Fantasy 1-6 Pixel Remaster to mario Kart Wii and so many others. I never thought I'd see (not literal) the day.
0x000xca0xfe 20 hours ago [-]
Interesting, I did the same with an old RTS and it also worked very well.
But instead of cloning their assets and rewriting their code, a dynamic x86 to WASM JIT recompiler with some web platform support code proved fast enough to run the original binaries. It is even playable on an old Android smartphone in Firefox.
The final web player is ~600KB and only needs the game installer as input, so people still have to buy/possess the game. I think this is a lot fairer for the original devs...
But the approach should work with all of those ~2000s simple 2D/light 3D Windows games, like Anno, C&C etc. that have less than 1GB of assets and no serious multithreading.
Newer ones probably won't fit into a browser tab anymore or would take too long to load.
The biggest problem for the smartphone browser version is extensive use of left+right mouse input of that era which is naturally not a good fit for touchscreens.
ticulatedspline 20 hours ago [-]
I think we're going to start seeing a huge avalanche of remakes and mods of classic games. I've been working on my own modding of a classic game I still play and the experience has been both fascinating and surprisingly fruitful.
Like most things the majority of these attempts will be garbage but some will certainly pan out into great new or extended experiences.
carterschonwald 21 hours ago [-]
reversing into a faithful modern implementation has a very important detail. everything complicated has an unambiguous unique reference semantics. there is no decision making on observable behavior of a program, just what the internal details of replication are
having that precise a grounding is literally what llms thrive in, esp across compactions.
nonamee 11 hours ago [-]
Very interesting article and great job! But does it also provide enough information so that someone else can replicate your results or is that a secret sauce? The article mentions some prompts, tweaks, skills, etc.
this has now become a piece of cake for the agent. The hard part has already been done. The question now is: how familiar is the agent with Forth?
georgeecollins 21 hours ago [-]
I don’t think reverse engineering a commercial game is legal. I don’t know if it is right or wrong, or if anyone can stop it.
stubish 18 hours ago [-]
This is one of the points he is making. It might not be legal, but it is certainly allowed with precedents being set in courts and policies being turned into law by politicians. He seems quite explicit in stating that what he is doing now with multiple 90s era games will be what happens with this years games next year, and IP protections are getting eroded at such a rate that even the big budget games get cloned with impunity. One of the quotes is about Rockstar's defense being a brand fortress and money.
Gonxa6282 19 hours ago [-]
Up to this point, it has never been actively stopped. Except for Take Two with the PS2 GTA games, but that was a cease & desist, so it's not determined as "illegal" either.
I think that it's legal as long as you're not decrypting stuff, but my understanding may be limited.
apatheticonion 11 hours ago [-]
Is that one of those "I know I can't win in court but you can't afford a legal defence" type of things?
jagged-chisel 20 hours ago [-]
Might be against the contract, but that’s not illegal.
georgeecollins 19 hours ago [-]
Something tells me you are not a lawyer who specializes in IP law.
Joker_vD 16 hours ago [-]
Which IP law? The EU has them somewhat different from the US, you know.
...The unauthorised reproduction, translation, adaptation or transformation of the form of
the code in which a copy of a computer program has been made available constitutes an
infringement of the exclusive rights of the author. Nevertheless, circumstances may exist
when such a reproduction of the code and translation of its form are indispensable to obtain
the necessary information to achieve the interoperability of an independently created program
with other programs. It has therefore to be considered that, in these limited circumstances
only, performance of the acts of reproduction and translation by or on behalf of a person
having a right to use a copy of the program is legitimate and compatible with fair practice
and must therefore be deemed not to require the authorisation of the rightholder. An objective
of this exception is to make it possible to connect all components of a computer system,
including those of different manufacturers, so that they can work together. Such an exception
to the author's exclusive rights may not be used in a way which prejudices the legitimate
interests of the rightholder or which conflicts with a normal exploitation of the program.
...Protection of computer programs under copyright laws should be without prejudice to the
application, in appropriate cases, of other forms of protection. However, any contractual
provisions contrary to the provisions of this Directive laid down in respect of decompilation
or to the exceptions provided for by this Directive with regard to the making of a back-up
copy or to observation, study or testing of the functioning of a program should be null and void.
...The person having a right to use a copy of a computer program shall be entitled, without
the authorisation of the rightholder, to observe, study or test the functioning of the
program in order to determine the ideas and principles which underlie any element of the
program if he does so while performing any of the acts of loading, displaying, running,
transmitting or storing the program which he is entitled to do.
