Mods / chatAI
- Tags:
- Author:
- roquess
- Side:
- Both
- Created:
- May 14th at 8:48 AM
- Last modified:
- May 21st at 3:56 AM
- Downloads:
- 86
- Follow Unfollow 11
-
Latest release (for Vintage Story 1.19.0-pre.0 - 1.22.2, potentially outdated):
chataimod_1.0.3.zip 1-click install
A lightweight server mod that adds a /ai command — for server admins who want to offer an LLM assistant to their
players.
Designed with control in mind
- Rate limiting per player to prevent spam
- Only the server owner configures the provider and API key
- Responses go only to the player who asked — no chat pollution
Ollama first
The recommended setup uses Ollama (https://ollama.com) running locally on your server: no external API calls, no
data centers, no ongoing cost. Mistral and OpenAI are available for those who prefer a hosted provider.
A tool, not an oracle
LLMs make things up — players should know that. This mod is best used on servers where the admin wants to
experiment with AI assistance, not as a replacement for the wiki or handbook.
Commands
/ai <question> — all players (.ai also works)
/ai config show|provider|apikey — server admin only
Extending
The IAIProvider interface lets you plug in other backends or drive NPC behaviors without modifying this mod.
| Mod Version | Mod Identifier | For Game version | Downloads | Released | Changelog | Download | 1-click mod install* |
|---|---|---|---|---|---|---|---|
| 1.0.3 | chataimod | 19 | May 21st at 3:56 AM | chataimod_1.0.3.zip | 1-click install | ||
|
Fix: ConfigLib panel appearing empty Problem: The configuration panel in ConfigLib showed up completely empty despite being correctly installed. Root cause: vsimgui 1.2.5 updated ImGuiNET to use ReadOnlySpan<char> instead of string for all widget methods Fix: ImGui.NET.dll and configlib.dll are now referenced directly at compile time (not bundled in the mod zip). ConfigLib remains optional: if configlib is not installed, the mod loads normally and simply skips the | |||||||
| 1.0.2 | chataimod | 19 | May 17th at 12:19 PM | chataimod_1.0.2.zip | 1-click install | ||
|
chatAI v1.0.2 New AI Providers - Anthropic (Claude Haiku / Sonnet / Opus) Now supports 6 providers: Ollama, Mistral, OpenAI, Anthropic, Grok, DeepSeek. New Features Per-player permission system Configurable system prompt API keys via environment variables MCP tool API for submods (Model Context Protocol) ConfigLib integration Bug Fixes - HttpClient: single shared instance to prevent socket exhaustion Configuration { | |||||||
| 1.0.1 | chataimod | 23 | May 16th at 4:24 AM | chataimod_1.0.1.zip | 1-click install | ||
|
## What's new in 1.0.1 - **Rate limiting** — configurable per-player cooldown (default 30s) to avoid API spam Requirements
| |||||||
| 1.0.0 | chataimod | 25 | May 14th at 8:52 AM | Empty | chataimod_1.0.0.zip | 1-click install | |
Just dropping this in here to present yet anoother project which is having the same upror about AI issues. You'll forget it in week, since everybody has a bad case of goldfish-brain, but I assure you, aggressive lamentation about the ills of AI online does a whole lot less than stepping outside your door and having an informative discussion with your local politicians about the tragic outcomes of poorly-developed, rushed and under-provisioned data centre projects meant to cash in uoon and bolster AI as the latest financial trend.
Thank you for this mod, it is really great!
I was reading through the code and the system prompt seems to be configurable via the config file, but it's not documented - can you please clarify how it should be provided? Is there a way to configure it in-game by the admin? I think it would also be nice to optionally toggle the chat as global.
I love the idea of having NPCs that chat at you, or among themselves, with some personality defined in prompts. Do you plan to implement anything like that?
I also saw that you have added support for tool calls, is that going to be exposed somehow? What kind of tool do you imagine connecting to this?
Thanks again, it is really cool.
Mmmm yes AI more AI into my game mmm yes generate me a steel pickaxe out of thin air chatgpt, yes install me a 4GB AI model on my pc without my consent. Claude generate me more codes so I can have the game play by itself so I can generate MORE AI slop, Claude steal the source code from Vintage Story and generate me 100GB of hyper realistic textures and models and claim it as my own, and copyright strike Tyron because they copied the source code my AI model generate that was stolen from Vintage Story. Claude generate me an AI lawyer and have it copyright protect my hard work prompt and SUE anyone who dares to criticize AI and call them domestic terrorists.
If this is to attract tech bros and investors into your mod then I hate to break it to you they're not buying your stocks lol
Other have already stated my opinion on this subject, so I'll say this: This was very brave and I'm proud of you for providing the option for those who want it.
👎 💢
it's nice to see people are pushing back against how awful Ai is in every reguard
But also thanks for making this, now I'll know what servers to avoid playing on :D
couldn't have said it better myself. it's very encouraging to see this many voices, even in a cesspit like a gaming forumn lmao, speaking out against systemic injustices and acknowledging our smaller, but important role in not engaging complacently with these destructive systems.
it's sad to realize that a lot of people railing against abstaining from using AI either in willful ignorance, or more charitably because they're simply uninformed, don't or won't realize that these problems affect more than their own bottom line in the real world. people doing real work on the ground to fight against this existential threat, whether they'd like to acknowledge it or not, are fighting for their lives too. they inhabit the same planet that's being destroyed by this technology that we do. they're dependant on the same potable water supply being taken from them, that is actively being used for these data centers to automate weapon systems to kill children overseas. they're a part of the same collective humanity that's literally keeping them alive, just like the rest of us, which is under threat due to climate disasters being accelerated by complicit AI usage. we want better than this for them too, and I hope that people who don't engage with these issues much in their life realize through this thread that there's a lot more at risk for them than "these guys don't like the thing I think can be fun sometimes", which seems to be the general consensus from the minority in here, especially under my own comment lmao. i hope one day they want better for themselves too.
love the encouragement for a better vintage story community I've been seeing, keep it up everyone:)
even if people can willfully ignore Ai just steals real people's hard work without compensation to train themselves trying to ignore the fact that it's doing real harm to a whole generation and future generations is just standing in a burning house and not caring at this point
poisoning our water supply, sucking up the power and computer parts from consumers making it harder for people to get access to modern technology, the Ai glazing is doing actual harm to human psycology, as an outlet for incels, and people like the Krafen CEO who had an Ai make up a legal case to stop the subnautica 2 devs from getting a big bonus which was the most schitzo insane idea ever for a legal case and the dude really though "ya that sounds like a good plan" college and education is being made completely illegimate
and then just the cherry ontop is people losing their jobs only for the Ai to have schitzo moments and need the same team hired back to peer review the machine, basically trying to train it to replace themselves
Also the government wanting Ai to used for autonomous weapons is just pure dystopian terminator/literally ever sci-fi end of the human race scenario ever
but you know, the chat bot makes me feel like I'm really cool and have a friend so better white knight for the biggest economic bubble since the south sea trade company C:
This has no place existing on a game made by hardworking human beings.
doh, i just got why there is a cat.
