This was made in 2 days and 91% of the Max 20x plan, as the author stated on the Reddit thread, so roughly ~$200. Supposedly, existing free assets were used and weren't generated.
I'd say demos like these stand to profit the most from LLMs, if the goal is to make as much as possible in a few days: a barrage of quests are easy to generate, so are gear choices, and some skills for the initial 9 classes to pick from. A human would generally spend a lot of time here, thinking about whether the class/skill choices fit their world, what type of progression is fun and isn't. It's also where player testing would be important for a game to set good pacing and balance the difficulty.
Of course, the game itself is barely playable, it randomly stutters when I walk too far away from camp, the character controls are unintuitive, etc. A lot of this stuff could be chipped away by spending more time on the project and testing it yourself, getting a feel for what you want the game to be. That by itself should require a game to take more than a few days, if we expect others to play it and enjoy it. Something simple like movement controls could take many game iterations to iron out, and those aren't hard technical tasks.
Still, I can't entirely wrap my head around the fact that I live in a world where a machine can create this with minimal intervention by humans, and do a somewhat OK job at it, to the point where I'm willing to spend 10 minutes playing it.
> Still, I can't entirely wrap my head around the fact that I live in a world where a machine can create this with minimal intervention by humans, and do a somewhat OK job at it, to the point where I'm willing to spend 10 minutes playing it.
I don't know what that means. I can post something random on social media and there is a chance some person will spend 10 minutes on it. I don't need an LLM nor any money for that effect.
> a world where a machine can create this with minimal intervention by humans
"Create" is doing a lot of lifting here. As you (and the original author) mentioned, almost everything was assembled from downloaded free assets and libraries. Almost everything is a copy-paste. The Fable part is linking and debugging until it doesn't crash.
The main value proposition of LLMs is to wash the credit away from the giants and take it for yourself.
I wish we would give credit to Kenney [1] for making sick asset packs, mrdoob for making THREE.js [2], etc. than Fable for running curl/wget...
This demo actually kinda blows my mind and makes me want to purse a game idea I had that wanted this exact aesthetic and capability
It gets said ad nauseam but a lot of software development is remixing. Think about how much gaming innovation happened in the Warcraft and StarCraft map editors. The Birth of tower defense, moba, and probably many more.
“Boars are attacking the town! Kill 200 to thin the herd.”
“We tried everything to cure the boars, but we couldn’t. If they go any farther, they’ll infect everything. Kill 200 to stop them from infecting the forest.”
“You are inside your own mind. Your fears surround you. Fear of the boar. Free yourself. Kill 200 dream boars.”
> by spending more time on the project and testing it yourself
> a human would generally spend a lot of time here,
This was always the hard part - game engines, asset libraries and all other services / SDKs were always making code-generation cheaper every generation.
Releasing a bug-free, thoughtfully built product requires a lot of attention and product skill.
>Still, I can't entirely wrap my head around the fact that I live in a world where a machine can create this with minimal intervention by humans, and do a somewhat OK job at it, to the point where I'm willing to spend 10 minutes playing it.
What is the point of all these projects like this, except to say that you did something? (that you didn't actually do but half-assed a minimally plausible version of, so hey)
It's tale as old as time at this point: the LLM produced something sort of shaped like some other software, and did it in an impressively short amount of time, but it's basically impossible to bring a codebase made in this way up to production standards, or to maintain it in any reasonable way. Nobody's gonna pay money for this, or want to play it instead of Warcraft. Why should anybody care?
Everytime i read one of these vibe coded projects i wonder: Is AI capable of building well structured programs? Designs with strong separation of concerns. Clean code. Short, well defined functions.
This is not how I'd design much of this. Does that matter? AI and whatever training data used seems to differ.
Yes, you just need to tell it to do that. Before I start I go back and forth talking about the architecture and making design docs. If you start with good bones, things go much better than stepping through adding things and ideas ad hoc.
I've built very large structured programs with claude. Talking about the structure is indeed an important part of the exercise. It's also an important factor in success. Context is limited and separation of concerns is an essential part in the LLM being able to do it at all. The chunk of "what needs to be done" needs to be small enough for it to be able to recall and reason about. Bad architecture will result in spinning your wheels constantly changing spaghetti soup that never meets spec.
Building a CAD kernel one of the essential pieces in getting from vaguely working to closing an extremely large number of gaps was some rather strict separation of concerns – without it we were just stuck on perpetual rearchitecting switching from methodology to methodology opening new gaps with each attempt to close others.
It's impressive that Fable 5 was able to resynthesise something like this from its training data, but I am really not looking forward to more of this. What's the point?
I built this with Fable over a couple of days, on the side. It's a vanilla-WoW-flavoured micro-MMO in the browser: nine classic classes, three zones, a 5-player instanced dungeon, parties/duels/trades, and persistent characters. Free to play: https://worldofclaudecraft.com — and fully open source (MIT):
Honestly the most mind-blowing part for me was how much it shipped that I never asked for. The level of polish and completeness coming out of the model genuinely surprised me — quest logs, threat metrics in the combat log, eating/drinking, spirit release on death.