...In the absence of specific contractual provisions, the acts referred to in points (a) and (b)
of Article 4(1) [reproduction and alteration of the computer program — JvD] shall not require
authorisation by the rightholder where they are necessary for the use of the computer program by
the lawful acquirer in accordance with its intended purpose, including for error correction.
...The authorisation of the rightholder shall not be required where reproduction of the code and
translation of its form within the meaning of points (a) and (b) of Article 4(1) are indispensable
to obtain the information necessary to achieve the interoperability of an independently created
computer program with other programs, provided that the following conditions are met...
It's a very narrow exception, but it exists, and the the laws of the EU members follow this directive.
AdeptusAquinas 19 hours ago [-]
If they can argue its copyright infringement, then you can be sued for it. Sometimes you can get away with just a rebuild of the engine without the assets (do not try to copy the assets!) but even then its up to the company's lawyers if they want to push it. Certainly if you try to sell your remake things get spiky. Can look at all the legal disclaimers on things like OpenSAGE as an example. But plenty of companies are cool, as long as they see no competition to their own copyright: https://en.wikipedia.org/wiki/List_of_game_engine_recreation...
gmerc 13 hours ago [-]
OpenAI and anthropic make the case that if the model decided to reverse engineer the game itself after getting prompt injected, I mean inspired, “GPU did it” is a perfectly valid defense and we should invite the owner of the GPU to write future regulation
AdeptusAquinas 13 hours ago [-]
Seems fair. If you have a thousand dollars and try to clone a million dollar game, you're sued so hard your grandchildren need lawyers. If you have a billion dollars and try to clone a million dollar game, they call you a genius and invite you to the white house (for what that's even worth any more).
owebmaster 9 hours ago [-]
Don't delude yourself thinking you can use their same argument.
georgeecollins 17 hours ago [-]
There is also trademark law to consider.
jnaina 16 hours ago [-]
am going to try vibe-build an Apple tvOS version of Jeff Minter's T2K, complete with the Atari Jaguar blitter raster special effects
jnaina 7 hours ago [-]
damn. sonnet 5.5 pulled it off within 20 mins and sideload it to my apple tv via xcode. fluffy 60 fps to boot.
Maybe I missed this but to me the economic implications are obvious: net loss for game companies, neutral for ai companies, win for gamers probably
keeda 3 days ago [-]
Cross posting from another thread: 17 games fully or partially reversed and ported to the web by one very capable dude in a few weeks with frontier and open weight models. He also did this thing where he “spliced” different game concepts and turned WoW into an FPS: https://news.ycombinator.com/item?id=49652089
In case you mistake him for an “AI booster”, you should see this guy’s other posts, he is extremely critical of the AI frontier labs and Tech Bros and this whole AI bubble.
But he is also a very experienced game developer, and under his control the latest models have pulled off these crazy feats. What TFA is the analysis of a domain expert and what it means for the industry.
zahlman 23 hours ago [-]
Just so you know, the other submission you linked here got killed (i.e. is not visible without showdead on).
orangedog 23 hours ago [-]
Why was it killed? That sounds very interesting.
Does anybody have a link to it? I can't find it on the blog.
They're all [dead] (you'll need "showdead" on in your profile).
keeda 22 hours ago [-]
Whoa TIL. Too much low-quality stuff being cross-posted from LinkedIn?
Jtsummers 22 hours ago [-]
Lots of would-be business influencer types post (or posted? I've never paid much attention to it) on there. There are a number of other sites that are also auto-killed by the system. Sometimes those submissions will come back to life, but it has to be found by a mod or vouched for by enough users.
Why do AI boosters always tell blatantly obvious lies? I guess the answer is because lying is rewarded, but then the question is: why does our society reward lying so heavily? Why does HN? You would think HN draws logically-focused, high-information users who enjoy facts. You would think.
Let's take this claim:
> the state of the art on 3d asset generation already changed dramatically. GPT-6 Astra, GPT-6 Sol and finally Claude Opus 5.5 all moved closer into "Senior 3d artist territory"
I opened Codex, gave Astra Max access to Blender and a model, shaders, materials from Persona 5 and confirmed that it could work with it. Then I gave it reference artwork of an unrelated character and prompted it to use the Persona 5 model as a reference and model the new character in that style. The result: https://d.uguu.se/BOuGsmyR.png
Yes, it is nifty a language model can do anything at all with that prompt. But come now. Senior 3D artist. Give me a fucking break. Lie after lie after lie. Do you enjoy the feeling of clout you get from lying and people sharing your fabrications in a panic? Is that it?