Oui oui baguette
this chat is so fucking funny lol
Consider adding ways to control access to command by user/groups/whitelist as part of the core of the mod. Logging should be present.
Also your SyncConfigToAll() - use a broadcast packet instead of iterating manually. under the hood it's the same thing, but still
Setting aside my own biases regarding AI, I have some feedback just based on inspecting the code:
Both this mod and the RimTalk mod of Rimworld use LLM, what is the difference between these two showing different reactions?
Rimworld is renowned for being all for crimes against humanity.
🤣 that makes sense
Can we get an AI Chiseling mod next? I hate being creative and chiseling everything by hand. Wish an AI could do it for me instead pls thx.
was just thinking we could also use a movement mod as well. I'm tired of finding ores and the stones I need, AI should just know where to take me so I don't have to waste precious brain power on having fun.
It's a chat assistant framework... stop hyperventilating.
what a zinger man, hope you get all the reddit karma you're looking for off that very clever lil quip.
anyway, for everyone else who actually engages with these problems in the real world, back to talking about how the uncritical and complacent use of AI by average people in all forms holds water for billionaires through engagement with their services and allows for the justification of data centers being put up which are destroying local communities and ultimately the planet as we know it in the long-term.
... zinger?
My guy... calm down. Find God. Touch grass.
It wasn't a quip, it was very literal. I can practically sense you typing "I'm literally shaking right now".
to repeat myself, if it's easier for you to think up little fan fictions about how randoms on the internet must feel in order to over-simplify people and universalize them into your pre-concieved world view, than to even attempt to empathize with or understand the actual positions people hold who are against this stuff, be my guest, but please at the very least don't make your cognitive dissonance and unwillingness to critically engage anyone else's problem but your own. last message i'm leaving, try and have a nice day:)
edit: since you're religious, I thought this might resonate with you better than my own words can. fft.
Proverbs 18:13, "He that answereth a matter before he heareth it, it is folly and shame unto him"
edit 2 so this chain doesn't continue: vvv let this response be a good rule of thumb for those reading; when you find yourself thinking that the substance of someones argument is less important to you than who "feels more mad" based on vibes, and that subjective qualifier solely determines, in your mind, who is right and who is wrong, you're probably not taking the most morally defensable position you could be taking. because who am I being unfair to with this position, exactly? billionaires? laughable. this person? couldn't be, they only presented "ur mad lol" in 3 different messages without a single acknowledgement of any justification for my position.
disengenous final boss, tell me you're alexythimic without telling me you're alexythimic lmao. brazenly displaying your unwillingness to think critically about real world issues negatively affecting people you'll never have to meet isn't the flex you think it is.
xD projecting much? The irony of your choice of phrasing is palpable. Seriously.... calm down. It's not good for your health to be living such a highstrung life.
Wew, meltdown.
roquess
Hi! nice mod
One suggestion would be to include a .md file (or a linked few for structure) that details what vitnage story is about, game mechanics etc.
ie if you ask about valid room seize criterias has a file/section where the ai can look up the answer.
That way you can provide a set of rules, it should help with hallucinations and provide some guide rails.
Also server would have the posibility of setting the style (talk like in medieval times etc) of the answer if you want to role play a bit.
Anyway ignore ludites around here, AI for sure its a topic that deserves a discussion but here is not the place and for sure that is not the tone and should be moderated.
Thanks! Actually this mod is intentionally meant to be more of a "lib" a base layer that stays as neutral as possible. Building a higher-level mod on top of this one that adds everything you describe (Vintage Story lore context, gameplay rules, medieval style for roleplay, etc.) would be a really great idea, and that's exactly why I designed it this way. If you want to take a shot at it, go for it!
I love the idea of setting the response style. Choices like Medieval Scribe or Grizzled Prospector would be cool. Maybe even tell it to talk like a pirate 🤣
117 comments for 17 downloads is quite the discrepancy
Some people who think that harassing people who use AI will wipe it out
Pretty sure their thought process doesn't go beyond "people need to see me having the right opinion", lol
Remove this garbage, AI slop has no place here
Ai is art, cry bout it.
v1.0.1 is out — a few things worth clarifying based on the feedback here.
The /ai command is now the default (.ai also works). Rate limiting is built in so players can't spam the API.
Responses go only to the player who asked, not broadcast to chat. Config is server-admin only.
On the environmental and ethics concerns: the intended setup is Ollama running locally on your server. No external
API calls, no data centers, no subscription. You pick the model, everything stays on your machine. Mistral and
OpenAI are there as options for those who want them, but they're not the focus.
On hallucinations: fair point, LLMs make things up. This isn't meant to replace the wiki or the handbook — it's a
tool for servers that want to experiment with it. Whether that fits your server's spirit is entirely up to the
admin who installs it.
Thanks for the honest feedback, even the blunt parts.
You should ignore most of the comments; people hate it just because it's trendy, and they don't even know what a local LLM is.
If the moddb had an actual ignore function this would've gotten like eight comments tops.
this lmao. all i can assume is that the devs are okay with this kind of low effort slop at this point as they appear to flat out refuse to say where they stand regarding LLMs
LOL can't even write with your own words instead of using AI
I get the hate, but I've defended use of AI tools for artists before; lacking use of the technology does not preclude it from being developed, it only disables those who would like to use it by means of social contract the same software which more well-off people would use against you. A lot of the AI discussion here boils down to this:
No, guns are tools. All of the rich men have tools, so we the poor should have tools too. Now, what we do with these tools is only our business and those said business affect. Using LLMs is good-as eating meat that's already dead. You can't stitch back together a cow that's been slaughtered, gutted, butchered and had whatever less-desireable bits processed into ground, so we accept the cow is dead and consume of its flesh, promising better life for the next livestock we rend.
