ChatGPT o3 Mini-High (Can it code for DECORATE?)
Moderator: GZDoom Developers
Forum rules
Before asking on how to use a ZDoom feature, read the ZDoom wiki first. If you still don't understand how to use a feature, then ask here.
Please bear in mind that the people helping you do not automatically know how much you know. You may be asked to upload your project file to look at. Don't be afraid to ask questions about what things mean, but also please be patient with the people trying to help you. (And helpers, please be patient with the person you're trying to help!)
Before asking on how to use a ZDoom feature, read the ZDoom wiki first. If you still don't understand how to use a feature, then ask here.
Please bear in mind that the people helping you do not automatically know how much you know. You may be asked to upload your project file to look at. Don't be afraid to ask questions about what things mean, but also please be patient with the people trying to help you. (And helpers, please be patient with the person you're trying to help!)
-
- Posts: 968
- Joined: Mon Jan 20, 2014 6:46 pm
- Location: LV-426
Re: ChatGPT o3 Mini-High (Can it code for DECORATE?)
I saw that GROK 3 was able to reproduce the game Asteriods . So I tried asking it the same prompt I asked ChatGPT o3 Mini-High.
Prompt:
"Make a custom enemy DECORATE script for Doom II of a DoomImp that fires three Doomimpballs in a row that I can copy and paste in Slade."
Result:
ACTOR TripleShotImp 3001 { Health 60 Radius 20 Height 56 Mass 100 Speed 8 PainChance 200 Monster +FLOORCLIP SeeSound "imp/sight" PainSound "imp/pain" DeathSound "imp/death" ActiveSound "imp/active" AttackSound "imp/attack" Obituary "%o was scorched by a triple-shot imp." States { Spawn: TROO AB 10 A_Look Loop See: TROO AABBCCDD 3 A_Chase Loop Missile: TROO EF 8 A_FaceTarget TROO G 6 A_CustomMissile("DoomImpBall", 32, 0, 0) TROO G 6 A_CustomMissile("DoomImpBall", 32, 0, 0) TROO G 6 A_CustomMissile("DoomImpBall", 32, 0, 0) Goto See Pain: TROO H 2 TROO H 2 A_Pain Goto See Death: TROO I 8 TROO J 8 A_Scream TROO K 6 TROO L 6 A_NoBlocking TROO M -1 Stop Raise: TROO MLKJI 8 Goto See } }
Prompt:
"Make a custom enemy DECORATE script for Doom II of a DoomImp that fires three Doomimpballs in a row that I can copy and paste in Slade."
Result:
ACTOR TripleShotImp 3001 { Health 60 Radius 20 Height 56 Mass 100 Speed 8 PainChance 200 Monster +FLOORCLIP SeeSound "imp/sight" PainSound "imp/pain" DeathSound "imp/death" ActiveSound "imp/active" AttackSound "imp/attack" Obituary "%o was scorched by a triple-shot imp." States { Spawn: TROO AB 10 A_Look Loop See: TROO AABBCCDD 3 A_Chase Loop Missile: TROO EF 8 A_FaceTarget TROO G 6 A_CustomMissile("DoomImpBall", 32, 0, 0) TROO G 6 A_CustomMissile("DoomImpBall", 32, 0, 0) TROO G 6 A_CustomMissile("DoomImpBall", 32, 0, 0) Goto See Pain: TROO H 2 TROO H 2 A_Pain Goto See Death: TROO I 8 TROO J 8 A_Scream TROO K 6 TROO L 6 A_NoBlocking TROO M -1 Stop Raise: TROO MLKJI 8 Goto See } }
-
- Posts: 968
- Joined: Mon Jan 20, 2014 6:46 pm
- Location: LV-426
GROK 3 produced viable DECORATE script on 1st attempt
Wow! it actually worked on the first attempt, without having to provide reference.
Seems miniscule, but compared to ChatGPT, this a major leap forward.
Seems miniscule, but compared to ChatGPT, this a major leap forward.
-
- Posts: 20
- Joined: Sat Oct 26, 2024 7:42 am
- Preferred Pronouns: No Preference
Re: ChatGPT o3 Mini-High (Can it code for DECORATE?)
I think the community should focus more on this
-
- Spotlight Team
- Posts: 1109
- Joined: Mon Nov 25, 2019 8:54 am
- Graphics Processor: Intel (Modern GZDoom)
Re: ChatGPT o3 Mini-High (Can it code for DECORATE?)
I don't think it should because it stiffles creativity and for what it does the energy requirement is far too high.
People will (and are) using it as a shortcut because why care about learning how to code when ChatGPT gives you a real crappy often-not-working equivalent?
-
- Posts: 127
- Joined: Sun Feb 25, 2018 10:53 am
- Location: Tatarstan
Re: ChatGPT o3 Mini-High (Can it code for DECORATE?)
Has anyone tried drawing a set of sprites with states and directions using AI?
-
-
- Posts: 26874
- Joined: Tue Jul 15, 2003 4:58 pm
- Location: Scotland
Re: ChatGPT o3 Mini-High (Can it code for DECORATE?)
