Hello! This is a script made by ChatGBT, If you dont know already, ChatGBT is a chatbot that can write scripts. https://pastebin.com/raw/n4fGdBUC Do not remove the water mark i put there I am not a mod maker so im hoping someone will compile this.
That's... Not how vehicle work in beamng lol It's just a mix and match of different coding languages and makes no sense.
Its not for a car -_- And openAI says this: This mod will create a new vehicle that spawns on the map with a unique name and properties I also said write in lua
This isn't all thats needed. You need jbeams and many other things that aren't included in this code.
May I know what you are trying to make? This code doesn't make any sense and it will not work in BeamNG
Bruh you blindly believe that ChatGPT can write a working code... It definitely doesn't know LUA, this is not LUA at all, doesn't even seem to be any programming language, just some weirdly formatted JSON Although I kinda want to see what the mod would look like if the code was written properly, definitely nowhere near close to a car lol
That's not how it would work either. That AI has no clue of how to code for BeamNG and you're blindly believing it I think? None of what's in that code make any sense and none of it would work.
Also of note: to make a fully working car, you need like hundreds of thousands of lines of code in total, and this is less than 100 lines lol --- Post updated --- So this is what it would look like in Jbeam format most likely And this is the "car" in game It doesn't resemble a car in any way at all, it's just a weird twisted rope lol There I made it a mod, it's completely useless, good luck trying to get ChatGPT to make a real car
I like how you went through all this just to prove that ChatGPT is completely useless for this lol You also rEmoVeD thE wAteRmArK!!
Pretty sure ChatGPT pulls code from online, and guess what? there is none! people don't upload their .jbeam files online(except for attachments, it probably doesn't check those), and that's why it has no idea how to make a working vehicle. That's how machine learning works, it observes, it learns. Plus the fact that BeamNG isn't the most popular game, it wouldn't have much data on it other than "BeamNG is an advanced soft-body physics game".
The latter is true, there's possibly no code related to jbeam in GPT models offered by OpenAI, ChatGPT included. The former, however, not so. ChatGPT doesn't have access to the internet, so it can't just look things up (there is a small chance that a random link it outputs is valid, though). Neither does it have any factual knowledge – instead of relying on any accurate "facts" it operates solely on what it was fed and categorizes items according to semantic meanings and parameters (not semantic meanings and overall knowledge like humans do). Due to this, it's great at outputting text that is eerily intelligible for human readers, but it lacks any substance. For anyone curious, I also asked it to generate some jbeam code with a more appropriate model and parameters: Spoiler: Playground generated code Code: { name = "example_object", mass = 1000, drag_coeff = 0.2, restitution = 0.5, friction = 0.8, nodes = { {pos = {0, 0, 0},}, {pos = {1, 0, 0},}, {pos = {1, 1, 0},}, {pos = {0, 1, 0},}, }, beams = { { nodes = {0, 1}, stiffness = 1000000, damping = 0.1, }, { nodes = {1, 2}, stiffness = 1000000, damping = 0.1, }, { nodes = {2, 3}, stiffness = 1000000, damping = 0.1, }, { nodes = {3, 0}, stiffness = 1000000, damping = 0.1, }, }, } It looks like it is syntactically correct, but I can't say what language it is. You can also give it some examples, and it will try to imitate them. Spoiler: Find and replace GPT3 started after // new code Code: Here is an example BeamNG material code for a vehicle called "bolide". Create a new one for "covet". // my code { "bolide_main.skin.bolide_main_skin_street": { "name": "bolide_main.skin.bolide_main_skin_street", "mapTo": "bolide_main.skin.bolide_main_skin_street", "class": "Material", "persistentId": "badce8fc-be86-42ae-b36a-4f8b9e1adbe1", "Stages": [ { "ambientOcclusionMap": "/vehicles/bolide/bolide_main_ao.data.png", "baseColorMap": "/vehicles/bolide/bolide_main_b.color.png", "metallicFactor": 1, "metallicMap": "/vehicles/bolide/bolide_main_m.data.png", "normalMap": "vehicles/bolide/bolide_main_nm.normal.png", "roughnessMap": "/vehicles/bolide/bolide_main_r.data.png" }, { "ambientOcclusionMap": "/vehicles/bolide/bolide_main_ao.data.png", "baseColorMap": "/vehicles/bolide/bolide_main_skin_street_a.color.png", "clearCoatFactor": 1, "colorPaletteMap": "/vehicles/bolide/bolide_main_skin_street_bb.color.png", "colorPaletteMapUseUV": 1, "diffuseMapUseUV": 1, "instanceDiffuse": true, "metallicFactor": 1, "metallicMap": "/vehicles/bolide/bolide_main_skin_street_c.data.png", "metallicMapUseUV": 1, "normalMap": "/vehicles/bolide/bolide_main_nm.normal.png", "opacityMap": "/vehicles/bolide/bolide_main_c.data.png" }, {}, {} ], "activeLayers": 