I wish we could talk rationally about the capabilities of tools. Public discourse has degraded beyond reason. As one of the apparently extremely tiny minority of people who are grounded in reality, who objectively evaluates claims with their own two eyes, it is not very enjoyable to be bombarded with conversations from and between liars day in and day out. What is the purpose of this liars' society we have built?
gmerc 15 hours ago [-]
Author here. I’m as far from an AI booster as it gets (look me up) and I worked in the industry for almost 2 decades. I know how assets look like, especially from outsourcers and internal outsourced studios on titles like Assassins Creed.
I’m telling you Astra+ is massively improved under proper guidance and it’s going to be good enough for 80% of games. Maybe not your GTA, Kojima or Call of Duty but that claim was never made. We have the pipelines of tools to deal with mass produced assets. Especially if you pass reference assets not just pictures which all game studios have aplenty.
There’s no way to assume the technology, given training data, will not solve this completely either.
Have a look at myroom.isfine.org - an 1 day experiment with one of said pipelines.
anonymous908213 13 hours ago [-]
> know how assets look like, especially from outsourcers
"I work for AAA studios that are routinely mocked for how low effort their productions are, and I know how cheap we can go" is a very different claim from "senior 3D artist". I don't deny you'll be able to cheapen out more than ever before and produce even sloppier games.
> Electronic Arts and Ubisoft and led game industry aligned teams at Facebook/Meta
It's amazing you can boast about your resumé, it's a who's who of deplorables known for shameless moneygrubbing. No wonder you're so enthusiastic at the prospect of replacing quality art with the cheapest slop imaginable.
> I’m telling you Astra+ is massively improved under proper guidance and it’s going to be good enough for 80% of games.
You think that model I generated is good enough for 80% of games (note also this is walking back your claim of senior 3D artist)? Let's not be liars. If even a junior artist produced that model, they would be let go immediately. I did also pass a reference asset, you know. If I'm wrong, show me. Not a room full of textured boxes that might as well be a mixup of existing asset packs - show me an actual novel human character model generated from concept artwork references. Your words have zero credibility without proof.
> There’s no way to assume the technology, given training data, will not solve this completely either.
This is a religious belief, nothing more. I have been hearing this every day for the past several years. "Trust me bro, just three more months and then it'll be AGI!"
gmerc 12 hours ago [-]
You know, why don’t you actually try it yourself? It’s not my job to convince you of anything - if anything the stations of my career should give you a very uncomfortable feeling that I may actually know what I’m talking about. And I suspect that’s the very reason for your reaction, it’s really creating some serious pressure on your cognitive dissonance.
glimshe 11 hours ago [-]
Never mind that troll.
Given your work described on this great article, how do you see a path where the LLM produces a game replication that a human could directly understand and manually improve - even if arguably unnecessary?
Take a game like Wasteland, which has a terrible modern port. I'd be interested in creating a modern port as a hobby while manually making improvements myself. But that would require code that is reasonably modular which a human can understand. Is this possible using your method and current models?
gmerc 9 hours ago [-]
I think it’s possible, yes. RPGs are tough because they require a lot of detail but for example Ultima 6 “modernized” to Duke Nukem style 3d https://u6.tooclever.org/map3d or Syndicate or Dungeon Keeper all were no problem to do with modern models
glimshe 4 hours ago [-]
What do you mean by "because they require a lot of detail"? I assumed that most classic RPGs are an engine reading tiles and some intermediate scripting/config language for the behaviors.
anonymous908213 12 hours ago [-]
> why don’t you actually try it yourself?
I literally did and posted a link to the model. Honestly, fuck off.
This is exactly what I mean. A society of liars. Backed into a corner so all you can do is bring out another lie, pretending I didn't already do this in the conversation. If I try to bring reality into anything, you will deny what's literally right in front of you and continue living in a fantasy world of your own making, detached from any concept of proof or evidence. People like you have completely destroyed the value of public discourse. Congratulations.
gmerc 12 hours ago [-]
I think my work speaks for itself.