It's bad enough with the variation in height and pose between different angles/states with the default sprites. Imagine them having different numbers of fingers, limbs, whatever, on every frame.



-
- Posts: 9
- Joined: Fri Jan 17, 2025 5:41 pm
Re: ChatGPT o3 Mini-High (Can it code for DECORATE?)
AI may not be cool, but c'mon, atleast try to "fix" or "cleanup" the code, doomworld has already banned everything AI related (except for neural upscaling i believe)
not trying to stir up some trouble, but this is my opinion to be honest
not trying to stir up some trouble, but this is my opinion to be honest
-
- Posts: 827
- Joined: Thu Apr 25, 2013 12:21 pm
- Preferred Pronouns: They/Them
- Operating System Version (Optional): Windows 11 for the Motorola Powerstack II
- Graphics Processor: nVidia with Vulkan support
- Location: The Corn Fields
Re: ChatGPT o3 Mini-High (Can it code for DECORATE?)
AI stuff should be banned from this forum too to be honest. No, not taking any feedback.
-
- Posts: 127
- Joined: Sun Feb 25, 2018 10:53 am
- Location: Tatarstan
-
- Posts: 13554
- Joined: Wed Jul 16, 2003 3:52 am
- Preferred Pronouns: He/Him
- Operating System Version (Optional): Windows 11
- Graphics Processor: nVidia (Modern GZDoom)
- Location: Germany
Re: ChatGPT o3 Mini-High (Can it code for DECORATE?)
Give this another 2-3 years and it will be able to deliver perfect code.
-
- Spotlight Team
- Posts: 1109
- Joined: Mon Nov 25, 2019 8:54 am
- Graphics Processor: Intel (Modern GZDoom)
Re: ChatGPT o3 Mini-High (Can it code for DECORATE?)
You wanna know why? Because its just the next evolution in user-laziness for folks who think that this will make them on par with John Carmack. In similar gripes as to why Oblige maps aren't allowed to be on /idgames, AI-slop shouldn't be allowed anywhere either way. Its a creative stiffler.krashregister wrote: ↑Sun Mar 30, 2025 6:40 pm AI may not be cool, but c'mon, atleast try to "fix" or "cleanup" the code, doomworld has already banned everything AI related (except for neural upscaling i believe)
not trying to stir up some trouble, but this is my opinion to be honest
Agreed. All this does is make lazy people appear to have done any effort in the realm.SanyaWaffles wrote: ↑Sun Mar 30, 2025 6:55 pm AI stuff should be banned from this forum too to be honest. No, not taking any feedback.
Its the same reason why i vehemently dislike the Ghibli-AI trend going on. Ghibli's artstyle is about heart and soul. Just recreating that through AI is boring.
The question should be if we should want this.Tormentor667 wrote: ↑Tue Apr 01, 2025 11:50 am Give this another 2-3 years and it will be able to deliver perfect code.
-
- Posts: 827
- Joined: Thu Apr 25, 2013 12:21 pm
- Preferred Pronouns: They/Them
- Operating System Version (Optional): Windows 11 for the Motorola Powerstack II
- Graphics Processor: nVidia with Vulkan support
- Location: The Corn Fields
Re: ChatGPT o3 Mini-High (Can it code for DECORATE?)
Rich coming from the person with the "inspiration folder".Tormentor667 wrote: ↑Tue Apr 01, 2025 11:50 am Give this another 2-3 years and it will be able to deliver perfect code.
If this shit's gonna be allowed on the forums then I want nothing to do with the forums.
EDIT: removed an erronious statement, apologies.
-
- Posts: 127
- Joined: Sun Feb 25, 2018 10:53 am
- Location: Tatarstan
Re: ChatGPT o3 Mini-High (Can it code for DECORATE?)
It will happen anyway.
Will there be a ban on this forum or not?
-
-
- Posts: 26874
- Joined: Tue Jul 15, 2003 4:58 pm
- Location: Scotland
Re: ChatGPT o3 Mini-High (Can it code for DECORATE?)
At one extreme, an entire project constructed by AI (not possible yet, but one day); at the other someone trying to put together their own project, but struggling with one line of code, so they turn to AI and get a working solution. Are they both equally unacceptable? if not, where is the line? How much AI is too much? How can you judge? Is your too much the same as someone else's too much?
Is getting AI to produce a graphic better or worse than a modder "borrowing" someone else's work? (Neither require the modder to have any pixel-pushing skills.) What if a modder puts together an entire project, but decides to AI generate the titlepic, and gets something that they like? Is that too much? What if they produce a rough titlepic themselves and get AI to "improve" it (subjective)? Is that too much?
Is, perhaps, getting good at writing AI prompts a skill in itself? It certainly requires trial, error and learning.
I don't know the answer, but "AI" is here (as in, it already exists, not specifically on zdoom.org), people are using it, and its use is increasing. Personally, I don't think it's a tide that can be held back, but people can get more savvy about it - both in detecting and rejecting the crap, and using it in productive ways. I'm not convinced it's all bad; I'm not convinced it's all good. As with anything where there are extremes, the answer probably lies somewhere in the middle. What humans are very good at doing (in many, many situations) is falling out with each other over where the compromise should be. (So, let's not do that.)
