CHATPGT-4 in stellaris

[Stellaris enjoyer]

Is anyaone can integrate Chatgpt-4 in stellaris. For example you can communicate with other country as you want and say them what you want. Their answer will be based on the principles of the state.

[corisai]

Nguyên văn bởi Stormwind:

1. Dynamic and Responsive Dialogue: ChatGPT can generate dialogue that adapts to players' input.

Default? Don't adapt at all it's an illusion - system don't allowed to use input for self-education.
Allowing it to educate on player promt without any kind of censorship could end with abrassive players teaching AI to speak nasty things => obvious ♥♥♥♥ storm on devs.

Nguyên văn bởi Stormwind:

Indeed, integrating AI would require significant work from the developers.

And on top of it - will turn game into online one (or even more work to bypass it with scripted answers when there is no internet connection). Current LLM are faaaar from being compact or lightweighted :)

[corisai]

Nguyên văn bởi Sputnik:

While I certainly expect humanity to be finally snuffed out by AI, I still look forward to having a little fun with it along the way.

Don't worry AI for decades have a control of modern nuclear weapons in Russia and nothing happened (yet).

Nguyên văn bởi Sputnik:

In multiplayer games you'll never quite be certain just who is real, and who is a NPC.

C'mooon. Just a quick discussion on "dangerous" themes will show it easily. Mass-produced AI will never be allowed to have "own" opinion on politically dangerous questions.

Nguyên văn bởi Sputnik:

Heck, maybe even the Total War battle AI will move its troops out of my artillery fire!

Sorry but it is absolutely unrelated with current "AI" in form of LLM and impossible for them by design :) And don't forget - devs doing games for average player to be capable to win & having fun. So you will find complex and really efficient AI only in small niche games (like now, lol).

[rafael.ramus]

Nguyên văn bởi corisai:

Nguyên văn bởi Sputnik:

While I certainly expect humanity to be finally snuffed out by AI, I still look forward to having a little fun with it along the way.

Don't worry AI for decades have a control of modern nuclear weapons in Russia and nothing happened (yet).

That thing scares me to no end.

[Stormwind]

Nguyên văn bởi corisai:

Nguyên văn bởi Stormwind:

1. Dynamic and Responsive Dialogue: ChatGPT can generate dialogue that adapts to players' input.

Default? Don't adapt at all it's an illusion - system don't allowed to use input for self-education.
Allowing it to educate on player promt without any kind of censorship could end with abrassive players teaching AI to speak nasty things => obvious ♥♥♥♥ storm on devs.

Nguyên văn bởi Stormwind:

Indeed, integrating AI would require significant work from the developers.

And on top of it - will turn game into online one (or even more work to bypass it with scripted answers when there is no internet connection). Current LLM are faaaar from being compact or lightweighted :)

That's a valid point you're making. When I say "adapt", I don't mean that the AI learns from individual players' input during the game in a way that changes its future behavior or that it could be 'taught' to behave poorly.

Instead, "adapt" in this context means that the AI can generate responses that are relevant to the specific input it's given in the moment. For example, if a player asks an NPC about a particular topic, the AI can generate a response that addresses that topic.

As for potential misuse, it's indeed crucial to have systems in place to prevent inappropriate use of the technology. AI should be guided by a carefully designed set of rules and restrictions defined by the developers to ensure that it behaves in line with the game's guidelines and maintains a respectful and engaging environment.

[Chrisgk]

Nguyên văn bởi Immortalis:

I'll just stand by my very first line, hoping to be proven wrong in the near future.

Nguyên văn bởi Immortalis:

It won't work. At least not with this level of technology.

oh ♥♥♥♥ i have made a gpt 2 model based mod am i a genius or something or you are just arrogant ?

[CrUsHeR]

You really need an AI to talk to other empires?


Xenophobes - "We kill U all"

Xenophiles - "Send Nudes Plz"

Spiritualists - "Praise the Lord... or else"


I mean, that's all you need to hear from them. This is a resource management game with strategy elements, not a dating sim.

If anything, it could be interesting to see a neural network take control of an actual player empire. But that wouldn't work in realtime, perhaps in a turn-based game with few players, few units, few options in general.

[Sputnik]

I think you guys are far underestimating an emerging technology. Just off the top of my head we can expect stuff like this:

• Talking to an AI empire freely, just like it was a player
  
• Not knowing for certain if you're talking to AI or a player.
  
• Procedurally generated narratives based on an AI's 'needs' (e.g. to some backstory or satisfy some hidden character quirk).
  
• An opponent with machine intelligence that counters your preferred playstyle, forcing you to play a flexible strategy (for once) just to keep ahead of the game

Now go right ahead tell me that's all just a pipe dream like first person shooters were
(they weren't)

Or tell me on-line players are already that talented and interesting
(they aren't)

[Cinnamoon_dragon]

It's best to use neural network AI as a game master to generate roleplaying situations, from what I've seen AI can even generate quite a meaningful plot for a D&D party with forks of choice for your actions, in the spirit of choose your adventure, as well as simulate a die roll. People only complain that the AI ​​is very censored, so not every deep fantasy can be realized.