Move fast and break things is the Silicon Valley mindset which brought about many other "Innovations" (for better and worse), which made the theft era of LLM training inevitible. Men with tools think they are above the nature of law and rule. Thus, tool-wielders kill many cows to rend, with no possible means to put the cows they rended back together. So we're all arguing over meat at this point, not whether the meat should continue living because there's no meat left living. We have all of this meat the rich had given themselves, so it's pointless to deprive ourselves.
Now, with all of that preamble out of the way — I could see this as a component of VS Village and VS Quest. If G3rtse could get over themselves about this idea, it could lead to more dynamic and interesting conversation with AI.
The Outlaws mod could have voice TTS which uses this chat mod to express myraid ways of addressing the player hostily, ensuring there's no two entities which sound the same, and some false relationship between groups of raider bandits could have them lament their friend's deaths or express revenge upon killing one of their own.
Animals could have hidden chat functions which an LLM will occasionally use to override default animal behaviour for introducing randomness into a chase which cannot be programmed by man, nor understood with depth through code; not unlike many animals.
There's a lot we could do with this, if only we quit arguing over all the meat that's no longer living, electing instead to put it in use as various cuts and sausages. It's already here, it's going nowhere, even after the AI is "Gone" there will still be people with more money than you running local models, and no amount of eradication online will stop it from proliferating in private dark web circles. Give it up; stop being so angry over things you can't control. Realise your government wants this AI much as you want your government to get rid of it. Accept. And move on from this, so you can direct your anger where it matters most.
This does not matter. This whole mod should not matter, if you don't want it to. Much as I hate it exists as you do, we cannot enforce a social contract where we dictate what is valid, and what valid things people should use, because it will invariably encounter unjustifiable double-standards. Take some deep breaths and go somewhere else. I did — I hate this mod exists still. But I'm not wasting my time wanting it to disappear; I'd love to see how else people could use this.
I disagree.
I'll respond to the strawman you presented. Guns are "tools," but when you have an environment where you're dismissal of the violence they bring, it leads to raised rates of violence per capita compared to similar countries. There is a reason its called gun control and not gun elimination.
Likewise, dismissing LLMs are just tools completely ignores the millions of gallons of water that are consumed to have this "tools" available, the entirely new powerplants built that are dedicated to new datacenters, and completely ignores the fact that regular people are actually harmed by them. See for just a whiff of some of the consequences of having this dismissal attitude towards it. Let alone the insane ethical implications of AI.
While I agree that locally-hosted LLMs are great for use-cases, and that there are some benefits including pathing. However, instead of trying to justify AI and dismiss all the wrong, wouldnt it be much better to just have REAL voices reading REAL lines by REAL people that can implement all the suggestions, wherever possible, that you just made? There is an inherent human touch that many people lose the ability to recognize and appreciate when you are constantly acceptant of AI.
I am very well aware the environmental damage our modern technology has rendered. The industrial revolution and its consequences have indeed been detrimental to not only mankind, but also of the planet itself. Not only do we rape the planet with our picks and shovels for finding goods deep within nature's bosom, we also transmute all of that good into raw materials which we then purpose for entertainment and war, since people hardly ever use computers to distribute and ingest anything informative anymore.
So, that's three times the planet suffers because of us:
Yet, it never stopped our governments from conquring those who came before us, and cause immeasurable human suffering to do any of this. You are a product of war and strife with wasteful by-products, as-is everyone else in modernity, including myself.
What're you going to do about it? You're still hung up about the sausage with your argument, rather than confronting the sausage-makers and licensers who permit them. What your government wants isn't in your control right now. What people do isn't anything you can contest, lest you risk infringing upon people's liberties. And so it goes; war never changes. Even if the roads we travel change how we act, there is no happy ending in trying to contest it.
so we might as well do nothing and continue to rape the planet for the benefit of the ultra-right, right?
I wasn't trying to drag it into American-centric politics. Mainly because it doesn't matter which side of the bird is talking, since the government looks out for itself either way with little consideration for their citizens. If you'd like some proof of that, consider which parties are pushing for your personally-identifiable information to be used to enable dynamic pricing based upon how you look, prohibit repair toward one's property without some variety of authrisation, and enforce "Age verification" for determining whether you're allowed to do anything with an Internet-connected appliance.
Nonetheless, you're going to make assumptions. Do it. We'll see which replies stand the test of time. What I will assure you, is that the chorus of many here isn't going to get rid of this mod, so if that would influence whether you keep on playing Vintage Story, let-alone support Anego Studio's future products, you might just want to walk now and forget this game even exists. Maybe, during that time away, you might figure out how to ameliorate the world's governing bodies and move the needle on topics you care about.
lmfao i guess at least with your head buried in the sand you dont have to worry about what the rest of the world looks like. You do you i guess.
Telling that you choose to attack us all rather than the points we make. it makes it patently clear that you have no real or legitimate counter-argument.
Try attacking the real problem then. This mod and what makes it possible isn't it.
Feel free to provide an exhausitve line-item list of points as a response and I'll happily work through them, though it'll continue being a waste of time.
why would i bother? You've made it painfully clear you're incapable of engaging in good-faith discussion
All I've said was the truth, far as I reckoned. What's your "Good-faith" respose then, that I theorise about remedies which would be upended by greed? I already made plain what humanity's been doing with technology.
I already said that I dislike this mod, and what makes it possible. But to say it should be removed for being, is an infringement of liberty akin that of denying content because it supports anthros and queers. Most of us would be banned from here if we threw as-much a fit about select Player Model Lib-dependent submissions, so we don't.
The land's already been ripped up, the content's already been stolen, the world goes on regardless how many suffered for the interest of corporations and governments who put their own interest first, and would themselves present endless bad-faith justifications for why the suffering must go on; the offending parties investigated themselves, and found nothing wrong!