Aside:
At an AI seminar last week, the presenters shared my opinion: that what we have now, and what people are calling AI is not actually AI; it's just increasingly efficient analysis of data to give increasingly human-like responses. They said that eventually, all tasks that humans can do could be broken down into steps and referenced against a database - e.g. judging a person's emotional state can easily (and already - to a certain extent) be done by looking at facial expressions, words used, tone and volume of voice and so on, and then comparing them to template examples. The AI decides you are unhappy and asks if they can do anything to help. It might even feel like a human enquiry. But is that intelligence? The presenters referred to it as quasi AI - giving the impression of intelligence, but not actually being intelligent. I guess the question is, is that all we as humans do too, but in a far quicker, more natural feeling way? Would some superior being look at us, see us making our decisions and drawing our conclusions and laugh at how we believe we are intelligent and in control of it all when all we are doing is comparing our situation to a lifetime of gathered data (and probably being influenced by whether we are too hot/to cold, hungry etc); that (perhaps) our responses are as predictable to them as an "AI" can be to us.
What the presenters often spoke about was not that a program was an AI, or was using AI; more that it was made "with a pinch of AI." Interesting at least.
Is getting AI to produce a graphic better or worse than a modder "borrowing" someone else's work? (Neither require the modder to have any pixel-pushing skills.) What if a modder puts together an entire project, but decides to AI generate the titlepic, and gets something that they like? Is that too much? What if they produce a rough titlepic themselves and get AI to "improve" it (subjective)? Is that too much?
Is, perhaps, getting good at writing AI prompts a skill in itself? It certainly requires trial, error and learning.
I don't know the answer, but "AI" is here (as in, it already exists, not specifically on zdoom.org), people are using it, and its use is increasing. Personally, I don't think it's a tide that can be held back, but people can get more savvy about it - both in detecting and rejecting the crap, and using it in productive ways. I'm not convinced it's all bad; I'm not convinced it's all good. As with anything where there are extremes, the answer probably lies somewhere in the middle. What humans are very good at doing (in many, many situations) is falling out with each other over where the compromise should be. (So, let's not do that.)
Aside:
At an AI seminar last week, the presenters shared my opinion: that what we have now, and what people are calling AI is not actually AI; it's just increasingly efficient analysis of data to give increasingly human-like responses. They said that eventually, all tasks that humans can do could be broken down into steps and referenced against a database - e.g. judging a person's emotional state can easily (and already - to a certain extent) be done by looking at facial expressions, words used, tone and volume of voice and so on, and then comparing them to template examples. The AI decides you are unhappy and asks if they can do anything to help. It might even feel like a human enquiry. But is that intelligence? The presenters referred to it as quasi AI - giving the impression of intelligence, but not actually being intelligent. I guess the question is, is that all we as humans do too, but in a far quicker, more natural feeling way? Would some superior being look at us, see us making our decisions and drawing our conclusions and laugh at how we believe we are intelligent and in control of it all when all we are doing is comparing our situation to a lifetime of gathered data (and probably being influenced by whether we are too hot/to cold, hungry etc); that (perhaps) our responses are as predictable to them as an "AI" can be to us.
What the presenters often spoke about was not that a program was an AI, or was using AI; more that it was made "with a pinch of AI." Interesting at least.
-
- Posts: 13554
- Joined: Wed Jul 16, 2003 3:52 am
- Preferred Pronouns: He/Him
- Operating System Version (Optional): Windows 11
- Graphics Processor: nVidia (Modern GZDoom)
- Location: Germany
Re: ChatGPT o3 Mini-High (Can it code for DECORATE?)
I completely share what Enjay is talking about.
To go in further details as some of you might know I am working as a photographer and graphics designer as a full time job and I can tell you that the so called AI will take more and more space in all of our lifes in the very near future.
You can decline that change or try to avoid it but it will be similar with trying to avoid smartphones, or computers at all. Other people will use it anyway and be more efficient, faster and better than you.
The thing is: behind every ai-creation there is a human being prompting. I am not a coder or programmer but if I have a good idea and ChatGPT or similar tools can create the proper code for it, why not? If I am not a good sprite artist but Midjourney is able to create a full sprite set out of a prompt that was my creative idea, why not?
People fear losing jobs but jobs will change. And so will creative work change. Let that happen, use it to your benefit and stay creative, that’s what it is all about.
To go in further details as some of you might know I am working as a photographer and graphics designer as a full time job and I can tell you that the so called AI will take more and more space in all of our lifes in the very near future.
You can decline that change or try to avoid it but it will be similar with trying to avoid smartphones, or computers at all. Other people will use it anyway and be more efficient, faster and better than you.
The thing is: behind every ai-creation there is a human being prompting. I am not a coder or programmer but if I have a good idea and ChatGPT or similar tools can create the proper code for it, why not? If I am not a good sprite artist but Midjourney is able to create a full sprite set out of a prompt that was my creative idea, why not?
People fear losing jobs but jobs will change. And so will creative work change. Let that happen, use it to your benefit and stay creative, that’s what it is all about.