[corisai]

Nguyên văn bởi Sputnik:

I think you guys are far underestimating an emerging technology.

No, player's are overreacting to hype. LLM like ChatGPT are very limited in capabilities and mostly useless in current way for anything but messing around and having a fun. Plus many big failures in AI development are not known to people outside of industry.

The same situation happens with any new tech - common people overestimating it capabilities A LOT. First nuclear reactor? Fantasies about nuclear-driven robots in every house. First computer networks? Fantasies about SkyNet. And so on.

[Sputnik]

Nguyên văn bởi corisai:

Nguyên văn bởi Sputnik:

I think you guys are far underestimating an emerging technology.

No, player's are overreacting to hype. LLM like ChatGPT are very limited in capabilities and mostly useless in current way for anything but messing around and having a fun. Plus many big failures in AI development are not known to people outside of industry.

The same situation happens with any new tech - common people overestimating it capabilities A LOT. First nuclear reactor? Fantasies about nuclear-driven robots in every house. First computer networks? Fantasies about SkyNet. And so on.

I don't think my suggestions for game applications QUITE falls in with your examples of catastrophizing.. But, I'll ponder that here in the warmth of your gaslight.

In my (now distressingly long) experience, your "common people" follow this pattern in the face of radical change:

1. Bag the approaching New as impossible or irrelevant
   
2. Once the new becomes a commonplace, say it was all just commonsense and predictable.
   
3. Roost on this new orthodoxy, crowing denials at the next wave of change looming over them

[corisai]

Nguyên văn bởi Sputnik:

In my (now distressingly long) experience, your "common people" follow this pattern in the face of radical change

Problem that there is nothing really radical with LLM or ChatGPT itself. Aside from a nice tool to automatize SEO :)

There is no doubt that AI technology will change the world - someday. But like with fusion reactors - we could need an another century for it.

And yes, you're obviously a common player not related to AI industy. It's because of this text:

Nguyên văn bởi Sputnik:

An opponent with machine intelligence that counters your preferred playstyle, forcing you to play a flexible strategy (for once) just to keep ahead of the game

We already could create a very skilled AI player that will win against world top players. Major issue that most players DON'T like to lose their game and will NOT like to play against such AI.

Creating AI that will be a bit behind of player is much more complex task (~equal to create individual AI for every player) that will require a lot of hardware resources and time (that's also a big problem for average player).

[Sputnik]

Nguyên văn bởi corisai:

And yes, you're obviously a common player not related to AI industy. It's because of this text:

Nguyên văn bởi Sputnik:

An opponent with machine intelligence that counters your preferred playstyle, forcing you to play a flexible strategy (for once) just to keep ahead of the game

We already could create a very skilled AI player that will win against world top players. Major issue that most players DON'T like to lose their game and will NOT like to play against such AI.

You're quite right- I forgot to add that key crowing: point :

4. "The impending change won't come because there's really no call for it"

[Sputnik]

Nguyên văn bởi https://career.paradoxplaza.com/jobs/2984952-content-designer-stellaris:

Content Designer - Stellaris

Your Mission
As a Content Designer on Stellaris, it is your mission to conceive and implement captivating narratives, allowing players to explore their fantasies on a galactic scale. Additionally, you will be designing mechanics that will challenge and delight hardcore fans and new players alike.

Important information about the process:
When you apply for the position, you will receive an invite to a digital interview with the chatbot Hubert, which takes about 10-15 minutes. We would like you to respond to the interview within 48 hours if possible. Please check this article before the interview for tips and tricks: https://www.hubert.ai/insights/advice-before-your-first-hubert-interview

Can anyone else spot the glorious irony here?

[Seva Adom]

So one of my neighbours is a developer/owner of a game that is in the top 30 player count steam games.

He says one of the most cost intensive parts of developing the game is new levels, I suggested he implement AI into level design and patching. He said not a chance all his developers have been warned not to even visit website, the punishment will be severe if they break the rule.

Big firms are terrified of using AI in product development as they are scared of getting sued to hell and back for copyright infringement.

So say in your Stellaris game the AI starts chatting like the Borg or maybe the empire from starwars, without specific licencing it'll cost the developer millions in litigation. Same with development of FPS, what if your FPS shooter rips of DE_Dust 2 from CS, ect. ect.

Open source AI is great help for start ups, sensible established developers give it a wide birth.

I'm looking forward to playing licenced Star Trek Universe on Stellaris as well as Starwars :)

[pete3great]

I'd *love* to see Paradox have to pay out all the authors ChatGPT stole from.