12 hours ago [-]
brookst 17 hours ago [-]
> Why do AI boosters always tell blatantly obvious lies
While he's never been dismissive of the technology itself, it is fascinating to see this kind of work from him even as he rails at all the hype around AI.
simianwords 14 hours ago [-]
> While he's never been dismissive of the technology itself
Hmm,
> Remember the relentless “AI is beating everyone in Math Olympiads” and “it’s all over for humans, no point in learning STEM? Well…
> The 2025 Math Olympiad problems are out and naturally people tested the leading LLMs on those… and… spoiler .. they all fail, badly.
>That’s it LLMs are all about storage in training data which enables recombinatorial retrieval. Don’t expect them to solve or respond to new things like a human.
> They fail with novel and unknown tasks because their ability to generalize is vastly overstated by the people invested in the technology.
> The value of LLMs is in their data. No data, no answer. You have a novel piece of work that’s never been done before? OpenAI will insist it’s worth nothing because it’s a spec in the Ocean, yet that’s not true at all.
Its up to you to verify that he has dismissed the technology and that he has been wrong about its performance in mathematics.
Edit:
Even without lean I can solve IMO problems using Sol/Fable today.
You made a categorical statement about LLMs
> That’s it LLMs are all about storage in training data which enables recombinatorial retrieval. Don’t expect them to solve or respond to new things like a human
> They fail with novel and unknown tasks because their ability to generalize is vastly overstated by the people invested in the technology.
It was not contextual but categorical. And it is false today.
gmerc 13 hours ago [-]
Math Olympiads benchmarks at the time were reliant on training data stuffing / Benchmaxxing and this was the absolutely correct statement at that time in that particular context (Dramatic drop of performance on out of training set questions).
We hadn’t broadly moved into lean proof harnesses and brute force compute spend at the time yet and the statement itself continues to be true for LLMs on their own.
AdeptusAquinas 19 hours ago [-]
He mentions GTA a lot in this post. I feel if he reverse engineered and rebuilt GTA, then attempted to sell the result, that would be a fairly interesting case study in all the ways porting a game from the 80s to JS isn't quite the fait accompli for AI he thinks it is.
stubish 18 hours ago [-]
He is mentioning GTA a lot because the article is a warning to the big players. Today, you can repeatably clone games from 90s. Next year, Rockstar's moat against clones might be down to branding, lawyers and money. And the way the IP laws are going, lawyers might not be much help. And money might not help either if the clone is made by a behemoth like 10cent or Meta. Maybe Meta wants the Fortnight ecosystem but doesn't want to pay for Epic. Maybe 10cent wants a share of that GTA6 money and tweaks a GTA6 clone for the Asian market.
Game development, especially this kind of porting, is an extremely technically demanding task. But it is far more constrained and precise in its requirements than the sort of software development that happens outside of the game world. Most developers deal with vague and (importantly for AI) unverifiable requirements that change continuously as a project develops. AI is not as good of a tool for this type of programming because unlike the tasks demonstrated in the article, there is no verification function an AI can cheaply use to validate if it is approaching completion of the task.
In the example in the article, the bots can look at the entire memory state of the game, interact with it to see how the state changes, and then repeat the process on their working copy of the port. So long as the overall behavior of the port is getting closer to the original, the AI can validate its getting closer to its goals. Extremely impressive stuff. But this verification loop exists in a closed world. No equivalent exists for most programming disciplines. No equivalent can exist because shifting the requirements to fulfill the vague business goals is an essential job of a software developer.
So while I am extremely impressed about what AI can accomplish every day, I also have a loud voice in the back of my mind skeptical of the actual economic impact. Because the market keeps demonstrating that these technically simpler but vague requirement tasks have a lot more value than something like game dev, at least monetarily.
You can do N-version programming with coding agents and you get tests for free. It is easier to implement something than to test the same thing. This approach turns implementing into automated oracle for testing, takes less time (can implement in parallel), and every divergence between versions is either a bug fix or a requirement clarification. In order to make the N-versions more diverse we can use different model providers, programming language or libraries.
But none of this "implementing stuff" is making progress to solving the issue of unclear / changing / vague / missing requirements.
Game industry generates hundreds of billions in revenue. There is plenty of value in it. The value happens to not be in decompiling 80s/90s games you have no legal right to sell, of course.
It took claude an hour or two to reverse the gcode dialect and camera controls. Now I just tell it what to cut, load the material, and claude sorts out where to place the work, power, speed, variations for corners and stuff, and cuts. If there’s difficulty I pop the work on a scanner and claude diagnoses and adjusts.