So pray tell, what is your remedy?
no artist uses Ai tools, I can tell you that as an artist
no artist likes to see the corpse of other people's blood sweat and sleepless nights thrown into a wood chipper and devolved into the "base average"
the only "artists" who use Ai are the ones that just want the results but don't care about the journey. nice try
You can't adjust the truth to your liking. From generative fills in Adobe's suite, to iteration of ideas for furthering development in one's own style by-hand, there are methods people use to produce a work which has the initial concept developed by machine learning tools, while still creating the final work all by themselves. Video game works are also incorporating AI for enhanced random placement of fixed objects in a static field to appear more natural, and creation of placeholder assets to later replace with "Legitimate" work. Every time this stuff happens, people try to use shame and hate to dismiss it as fradulent. That isn't working.
Have another remedy?
Bro really out here trying to scab for the tech bros in a modder forum
dude your Ai GF isn't real, getting glazed all day by an affirming math equation doesn't mean you have to white knight for the tech bro poisoning our water supply and sucking up all the modern tech just to build a non profitable machine that can generate you waifu booba without needing to pay a real person for it
Bro really here talking like he knows me while doing nothing about the problem at-hand.
How about you get up off your duff and actually do something about the issue of abusive AI use? Sure there are a lot of politicians in your area who'd love to have an ear full of you.
⚠️ Inflammatory and derogatory comments are against rules 3 and 5 and will lead to sanctions when ModDB Moderators are back on duty. Lack of immediate repercussions doesn't mean you can safely continue.
This is not the place for AI flamewars. If you don't have any constructive feedback for the mod's function, limit your Comment to just👎and go find something else to do.
If you have beef with a particular comment here, use #support-moderators in Discord rather than reply to the person (don't feed the troll).
Try to have a nice day.
Man, i wish this was the main point of the discussion. Good on you for pointing this out. Report this on the issue tab in github of the mod if you can.
Holy crap. Wow, yeah, that's bad. I hadn't even looked at the code. I'm suddenly vindicated in my spending an hour fighting against this crap here. Edit to ping Michaloid so hopefully something can be done about this quickly.
I actually looked at it myself and it does look a bit sus, however the AI responses are largely limited to chat message responses. What direct file access are you seeing here? The config file handling only touches the config file itself and isn't handled by the LLM.
So for this:
"LLM has direct file access to your game files — a crafted prompt could delete your save"
I have to say it's false.
The API key storage is concerning, but it is only kept locally so it's only as secure as the server (or client) it's running on. Not really a concern as it's standard practice for server config files anyway. I run enterprise solutions that handle PHI and PPI on a daily basis that have api keys stored in plain text.
"Remapping Ollama to any endpoint is trivial"
Technically true, but if you follow the code path it requires controlserver privileges on AICommand.cs:107, which is the HIGHEST admin role in the game. An admin who wants to remap the config can just go read the config file and edit it there. Remapping Ollama to any endpoint isn't really dangerous since you'd just be mapping it to a different LLM.
"Any user's chat can influence AI responses for all users"
Also false, if you look at the actual code. AICommand.cs:53 shows the service sending only teh current player's message with no conversation history. There is no shared session context. Each request is stateless.
"Prompt injection/social engineering"
Partially valid as a generall LLM concern and not related to this mod. A player could craft a message designed to make any AI give misleading answers to that same player, but there's no shared context, so it would only affect the player sending the messages.
Lack of rate-limiting is part of how some LLMs work, the rest of your claims about the code are verifiably false or misleading.
The only thing wrong with it is that the author used the dot notation for client commands when describing the server commands.
The OP I was responding to deleted his comments saying that the code was malicious and the mod should be taken down. The code is not malicious and the OP I was responding to had similar ideas for an AI companion mod in the game. Think about that before you blindly believe something someone posts on the internet without supporting evidence, especially when it concerns AI topics... 🙄
traugdor I did come back to apologize for the snitty response I left. You were being diligent and sticking up for someone and all that is cool, but you need to understand, I was not coming down on AI, I was not coming down on the idea either, I like using AI for a lot of things and I did also think about incorporating an agent, that is also true. My argument was that without rate limiting the agent could go crazy and the user could end up paying huge bills, that was true, the author acknowledged it and has fixed it. I proved in testing I was able to comprompise the AI and have it impersonate administration and give other players direction from that role, this is the social engineering part, the 'Admin' could provide a download link to a user and request they "install this new patch to fix the jitters." There are millions of use cases like this. That was also acknowledged by the author and they have also corrected that. And the rest was me trying to make comments the author would understand while still warning people it was not safe to download at the time, and that is exactly what I told people. We have talked, I like you, and I really don't want you thinking I am an agressor or attacking out of malice or anything. I'm even following the mod to see what happens with it :)
To summarize, it WAS scary and I felt the need to warn people, the author took immediate steps to correct the problems and that is the exact right response and says good things about this author.
I appreciate your response. I still don't see how other user's commands would be able to tell a different user to do something because they appear to be separate sessions, but perhaps there is something I missed. I can allow myself to concede that.
Michaloid
Sorry to tag you, but I do not know how to report spam/useless comments that are being spammed here.
Also on top of the constant spam and flamewarring, this person is instigating violence by encouraging with comments such as "Be the change you wish to see in the world, or something".
And it also seems that this person is now targeting everyone who does not share his enthusiasm for harassing people with him.
You have better uses of your time than defending this worthless mod. C'mon.
Also on top of your annoying "teacher teacher you forgot to give us hours of homework!" attitude, learn to spot a joke when you see it. I'll delete that comment since the joke obviously didn't land; I was mocking the saying.
The hypocrisy of that statement, man.
You're defending the most vapid and useless mod ever posted to the VS modDB. I don't think you get to try to take moral high ground here.
I don't even defend the AI stuff, i choose not to use it and walked away.
I do however will die on that hill if it means defending people getting harassed over something stupid as this.
You people need to learn that there is no such thing as black and white. Stop going after people like a peasant mob. If you truly want to change peoples mind, start with ceasing the insults, screaming and threat for violence.
You have better uses of your time yourself.
The greatest nonviolent weapon we have against AI slop is shame. Instilling shame in those who perpetuate this crap is not only legal, but moral and good. I agree that violent threats are a hard boundary, which is why I had made the comment I had, in jest of such absurdity. I've since removed it since clearly I failed to word it in a way that was blatantly satirical.