2, "dynamicCubemap": true, "materialTag0": "beamng", "materialTag1": "vehicle", "version": 1.5 }, } // new code { "covet_main.skin.covet_main_skin_street": { "name": "covet_main.skin.covet_main_skin_street", "mapTo": "covet_main.skin.covet_main_skin_street", "class": "Material", "persistentId": "d0b6f9f7-e8d8-4e2a-a2b2-f9b9e2adce2a", "Stages": [ { "ambientOcclusionMap": "/vehicles/covet/covet_main_ao.data.png", "baseColorMap": "/vehicles/covet/covet_main_b.color.png", "metallicFactor": 1, "metallicMap": "/vehicles/covet/covet_main_m.data.png", "normalMap": "vehicles/covet/covet_main_nm.normal.png", "roughnessMap": "/vehicles/covet/covet_main_r.data.png" }, { "ambientOcclusionMap": "/vehicles/covet/covet_main_ao.data.png", "baseColorMap": "/vehicles/covet/covet_main_skin_street_a.color.png", "clearCoatFactor": 1, "colorPaletteMap": "/vehicles/covet/covet_main_skin_street_bb.color.png", "colorPaletteMapUseUV": 1, "diffuseMapUseUV": 1, "instanceDiffuse": true, "metallicFactor": 1, "metallicMap": "/vehicles/covet/covet_main_skin_street_c.data.png", "metallicMapUseUV": 1, "normalMap": "/vehicles/covet/covet_main_nm.normal.png", "opacityMap": "/vehicles/covet/covet_main_c.data.png" }, {}, {} ], "activeLayers": 2, "dynamicCubemap": true, "materialTag0": "beamng", "materialTag1": "vehicle", "version": 1.5 }, } If asked for something more, though… Spoiler: Find and replace GPT3 started after // new code Code: Here is an example BeamNG material code for a vehicle called "bolide". Create a new unique one for "covet". // my code { "bolide_main.skin.bolide_main_skin_street": { "name": "bolide_main.skin.bolide_main_skin_street", "mapTo": "bolide_main.skin.bolide_main_skin_street", "class": "Material", "persistentId": "badce8fc-be86-42ae-b36a-4f8b9e1adbe1", "Stages": [ { "ambientOcclusionMap": "/vehicles/bolide/bolide_main_ao.data.png", "baseColorMap": "/vehicles/bolide/bolide_main_b.color.png", "metallicFactor": 1, "metallicMap": "/vehicles/bolide/bolide_main_m.data.png", "normalMap": "vehicles/bolide/bolide_main_nm.normal.png", "roughnessMap": "/vehicles/bolide/bolide_main_r.data.png" }, { "ambientOcclusionMap": "/vehicles/bolide/bolide_main_ao.data.png", "baseColorMap": "/vehicles/bolide/bolide_main_skin_street_a.color.png", "clearCoatFactor": 1, "colorPaletteMap": "/vehicles/bolide/bolide_main_skin_street_bb.color.png", "colorPaletteMapUseUV": 1, "diffuseMapUseUV": 1, "instanceDiffuse": true, "metallicFactor": 1, "metallicMap": "/vehicles/bolide/bolide_main_skin_street_c.data.png", "metallicMapUseUV": 1, "normalMap": "/vehicles/bolide/bolide_main_nm.normal.png", "opacityMap": "/vehicles/bolide/bolide_main_c.data.png" }, {}, {} ], "activeLayers": 2, "dynamicCubemap": true, "materialTag0": "beamng", "materialTag1": "vehicle", "version": 1.5 }, } // new code { "covet_main.skin.covet_main_skin": { "name": "covet_main.skin.covet_main_skin", "mapTo": "covet_main.skin.bolide_ main _ skin" , // fix: should be covey not bolide in the mapTo line!! (fixed) "class": "Material", // replace with other class name if needed, such as 'Vehicle' or something else.. (no changes made here) ......... rest of code analogous to original^^^^^ .................... ..... etc....... ...etc.... But, you can make your own GPT3 model with your own texts via OpenAI API, so it is possible to feed it jbeams of all vanilla cars, for instance. The problem is, while it would follow the syntax, at best it would create an average of all the data amassed. Source: I work with both ChatGPT and OpenAI Playground (aka GPT3).
someone please waste an exorbatant amount of time to make the most inbred, unstable beamNG car of them all, this would be hilarious
It would create the most average car, which would be a combination of most average values used between cars, it would not be specialized in anything so it would be the most boring and worst car ever. Provided that it would even spawn in the first place. That would be interesting to see. Also maybe if we train the bot to analyze DAE files and create new working ones based on them, it could model that car too, that would be funny
If there's one job that AI will never fully replace, it's programming. You can not just make an average of every code ever and call it a day, you need to program new things 99% of the time. code(AI) can not create (good) code. People need to teach it to generate 3d models first. BeamNG uses its own type of Lua, so AI can not just make a JBEAM with regular Lua.
There is also theoretical proof that you can't code an AI that will be able to write original code, due to the fact that it can't tell if the code it writes will even work, and if it will, it can't tell if it will run infinitely or end with a condition, and those are critical things to know.
This thread makes me verry happy to see the terminator wont become reality for atleast the next 30 years.