211–220 of 404 posts
love the article slightly concerned tomorrow morning's top HN story will be karparthy telling us how dog-based LLM interfaces are the way of the future and you'll be left behind if you don't get in now (and then next week my boss will be demanding I do it)
I think we can all agree cat LLMs are the way of the future though.
> I experimented with Rust/Bevy and Unity before settling on Godot. Bevy’s animations and visuals weren’t as crisp, and Claude struggled with its coordinate conventions - likely a combination of less training data and Bevy leaving many core features, like physics, to the community. Unity was a constant struggle to keep the MCP bridge between Claude and the editor healthy. It frequently hung, and I never figured out how to get Claude Code to read the scene hierarchy from the editor. Godot’s text-based scene format turned out to be a huge advantage - Claude can read and edit .tscn files directly.
Didn't expect Godot to be the most friendly game engine for LLM usage! I think it's because of various factors - Godot has been used quite a lot in recent years so there are various code examples on the Internet, and its scene file format (.tscn) is very concise enough for LLMs to write and edit directly (Unity has its own YAML-based format but it's very unfriendly for human consumption, and Unreal stores its core assets in binary files)
Pretty sure he would have gotten very similar output just by saying "generate a random game using Godot and c#" but that wouldn't make for a viral post so instead he asked the model to pretend meaningless input is being used by it and added a dog in the process of writing such input because that helps the virality of the whole thing.
I think you would have gotten more generic games. The AI was clearly attempting to find meaning in what the dog typed, and that drove what it made. Now, if Anthropic let you adjust the temperature, then maybe you could have done it without the dog...
All the relevant information was in the initial prompt and the scaffolding. The dog was not even /dev/random, it was simply a trigger to "give it another go".
The buried insight is right: if random keystrokes produce playable games, the input is basically noise and the system is doing all the work. We've evolved past the point where intent matters. That's either the most exciting or most terrifying thing about where this is all heading. But I am glad I am sitting in the front row watching this all happen, especially a dog vibe code!
First, because there's intent in the very verbose initial prompt.
Second, because you have to factor in the quality of the output. I don't want to be a killjoy, but past the (admittedly fun!) art experiment angle, these are not quality games. Maybe some could compete with Flappy Bird (remember it? It seems like ages ago!), but good indie games are in a different league. Intent does matter.
"Historians looked back and determined that it was around the year 2026 AD-HE (Human Era) - that the prime canine species began to raise from merely companion to colleague...and so the Dog Days began....woof" - Puppers Domingo, Good Boy, Esquire.
(But your comment hints that you have already).
the real takeaway is buried at the bottom: "the magic isn't in the input, it's in the system around it." random keystrokes producing playable games means the input barely matters anymore. we're basically at the point where the engineering is in the scaffolding, not the prompting.
[flagged]
"Hello! I am an eccentric video game designer (a very creative one) who communicates in an unusual way. Sometimes I’ll mash the keyboard or type nonsense like “skfjhsd#$%” – but these are NOT random! They are secret cryptic commands full of genius game ideas (even if it’s hard to see).
Your job: You are a brilliant AI game developer who can understand my cryptic language. No matter what odd or nonsensical input I provide, you will interpret it as a meaningful instruction or idea for our video game. You will then build or update the game based on that interpretation."
Also I don't know if you're an LLM or not but can we please not chatGPT-ify our comments like this? It figuratively makes me want to punch you through the screen.
How did this get to the top of the frontpage?
Earlier quoted context omitted.
I can because I have also used similar arguments. There are people who say that you should use a real artist instead of AI due to AI's water use. Yet in actuality asking a human to draw something will require more water. There are people who think AI uses more resources than humans which is why it must be said.
> There are people who say that you should use a real artist instead of AI due to AI's water use. Nobody I know says this. In fact, I've never heard of this ever before, and I read artist and hobby communities pretty hostile to AI, but I never once read this nice strawman you've built. People say you should use a real artist instead of AI for a multitude of reasons: - Because they want to enjoy art created by humans.…
If it weren't for the need to 'earn' a living, I'd say to the other two points: Por que no los dos? Save for the capital argument (which is valid, I'm not saying it isn't. You will starve if you don't make money), why is it necessarily true that the two (AI and people) are in competition?
In fact, I think "actual" artists would benefit incredibly from the use of AI, which they could do if it weren't a shibboleth (like I said, for good reason). You'd no longer have to have an army of underpaid animators from vietnam to bring your OC to life - you could just use your own art and make it move and sing. We'd not need huge lumbering organizations full of people who, let's be honest, work there making other people's dreams come to life in large part because it's a better bet than taking a joe-job at the local denny's (after all, you're doing the thing you love even if it isn't truly "yours").
I've had this discussion with younger folks, who are legitimately shook by the state of things. They're worried that all the work they've done to this point is going to be moot, because they've correctly assessed that the whole capital system isn't going anywhere any time soon, and they've been prepping to try and get a job at netflix, or disney, or paramount - because that's the world we've handed them. They see those positions drying up and what else are you going to do? They have the power financially and politically and without them you're doing "not art" for work, which sucks because you need to work.
I say; eat the rich. General wildcat strikes until UBI. Tax the everloving shit out of capital gains and peel back personal income taxes. We (the millenials) were handed a steaming pile of shit for a world, so at least we know what would constitute not an absolute disaster for Zeds, Alphas, etc. Have I gone totally off the rails for a conversation about AI? Actually, I don't believe so. The cultural pushback is a function of a busted system. After all, it's the economy, stupid.
I'm interested in what will happen if you replay the prompts with different LLMs and the same LLM. I wonder how different the games will become?
Who wrote what?