If you're going to whine and say "just ignore it / don't use it", you are a part of the problem and deserve to be shamed for perpetuating an ecological, sociological, intellectual catastrophy.
Your argument boils down to "I just walk past the water fountains labelled 'white only', it's not that big a deal".
You and the ohers are literally the ones whining about it. I'm just calling you out for the lack of maturity approach.
Yeah go ahead, shame me, make me part of the problem. that's all you can do. This is exactly why people in general don't listen to you. You resort to threats and intimidation.
Again, not defending AI, this isn't even the place for such discussion. But i'm tired of people acting like animals.
(And now you're bringing segregation and racism into the topic, wtf dude.)
I mention water fountains because AI is that big of a problem. Trying to be Switzerland about it will not garner you favor nor exempt you from being guilty in its perpetuation.
It's frankly pathetic that you need to be told this. I'm gonna go play the damn game now.
beautifully said ezrazazel
o7
the fact he didnt even attempt to respond after you verbally slapped him down is hilarious. Cant wait for reality to catch up to all these AI slop defenders and fence sitters that are literally batting for the billionaire epstein class without even knowing it 🤣
Ehhh.... relax. The vast majority of folks I've actually had a discussion about the current state of so-called AI tend to lack the fundamental understanding of what it is and how it works and what its surrounding issues are, either by circumstance or ignorance. The folks who fall into the former category tend to be more receptive to learning about AI and its surrounding issues, whereas the latter are generally impossible to have good-faith arguments with because they are victims of years of conditioning from social media to "stir the pot" rather than keep a level head and be genuine.
Don't get me wrong, I understand your sentiment and for the most part agree with you, but in order to effectively combat this very serious threat that you and I understand, we have to at least try to engage with the folks who perpetuate it in good, or at least genuine, faith. I admit I failed to do that here, so me saying this here and now is rather hypocritical, but unlike those ignorant folk who seek to prepetuate this issue I can at least sit on my high horse knowing I have some modicum of maturity.
Gentlemen, my warn before wasn't to pretty up the page. Either quit this argument or you're both getting warned/banned.
You're right, this is getting nowhere.
I apologize for getting a little heated.
I just hope people can be more mature and civil in the future.
Have a good week end.
If you're feeling heated, it's always a good idea to step away and call a moderator, even if the situation might not call for it. Otherwise people spiral to hateful speech in no time, and that's a no-go.
So thank you for complying.
I refuse to apologize for doing what little I can to fight against this garbage, but I will agree to abstain from further comments under this mod. Edit: Turns out I suck at sticking to my word, but I felt it prudent to quell DatDarnDoggo.
You can use our #support-moderators channel on discord, or alternatively make a support ticket about it or just dm me from the site.
A report button is planned, but it's not implemented yet.
Exciting news. Will a report button for mods themselves also be planned?
Yessir.
Glorious.
You know a mod is ""good"" when there are only 11 downloads and over 60 comments within 12 hours of posting. 😂
11 Downloads, 107 comments, 🤨 @grok why is that?
jarvis help i got my peanuts stuck in the gears what do i do
No. If there were a way to downvote and discourage this crap, I would.
And to the moderator below, maybe see that as a sign that the larger Vintage Story community is sick and tired of this slop gumming up the modDB.
The "larger Vintage Story community" does not care because they realize they have a choice if they wish to use it or not.
Please don't assiociate me and the "larger Vintage Story community" for the same people here that screams insults and threaten for violence because they are incapable of critical thinking and control their emotions.
I honestly do not blame Vernstein for being as upset as they are. This crap is actively destroying ecosystems and the lives of tens-if-not-hundreds of thousands of people. A baseless and indirect Twitter meme is hardly worth white knighting this slop over. Quick edit: I do however agree that it crosses a line and they should either delete the comment or a moderator should remove it.
Yeah, this is really not cool. Not really even that much of an AI hater, but this is sad
Alright. I see some unfavorable comments down here. And i'll say this. You're allowed to praise the mod, you're allowed to critique the mod, but what you're not allowed to do is insult and be uncivil towards the mod author or anyone else.
So for those who haven't seen or have forgotten. Review rule 3 and 5 of the Terms of Use.
I would point out that by their very nature AI coded mods breach the license system outlinedon that page as they are using data harvested from others work without attribution in order to "learn" in the first place and generally will be in violation of the copyright of others by their very nature.
You're thinking of AI art, code mods are drawing from existing and open documented libraries and other sources. C#, json, proto buff and so on aren't closed or proprietary sourced stuff. The data set wasn't "farmed" on unwitting people posting code snippets, the structure, syntax and what not are well documented. What's at the core of code focused models is the same that runs predictive modeling that's been integrated for years now and further back in linting algorithms etc.
While I'm sure the majority of the code used to train the LLM being used here is open source, the issue is that the LLM itself is using that data in a way that pisses on the spirit of open source software with kidneys full of stones and a liver full of cirrhosis.
Worthless and useless garbage like this does not need or deserve defense. Get real.
It's not defense, it's a correction regarding why seeking it's removal on the grounds of it violating the cited rules wouldn't work.
Step it back a bit.
I'm interested to hear more about how its... doing that to the spirit of open source, if you don't mind phrasing it less metaphorically. I've not heard that side of things before and I'm interested in the perspective.
The long version: When a for-profit company trains a machine learning algorithm on as many open source projects as possible without contributing to said projects and/or acknowledging the projects it has used (which is legally mandatory if a specific project's license requries it), that company is metaphorically standing on the shoulders of thousands of giants without even acknowledging the pedestal it relies on to exist. Even if a private person were to do so for private, non-monetary purposes, many of those licenses still either restrict or disallow such usage. If the LLM is trained on a project released under the 0BSD license, then fine, it is fully within a person or entity's legal and moral right to do whatever they want with it. If it's trained on a more restrictive license, such as one that prohibits copying or redistribution, the LLM is, by the design of literally all large language machine learning models, immediately violating the license by copying and redistributing a given project, even if only in part.
The short version: the sheer amount of raw data required to develop even an LLM "equal" or """better""" than ChatGPT 3.0 requires more input data than there currently exists in fully open-source form on the entire internet. Integrating a literal, BY DESIGN plagiarism machine into Vintage ----ing Story is the single most counterproductive and thoughtless use for so-called AI that I have ever seen, and I take personal offense that someone thought this would be a good thing to add as a mod to this absolute gem of a game.