This is cool! I assume a lot of this was Fable orchestrating sub-agents with cheaper models, right? Something I noticed with Fable is that if it spun off three sub-agents in the cloud version of claude code, and then hit the 5-hour usage limit, all the work of those sub-agents would be lost (!). Did you run into the same thing?
One time though, I hit the limit when not running a sub-agent, and the agent resumed after the limit expired. Weird.
I’m on a phone so I can’t see what this does, but it reminded me of this great presentation of a game style agent manager AgentCraft: putting the orc in orchestration https://youtu.be/kR64LOqBBCU?si=d3IS7SVy2lv0hM_A
I was able to put in a username and password, choose a character and "enter the realm" but then within a second of the game graphics loading it crashes so I can't actually see what the game is about. Since it's called "World of Claudecraft" I thought it might be similar in concept to the AgentCraft video I posted. I still don't know if that's true.
Just a hotpatch maintenance window to release The Token Burning Crusade expansion. At this rate we're looking for a true WoC Classic release for the real fans in like a couple weeks.
This was made in 2 days and 91% of the Max 20x plan, as the author stated on the Reddit thread, so roughly ~$200. Supposedly, existing free assets were used and weren't generated.
I'd say demos like these stand to profit the most from LLMs, if the goal is to make as much as possible in a few days: a barrage of quests are easy to generate, so are gear choices, and some skills for the initial 9 classes to pick from. A human would generally spend a lot of time here, thinking about whether the class/skill choices fit their world, what type of progression is fun and isn't. It's also where player testing would be important for a game to set good pacing and balance the difficulty.
Of course, the game itself is barely playable, it randomly stutters when I walk too far away from camp, the character controls are unintuitive, etc. A lot of this stuff could be chipped away by spending more time on the project and testing it yourself, getting a feel for what you want the game to be. That by itself should require a game to take more than a few days, if we expect others to play it and enjoy it. Something simple like movement controls could take many game iterations to iron out, and those aren't hard technical tasks.
Still, I can't entirely wrap my head around the fact that I live in a world where a machine can create this with minimal intervention by humans, and do a somewhat OK job at it, to the point where I'm willing to spend 10 minutes playing it.
> Still, I can't entirely wrap my head around the fact that I live in a world where a machine can create this with minimal intervention by humans, and do a somewhat OK job at it, to the point where I'm willing to spend 10 minutes playing it.
I don't know what that means. I can post something random on social media and there is a chance some person will spend 10 minutes on it. I don't need an LLM nor any money for that effect.
> a world where a machine can create this with minimal intervention by humans
"Create" is doing a lot of lifting here. As you (and the original author) mentioned, almost everything was assembled from downloaded free assets and libraries. Almost everything is a copy-paste. The Fable part is linking and debugging until it doesn't crash.
The main value proposition of LLMs is to wash the credit away from the giants and take it for yourself.
I wish we would give credit to Kenney [1] for making sick asset packs, mrdoob for making THREE.js [2], etc. than Fable for running curl/wget...
[1] https://kenney.nl/assets [2] https://threejs.org/
This demo actually kinda blows my mind and makes me want to purse a game idea I had that wanted this exact aesthetic and capability
It gets said ad nauseam but a lot of software development is remixing. Think about how much gaming innovation happened in the Warcraft and StarCraft map editors. The Birth of tower defense, moba, and probably many more.
I've said this before, but LLMs are the next evolution of content consumption.
You no longer need another human to consume content, you just prompt your AI for the dopamine you want in that moment instead!
So basically what a normal person is going to do initially?
the difference people typically credit where things come from and not 'look at what the llm did' See how that works?
> a barrage of quests are easy to generate, so are gear choices
Yeah that sounds like WoW retail heh.
Kill 200 boars.
Kill 300 boars.
Kill 250 boars and use this sword.
Kill 251 boars and use this special sword.
I heard about a better sword over there. You have to get past the 200 boars.
Wow, thank you for saving me from the boars. Please take this Boar Bane sword. You should try it out on 200 boars.
Or, as a variation on the theme...
kill 200 brown boars.
kill 300 black boars.
“Boars are attacking the town! Kill 200 to thin the herd.”
“We tried everything to cure the boars, but we couldn’t. If they go any farther, they’ll infect everything. Kill 200 to stop them from infecting the forest.”
“You are inside your own mind. Your fears surround you. Fear of the boar. Free yourself. Kill 200 dream boars.”
Kill 3 boars, 3 spiders, 3 wolves. Forgot to tell you they share spawns so if you don't find any of X, kill Y & Z until X spawns.
:upsidedownface:
Don’t forget, “Kill the guy and get the thing, then take that to this other guy.”
The plans have weekly limits, but are charged monthly. So this is actually only about ~$45.
91% of weekly limit of 20x plan is about 1500 bucks in API equivalent.
> by spending more time on the project and testing it yourself
> a human would generally spend a lot of time here,
This was always the hard part - game engines, asset libraries and all other services / SDKs were always making code-generation cheaper every generation.