It is going to be very hard to control anything that’s in consumer hands, whether for good or bad reasons.
https://x.com/mitchellhynes/status/2104658015605379094
This is absolutely mind-blowing. And at the same time definitely overblown.
This stuff is impressive, but reversing is specifically what these models seems best at and porting new games to weaker hardware is way tougher than this. Not to mention making new games.
The real challenge with porting is, as I mentioned, porting modern games to constrained hardware. Specifically, porting demanding new PC games to current-gen consoles.
Trust me, operated by a skilled individual AI will have zero problem with that. I never make the case for amateurs doing these things.
Catalog is under “shop”.
It is a faithful port of Jordan Mechner's 1989 original game and the guiding idea was just one sentence: port the game exactly, rewrite the engine completely
I didn't write a single line of code or test, not even the readme. I don't know if it delivered swift, assembly or brainfvck as directed but the result is spectacular and I would never had done it in ten lifetimes
Coding is solved, human programming is dead, apps will cost close to zero and those who don't see the writing on the wall are hopeless
[1] https://github.com/kuyawa/prince
The work has already been done. Recreate this to be identical but just in a different programming language.
I'm going to say never.
One of the points the article makes is that copying is now essentially free. So if you spend money building something, someone else can copy it and use all the money they didn't spend marketing their version. And why license software anyway, when you can just clone it and even tweak it better towards your business cases? Will you be paying your Adobe subscription or buying the knockoff version of Photoshop?
That's not really what AI Boosters mean by "solving coding".
I would decompile the Google browser extension to extract the hash algorithm they used to sign page rank lookup requests, rewrite it in PHP and ship it to SEOs on weird forums.
They updated the hash every few months and I made enough money to pay for some summer activities as a teen.
Do you have an example of what you think AI is not capable of in programming? I can give it a try
i do not have an answer for your question, I may only extend from it tangentially. can you think of a way that generative ai could be offered such that end users ultimately experience a greater (rather than lesser) capacity to understanding their programming output? the loss of this understanding is one of many reasons I don't like to use generative ai.
The sane ones recognize that sales, service, and support have value. Is that teenager going to ensure 24/7 operation for a multinational?
The less sane ones think “if one teenager can build MS office, our thousands of SWEs can make thousands of, uh, help me out here…”
But instead of cloning their assets and rewriting their code, a dynamic x86 to WASM JIT recompiler with some web platform support code proved fast enough to run the original binaries. It is even playable on an old Android smartphone in Firefox.
The final web player is ~600KB and only needs the game installer as input, so people still have to buy/possess the game. I think this is a lot fairer for the original devs...
But the approach should work with all of those ~2000s simple 2D/light 3D Windows games, like Anno, C&C etc. that have less than 1GB of assets and no serious multithreading. Newer ones probably won't fit into a browser tab anymore or would take too long to load.
The biggest problem for the smartphone browser version is extensive use of left+right mouse input of that era which is naturally not a good fit for touchscreens.
Like most things the majority of these attempts will be garbage but some will certainly pan out into great new or extended experiences.
having that precise a grounding is literally what llms thrive in, esp across compactions.
https://github.com/s-macke/starflight-reverse
this has now become a piece of cake for the agent. The hard part has already been done. The question now is: how familiar is the agent with Forth?
https://n.uguu.se/yAvgVLPw.mp4
In case you mistake him for an “AI booster”, you should see this guy’s other posts, he is extremely critical of the AI frontier labs and Tech Bros and this whole AI bubble.
But he is also a very experienced game developer, and under his control the latest models have pulled off these crazy feats. What TFA is the analysis of a domain expert and what it means for the industry.
Does anybody have a link to it? I can't find it on the blog.
https://news.ycombinator.com/from?site=linkedin.com
They're all [dead] (you'll need "showdead" on in your profile).
It works for me in an Incognito window, but YMMV.
Let's take this claim:
> the state of the art on 3d asset generation already changed dramatically. GPT-6 Astra, GPT-6 Sol and finally Claude Opus 5.5 all moved closer into "Senior 3d artist territory"
I opened Codex, gave Astra Max access to Blender and a model, shaders, materials from Persona 5 and confirmed that it could work with it. Then I gave it reference artwork of an unrelated character and prompted it to use the Persona 5 model as a reference and model the new character in that style. The result: https://d.uguu.se/BOuGsmyR.png
Yes, it is nifty a language model can do anything at all with that prompt. But come now. Senior 3D artist. Give me a fucking break. Lie after lie after lie. Do you enjoy the feeling of clout you get from lying and people sharing your fabrications in a panic? Is that it?