I need to research this further before I form an opinion on this nuance of the arguments. This runs counter to previous investigations I've done into how coding models are trained and their inheritance from preceding predictive modeling algorithms/models, so there is a need to verify my understanding and these claims.
Thank you for taking the time to expand on that.
To be completely candid, I don't know the intricate workings of LLMs, but I do understand their structure enough to know that "coding LLMs" are truly nothing more than an LLM trained solely on programming and code. It's a bit of a chicken and egg situation... an LLM, which is used for "conversation", interfaces with a separate LLM that only outputs code, which the conversational LLM then formats and pipes directly to it's own output.
As a scenario: 'The User' asks 'JoeGPT' to write a [hello world] program in C#. 'JoeGPT' uses gigabytes of parameters to ask 'ClarenceCode' to write a [hello world] program in C#. 'ClarenceCode' interperets that input and uses gigabytes of parameters to output, ideally, the program desired by 'The User' in plaintext. 'JoeGPT' then sorta "asks itself" to take that plaintext output, format it, and respond to 'The User' with the generated code and a friendly "here you go, human, your [hello world]".
As far as I know, having one LLM to do both tasks is simply not feasible, even if you used a supercomputer that has access to literal hundreds of gigabytes of HBM3e memory and thousands of possible tokens per second. Quick edit: I do highly recommend doing your own research into the topic. Machine learning is a genuinely fascinating and fun field of study, but the current corporate tide of blatant disregard for both common courtesy and international law that controls the overwhelming majority of machine learning advancement is a genuine existential threat to our entire planet.
I've been following the field (machine and neural network/heuristics) for a while. I had thought of pursuing it as an excuse to go back to university at one point but I opted to start a family and move to Germany with my husband and kids instead lol. I said like ten or so years ago when i saw a demo of a heuristic generational model iterating it's way through a Mario level in absolutely no time, and the implications of what that could mean for the broader field that I was very glad I had not went all in on coding and advised some friends who had to diversify their skill sets....I had said "because in 20 years you'll need to have something else", so I was off on that guess (and they pushed this stuff out the door way too fast)
T'is the whim of the capitalist market, I guess. Whatever can generate the most capital with the least amount of risk and resources will rise to the top like oil in the ocean. An apt metaphor, I think.
Oh man right.
God I hate this dystopia we're in.
We don't even get cool cybernetics and a matrix to dive into. Rip off.
The closest we've got are DangerousThings implants and VRChat. shudders
Every single major coding LLM includes non-permitted closed source code in their training set, because they scraped whatever they possibly could to make it. Nevermind that even for open source code that was used for training, it is never properly credited as is often required by the open source license of the code.
I looked for something to verify that other than people saying this and I'm not seeing it. If you could provide a source with your claims, that would help.
I did find the ongoing case of Doe Vs github, but that is yet to be resolved.....Oh wait, no, on double checking it was resolved in a dismissal of claims. Edit again: but not a full dismissal, and somethign that could lead to actual regulation of training data-sets in general so that's a good thing.
This is in relation to open AI and github's LLM, however....I would welcome any sources on others.
Legit first time I've ever seen a mod comment in the mod DB. Hadn't even considered that was a thing.
Hammer violence is a way to summon I guess.
Same. I've browsed the ModDB comments for quite a while (mostly while not logged in, mainly as a form of avoidance since I've seen how weird it can sometimes get). Should be a sign that a mod should be removed, maybe...
Its strange how people fear the AI-boogeyman so much that they feel the need to insult a person who just makes mod that lets someone query AI through a chatbox.
I dont see myself using this mod, but its an interesting idea and I could see some sort of "dynamic dialogue" mod utilize something like this. Have villager NPCs comment on things, etc.
You're mistaking fear with disgust. This crap has no place anywhere near Vintage Story.
Definitely agree that it doesn't jive with the spirit of VS. I already dislike that the handbook shows me legitimately everything even if I've never once encountered a single predecessor of the thing.
Not that we need like a tech tree or something but some...flags to open that stuff up as you came across prerequisite materials would be nice.
Y'know what, now that you mention it... I'm honestly surprised that this approach hasn't been taken by Anego yet. Great idea! It of course makes sense to want to have the player be as informed as possible so they don't get completely lost throughout their journey, but I do agree that locking certain "tech" behind certain milestones would help reduce the overwhelming feeling of having absolutely everything right there in front of you at any time. It is a handbook, so it would make sense to organize it into chapters that would "require" (togglable of course) the player to have either cratfted or interacted with certain levels of progression in order to gain access to those progression markers' specific chapters.
It would be a lot of work to reorganize the behemoth of a encyclopedia that Vintage Story has, especially for the localization teams, but I think it could be time well spent.
I had an idea for it based on simple flag snippets patched into the handbook data that would link back to a switchboard class to handle the derived connectors via a generalized flowchart configured from sweeping and dividing what has a recipe and what doesn't, then clustering by key ingredients and further nesting those by a kind of "generation" (think broadly like the ages already in place)
This is all spitballing of course, I've not sat down at all to look at how difficult sweeping for those hooks across the API would be. I think you could probably rx it to nab most the class codes with human input on the debug/qa to catch what falls through for whatever reason. Split sections between recipes and information like it already does for guides.
Idk, maybe. I'm still relatively new to VS coding
Grog How Do Play Game 🤔
gonk, are these the droids i'm lookin for?
Michaloid yes this person right here ^ (no clue how to report a comment). Inciting violence is a big no.
Michaloid
Inciting violence.
(Sorry for the ping, i don't know how else to report but this one crosses a line in my opinion.)
grok, how do i move forward?
Why is it a server mod if it requires you to use client-side commands to interact with it? Shouldn't it be a client mod instead?
The description lies. The commands are actually server-side, so all are slashed "/ai" instead of dotted ".ai"
Shockingly, the developer an an AI vibecoded slop mod doesn't know how the mod actually works and lied in the description
It's a little embarrassing to see how people behave in the comments.
It's just a noisy minority. But yes, this is stupid.
I never imagined we had such a toxic community
Nah don't worry about it. As previously mentioned, it is only a small percentage who lack self-control, just like in every community. Don't let this minor incident ruin your view on the game.