Releasing a bug-free, thoughtfully built product requires a lot of attention and product skill.
>Still, I can't entirely wrap my head around the fact that I live in a world where a machine can create this with minimal intervention by humans, and do a somewhat OK job at it, to the point where I'm willing to spend 10 minutes playing it.
Does this excite you?
What is the point of all these projects like this, except to say that you did something? (that you didn't actually do but half-assed a minimally plausible version of, so hey)
It's tale as old as time at this point: the LLM produced something sort of shaped like some other software, and did it in an impressively short amount of time, but it's basically impossible to bring a codebase made in this way up to production standards, or to maintain it in any reasonable way. Nobody's gonna pay money for this, or want to play it instead of Warcraft. Why should anybody care?
I guess we'll just call anything MMO now.
What would you call it? A small multiplayer online?
Impressive for a few days of Fable time.
And here I am watching my 5-hour window disappear over a couple simple tasks in a CRUD app.
You can generate simple CRUD apps with free plans / old models.
Not surprised that it is impossible to play.
Everytime i read one of these vibe coded projects i wonder: Is AI capable of building well structured programs? Designs with strong separation of concerns. Clean code. Short, well defined functions.
This is not how I'd design much of this. Does that matter? AI and whatever training data used seems to differ.
Yes, you just need to tell it to do that. Before I start I go back and forth talking about the architecture and making design docs. If you start with good bones, things go much better than stepping through adding things and ideas ad hoc.
I thought one of the tenets of vibe coding was letting go of code quality, perhaps even to the point of not looking at the code?
I've built very large structured programs with claude. Talking about the structure is indeed an important part of the exercise. It's also an important factor in success. Context is limited and separation of concerns is an essential part in the LLM being able to do it at all. The chunk of "what needs to be done" needs to be small enough for it to be able to recall and reason about. Bad architecture will result in spinning your wheels constantly changing spaghetti soup that never meets spec.
Building a CAD kernel one of the essential pieces in getting from vaguely working to closing an extremely large number of gaps was some rather strict separation of concerns – without it we were just stuck on perpetual rearchitecting switching from methodology to methodology opening new gaps with each attempt to close others.
> 502 Bad Gateway nginx/1.24.0 (Ubuntu)
hugged to death?
Yeah RIP. Had a few performance issues which we're finding and fixing as we go
> 502 Bad Gateway
> nginx/1.24.0 (Ubuntu)
commenting so I remember to check again later when it's back.
https://www.reddit.com/r/ClaudeAI/comments/1u3m6a8/i_vibe_co...
Repo: https://github.com/levy-street/world-of-claudecraft
Ha, why is the creator's account banned?
Its reddit side ban not subreddit side so likely spamming en bulk triggered.
This is so cool. Where did the free assets / character models come from?
It's impressive that Fable 5 was able to resynthesise something like this from its training data, but I am really not looking forward to more of this. What's the point?
Creator here — fun to see this show up!
I built this with Fable over a couple of days, on the side. It's a vanilla-WoW-flavoured micro-MMO in the browser: nine classic classes, three zones, a 5-player instanced dungeon, parties/duels/trades, and persistent characters. Free to play: https://worldofclaudecraft.com — and fully open source (MIT):
https://github.com/levy-street/world-of-claudecraft
Honestly the most mind-blowing part for me was how much it shipped that I never asked for. The level of polish and completeness coming out of the model genuinely surprised me — quest logs, threat metrics in the combat log, eating/drinking, spirit release on death.
We already have some contributors on GitHub!
This is cool! I assume a lot of this was Fable orchestrating sub-agents with cheaper models, right? Something I noticed with Fable is that if it spun off three sub-agents in the cloud version of claude code, and then hit the 5-hour usage limit, all the work of those sub-agents would be lost (!). Did you run into the same thing?
One time though, I hit the limit when not running a sub-agent, and the agent resumed after the limit expired. Weird.
No this was all running on Fable. Used like 93% of my Max plan over the course of 2 days to get the initial version running
I’m on a phone so I can’t see what this does, but it reminded me of this great presentation of a game style agent manager AgentCraft: putting the orc in orchestration https://youtu.be/kR64LOqBBCU?si=d3IS7SVy2lv0hM_A
fwiw, it loads on my (i)phone
I was able to put in a username and password, choose a character and "enter the realm" but then within a second of the game graphics loading it crashes so I can't actually see what the game is about. Since it's called "World of Claudecraft" I thought it might be similar in concept to the AgentCraft video I posted. I still don't know if that's true.
It loads but you can’t interact
Well, this is outrageously impressive
Thank you!
hugged to death
Just a hotpatch maintenance window to release The Token Burning Crusade expansion. At this rate we're looking for a true WoC Classic release for the real fans in like a couple weeks.
Claude, add support for Firefox and Mobile views!
The rotation makes me motion-sick
i got a 502
Doesn't seem to work on mobile?
Aaand it’s down.
The right-click to attack isn't really friendly to clickpad configurations (like Apple laptops). Still impressive!