I wish we could talk rationally about the capabilities of tools. Public discourse has degraded beyond reason. As one of the apparently extremely tiny minority of people who are grounded in reality, who objectively evaluates claims with their own two eyes, it is not very enjoyable to be bombarded with conversations from and between liars day in and day out. What is the purpose of this liars' society we have built?
I’m telling you Astra+ is massively improved under proper guidance and it’s going to be good enough for 80% of games. Maybe not your GTA, Kojima or Call of Duty but that claim was never made. We have the pipelines of tools to deal with mass produced assets. Especially if you pass reference assets not just pictures which all game studios have aplenty.
There’s no way to assume the technology, given training data, will not solve this completely either.
Have a look at myroom.isfine.org - an 1 day experiment with one of said pipelines.
"I work for AAA studios that are routinely mocked for how low effort their productions are, and I know how cheap we can go" is a very different claim from "senior 3D artist". I don't deny you'll be able to cheapen out more than ever before and produce even sloppier games.
> Electronic Arts and Ubisoft and led game industry aligned teams at Facebook/Meta
It's amazing you can boast about your resumé, it's a who's who of deplorables known for shameless moneygrubbing. No wonder you're so enthusiastic at the prospect of replacing quality art with the cheapest slop imaginable.
> I’m telling you Astra+ is massively improved under proper guidance and it’s going to be good enough for 80% of games.
You think that model I generated is good enough for 80% of games (note also this is walking back your claim of senior 3D artist)? Let's not be liars. If even a junior artist produced that model, they would be let go immediately. I did also pass a reference asset, you know. If I'm wrong, show me. Not a room full of textured boxes that might as well be a mixup of existing asset packs - show me an actual novel human character model generated from concept artwork references. Your words have zero credibility without proof.
> There’s no way to assume the technology, given training data, will not solve this completely either.
This is a religious belief, nothing more. I have been hearing this every day for the past several years. "Trust me bro, just three more months and then it'll be AGI!"
Given your work described on this great article, how do you see a path where the LLM produces a game replication that a human could directly understand and manually improve - even if arguably unnecessary?
Take a game like Wasteland, which has a terrible modern port. I'd be interested in creating a modern port as a hobby while manually making improvements myself. But that would require code that is reasonably modular which a human can understand. Is this possible using your method and current models?
I literally did and posted a link to the model. Honestly, fuck off.
This is exactly what I mean. A society of liars. Backed into a corner so all you can do is bring out another lie, pretending I didn't already do this in the conversation. If I try to bring reality into anything, you will deny what's literally right in front of you and continue living in a fantasy world of your own making, detached from any concept of proof or evidence. People like you have completely destroyed the value of public discourse. Congratulations.
> Public discourse has degraded beyond reason
I rest your case.
While he's never been dismissive of the technology itself, it is fascinating to see this kind of work from him even as he rails at all the hype around AI.
Hmm,
> Remember the relentless “AI is beating everyone in Math Olympiads” and “it’s all over for humans, no point in learning STEM? Well…
> The 2025 Math Olympiad problems are out and naturally people tested the leading LLMs on those… and… spoiler .. they all fail, badly.
>That’s it LLMs are all about storage in training data which enables recombinatorial retrieval. Don’t expect them to solve or respond to new things like a human.
> They fail with novel and unknown tasks because their ability to generalize is vastly overstated by the people invested in the technology.
> The value of LLMs is in their data. No data, no answer. You have a novel piece of work that’s never been done before? OpenAI will insist it’s worth nothing because it’s a spec in the Ocean, yet that’s not true at all.
https://www.linkedin.com/posts/georgzoeller_proof-or-bluff-e...
Today you can solve IMO problems with ~$5.
Its up to you to verify that he has dismissed the technology and that he has been wrong about its performance in mathematics.
Edit:
Even without lean I can solve IMO problems using Sol/Fable today.
You made a categorical statement about LLMs
> That’s it LLMs are all about storage in training data which enables recombinatorial retrieval. Don’t expect them to solve or respond to new things like a human
> They fail with novel and unknown tasks because their ability to generalize is vastly overstated by the people invested in the technology.
It was not contextual but categorical. And it is false today.
We hadn’t broadly moved into lean proof harnesses and brute force compute spend at the time yet and the statement itself continues to be true for LLMs on their own.