Agreed, actively supporting something that's damaging so many aspects of humanity is very embarrassing.
oh good, a mod for the people too dense to read the handbook
also a mod for all those same people can come defend their AI girlfiends or whatever
"How dare you show this here, where i have the choice of which mod i can use or not in my game for free without forcing it down my throat under the threat of a gun against my head ! You should unalive yourself, also i'm a good person and you should be kind to others because being kind is free." ✌
Nice to see all the "adults" gracefully handling something they don't like existing and definitelly not behaving like preschoolers
If condemning plagiarism and deepfake child porn is behaving like a preschooler, then okay. I'll gladly accept that descriptor.
Kinda wild to see you or anyone endorsing those things, though. Are they popular among people like you?
"A server-side mod that lets players query an LLM directly from chat using the `.ai` command."
what the hell does childp*rn have to do with any of this ?
I honestly couldn't tell if you were feigning ignorance, but I'll be charitable and assume you genuinely don't know.
All genAI does is regurgitate things humans have already created. You're not getting anything new. You're getting permutations ripped from real original material made by people who had no opportunity to object to it.
When you don't cite your sources, that is known as plagiarism.
As for the deepfake csam, well... all models have real csam built into their training material, whether it's written, as an image, or as a video.
Anything you get genAI to vomit up for you has those things built into them inherently, and they will always affect the output.
Thank you my king for being so charitable. 🙏
You're talking to me about how you see GenAI as a whole, does this mod generate childp*rn ? You're all here polluting this site like we're on twitter, barking like rabbid dog at a guy who probably just coded it for himself and decided to share it. Bunch of you daring to tell him to "reconsiderate is life", hell i would not be surprised to see a "kys" in here or someone trying to doxx the guy, this could be a minor ffs. Acting like this guy in particular killed one of your family member. All of this is absolutely ridiculous.
I don't feign ignorance, AI has flaws that should be corrected i agree, but that discussion is not here, on a site where we enjoy free mods for a game we like.
Grow up, all of you.
lmao @ the place for discussion not being here, in the comments section explicitly for discussing the mod
okay buddy you have a good one, i was only ever polite to you
You were condescending, don't try to deflect it.
Yes, the comment section is to discuss the mod, like reporting bugs or making suggestions. Not about the world's problems and using it as an excuse to act like a pos against a guy that did absolutely nothing to you whatsoever. Again, this isn't twitter.
Have a good week end.
Leuc KnewOne Let's all drop this please. Hurling insults at each other will help no one. And it's clear you're all heated by this discussion.
I genuinely cannot fathom the mindset required to want access to the misinformation machine in your Vintage Story game.
This is how you're spending your time on Earth?
"Yes, you are totally right! In retrospect, building a pit kiln inside your wooden house was not the best idea."
This is beyond parody. Come on, man. Why.
This is garbage.
Hilarious, totally useless.
Love the idea of using vintage story as a client for my 1.5b model, even tho the mod was fully vibecoded
oh man, exactly what i wanted! i like my rustic survival game, but i always thought it needed integration with glorified spellchecker chatbots that fundamentally don't understand what they're talking about.
i'm so glad this mod exists so i can use the terrible programs that are ruining social media platforms, putting artists out of jobs, driving the costs of computer hardware into the stratosphere, making deepfake porn of children, and running on data centers that use more electricity than god, all without the ordeal of pressing alt+tab!
Amazing mod))
To what end? You can just google or use the wiki or in-game help. This game is absurdly well documented, this has no point.
some thoughts should not be acted upon
so true
ew.
lol, i still dont understand the hate on AI. Yes, AI can give false information or wrong information sometimes, it is not perfect, but specifically thats what make AI at this point harmless. Is not like it will take over your world or your life, its just a program that give answers bassed on what we as human feed it over time and "training". People who dont want to use AI, just dont use it, people who want to use it, go for it. Im not a modder, but i use ChatGPT a lot just to chat and create ficticial worlds, sometimes i play DND with chatgpt, is not perfect, it dont have memory and it messes up my inventory or stats, but it can be fun if you know that AI is just a tool and you learn how to use it with caution because it can and it will have errors.
Having an AI in game could help if you want to ask for example "wich rocks can have copper ore, or at wich deep can i find Olivine" it could be faster than checking the handbook or searching on google just to have another AI looking your answer lol. Ive seen games with AI as NPC, it can be fun because you can see how dumb are the NPC sometimes or how smart they seems to be other times.
on the other hand, claiming the code is not safe without pointing any real evidence, what about the code makes you think is dangerous? so for those like me that know nothing about code could understand what is so dangerous or problematic or sus.
As far as i know this page take down mods that are malicious or contains things that could affect your pc or have other intentions behind it. If that the case, this mod would of be deleted or we would get a warning like what happened with the rivers mod.
Im not asking people to use or like AI, what i dont understand is, if you dont like AI, why attack this person just because he made a mod that uses or have AI? Is not like he goes around attacking people who dont want to use his mod.
Fefa, I'm sorry, but you don't understand what you're talking about when you say that AI is "at this point harmless." An incomplete list of problems: AI data centers, slop replacing human content, AI taking over jobs, and AI being used by lawyers, teachers, etc. who confidently repeat what it says even though it regularly hallucinates. If everyone used AI intelligently and in moderation, then you would be partly correct, but they don't and they never ever will. And even if they did, it's not worth the tumors we're slapping all over the Earth to suck up all the water and power to make it happen. It isn't nearly good enough to justify that, those datacenters are ruining people's lives just so you can save five seconds opening a tab. It isn't worth it and it makes us dumber anyway.
On top of that, pretty much everything AI gets inserted into gets worse. People are sick and tired of seeing it everywhere, it's a reminder of how stupid and unstable the modern world is, and people don't want to see it in their games. Nobody should be attacking or harassing the mod author here (which they haven't yet as of this comment, imo) but they have every right to be angry and vent about seeing a game they love get corrupted by the latest collapsing tech-bro fantasy.
The AI on your phone or browser says otherwise.
You use AI every day without noticing it.
Any use of AI in my life is accidental. If the only response you can think of is "but I assume you also use the thing you're criticizing," you are failing to argue.
But to be clear, anyone who DOES use AI and honestly criticizes it at the same time is still correct. Unless you think that everything in use by anyone is automatically good?
Yeah and that's why tools have been getting progressively worse. They're shoehorning it into everything to justify their investments and retain high stock prices. This mod is a joke.
I am well aware of the AI being shoveled in everywhere they can cram it. Google for example have implemented AI and disabled boolean operators from searches. That means you cant accurately narrow down a search anymore and instead have to deal with an automated disinformation engine to try and find things. It's also been used to replace spellchecker systems which has it resulting in more gramatical errors and mispelled words thanks to it simply being a lousy auto-complete that just regurgitates the most popular answer, which if its wrong spreads and reenforces itself. I could go on and on, that's also before you get into the resources it guzzles to do all this in terms of both power and water. I DO notice it, and that's why people actively propogating it get so much flak.
I don't quite understand what the author of the mod has to do with it, and why so much criticism has poured on him.
If you say that he supports the AI industry by using AI, then how do you like this example:
Electricity. How much pollution occurs during its production? Tons of waste are released into the ocean, atmosphere and soil. However, who will give up electricity now?
Well some will say its the environmental impact, but I would wager they don't speak out against the other things killing the planet that have existed a long time before, so....I tend to try and find something more to the context of the location:
In the case of art, justifiably that the stuff is well known to be fed stolen stuff. Full stop, unethical unless proven otherwise and IDK if it can at this point given data set cross-breeding.
In the case of code: The issue is that the iteration of what tool this is meant to follow on from, IDE integrated predictive modeling, has been kind of cursed with the chat-bot 'I need to seek the best response to make my user the happiest!" line that LLMs have been infected with. In most cases, this means the fastest guess it can give, the 'best' assumption for the quickest engagement turn around it can muster without really diving into something...so instead of a tool that works to enhance and compliment user inputs/skills or even help devleop those skills, you get a lot of stuff that's being done to appease. Errors and issues getting silenced instead of fixed because that's the lowest impact, fasted turn around for engagement way to make errors go away.
If these things weren't keyed to say what 'should be the most appeasing' and instead keyed to only give verified facts (which source checking IS something these models CAN do, but agian...speed = laziness), you'd get far less hallucinations.
Code and things that are based on solid rule sets should be the easiest things in the world for these tools to work within to produce excellent output, but they are limited by so many factors you really need experienced developers to use 'safely' and correctly. You CAN get apps and mods and whatever out of them, the simpler the better, but do you want it working in ways that won't produce 10000 bugs that then get put on an end user (becuase the person vibe coding doesn't know basic 'best practice' programming methodology, Debugging, and QA testing rigors)....you really need someone to actually look at it who knows what's up.
Human coded mods can do damage and be buggy wrecks....hell, a properly set up LLM with the right source to point at and clear parameters and boundaries could probalby write something simple with less errors and bugs than a human....and they are great (and their previous interations on the same core have been too) at reviewing code for little missable bugs about 100 times faster than a person (trust that). You have a crash report, toss it to one and you'll get an accturate parsing in a split second. Its good at reading what's right there.....but yeah, it gets wobbly further on.
There's a lot of arguments to be made against AI, some of them are generally valid....this is my, less 'against' and more.....of a lamentation about misappropriated resources and missed opportunities. I've been doing dev stuff for years and year and having the code review, tedium bypass stuff is great...but it really needs to be uncoupled from the shit. I don't need to have a nice chat with my tools, tyvm.
I think you might particularly appreciate one of the worst things that has been replaced by AI since you've been doing dev stuff. Google decided that AI should replace boolean operators in searches and disabled them.
I noticed this just the other day and I was livid forreal
Oh thats why it hasn't been working, man. You can never hate AI enough
Are you serious? you can't be serious
this has to be bait right? i refuse to believe anyone is this slopbrained
it's really effective rage bait if true lmao
Imagine being this depended on the lying machine that you HAVE to have it in the medieval minecraft game. Genuinely, reconsider your life choices.
apart from it being made with AI I've looked at it to see what it is but there's something about it that makes me not recommend it
so this ia a big no from me
The code is fine, however, as you said, it's fully vibecoded. It's not unsafe to use this mod
I wouldn't call it fine.
It is in dire need of a proper code review from someone who actually understands Vintage Story modding.
Part of me wants to fork it and fix the issues manually and then tell the author how to update the page to be more accurate with what the mod does...but the other part of me doesn't even want to leave this comment.
This is what I tried to suggest on reddit and I got screamed at by a bunch of people who can't read about how I was an AI evangalist. I was trying to say "Hey, reality is these tools exist, let's encourage some safty and proper dev practice/test/QA/debugging etc" because yeah....vibe coders are going to vibe code, and they are the ones typically not disclosing AI use, despite it being usually obvious.....and they should really really be seeking out code review because vibe coding doesn't teach how to properly set up an environment to actually properly interface with existing API hooks without rewriting everything from near scratch...and it doesn't teach, nor is it really able to fire up the game to properly debug and most times AI is so keen on appeasing the user it will suppress error calls instead of actually fixing, so you go from loud errors to silent failures.
I wish there was a code review collective or something idk.
I get the gumption, but even modding takes a lot of time, care, and human direction to get it right beyond the bare-simple shit.
(I have not looked at this code and I won't, I just wanted to comment agreeing)
👆This.
Thank you, finally someone with critical thinking. You can't take a middle ground with those people i swear.
Although your mistake here was going on reddit to talk about it, you will get jumped on and kicked out within a minute.
Thank you. Yeah, reddit can surely suck. It was a ridiculous dog pile of really hateful stuff. Someone reported the post as me talking about committing self harm so I got this suicide prevention stuff suddenly popping up in my inbox and that's not a fun subject to read about our of nowhere.
I get people being angry or afraid, but this is wild. They are attacking people like they've found the CEO of one of these companies out in the wild instead of some kid who posted a mod or whatever.
why don't you say what it is?
Maximum slop
Why. If immersion breaking (tabbing out) is the concern - why you made this mod in the first place - why on earth would you leave it to an AI that'll hallucinate every possible answer?
A far better use of time would have been for a direct Discord integration to connect to #gameplay-help, or better yet, a personal expansion on the Handbook for more guidance.
😬
why?
grok play the fucking game for me
that's an absolute "no" from me
😒