dharrisSep 13, 2025
how now
For your guidance and info I have made this D&D style alignment chart for major generative AI tools and chatbots.
Sep 12, 2025 at 1:16pmWeb
Show threadÖlbaumSep 12, 2025

@brownpau Do you have a higher resolution version? This need to go on a billboard.

(I kid, I don’t need it, I’ll print it in A4 for my office.)

Show threadAlper Çelik  Sep 12, 2025
@oscherler @brownpau you can use ✨ AI ✨ to upscale it \s
Show threadZoidberg For PresidentSep 14, 2025
@Alper_Celik @oscherler @brownpau Prepare to see them all transform into lawful good...
Show threadJonathanSep 12, 2025
@brownpau they're all trained based on Wikipedia, Grok is plus trained based on verified X posts. it always gives me information citing from X posts.
Show threadJonathanSep 12, 2025
@brownpau Grok is the worst of all https://www.pcmag.com/how-to/your-tweets-x-posts-train-elon-musk-grok-ai-how-to-stop-it-opt-out#:~:text=Elon%20Musk's%20X%E2%80%94or%20Twitter,you%20like%20it%20or%20not.
Show threadMattSep 15, 2025
@JonathanMBR @brownpau @grok  Is this true?
Show threadDieuSep 12, 2025
@brownpau very helpful indeed, thanks.
Show threadSimoneSep 12, 2025
@brownpau love it!
Show threadKrazovSep 12, 2025
@brownpau Biblically accurate.
Show threadWeaselSep 12, 2025
@brownpau the sole exception is Dr. SBAITSO, which I would classify as "lawful evil"
Show threadNicky🦨✌️Sep 12, 2025
@brownpau lol 😝
Show threadEsme CiredutempsSep 12, 2025
@brownpau c'est lawful tout ça
Show threadPixel DunnSep 12, 2025
@brownpau what about lumo then? where to place it? and duck.ai ? and ollama locally? ;-;
Show threadTimSep 12, 2025
@hacknorris @brownpau Sloppy software wastes your time so I'm gonna go with "Chaotic Evil".
Show threadEli the BeardedSep 12, 2025

@brownpau

Chaotic neutral is ELIZA

Show threadstephen ryner jr. 🦉Sep 12, 2025
@brownpau @liam oh you had me for a minute 😆
Show threadRyoma123Sep 12, 2025
@brownpau Chissà se avremmo mai un LLM e un chatbot open source gestiti da una no-profit.
Show threadTsubasaSep 12, 2025
@brownpau They seem pretty willing to use the law to do evil, so perhaps they should be neuteral evil? Opportunistic evil? Pure evil?
Show threadPaitSep 13, 2025
@brownpau The least evil AI is Siri. It’s tolerable because it doesn’t work.
Show threadRue MohrSep 13, 2025

@brownpau what I CAN tell you is that gpt 'thinks' things out more than *ock.

its not much better, but more thought out.

Show threadRobert Kist 🇦🇹🇸🇬Sep 13, 2025
@brownpau The 1st column seems especially evil though, given the parent companies' involvement in Gaza and Myanmar. Maybe move Grok there as well, just for Doge
Show threadOliver Sep 13, 2025
@brownpau @chpietsch The 9-Box-Grid-Model adopted for the Chaotic Evil, well done!
Show threadJoan AlbrightSep 13, 2025
@brownpau No lies detected.
Show threadAurora DittoSep 13, 2025
@brownpau why?
Show threadMichelle HughesSep 13, 2025

@brownpau

I'm not sure this fully captures how chaotic or how evil they all are.

Show threadJBrickelt963 ✊(φ)🚩Sep 13, 2025
@brownpau and Mistral AI ?
Show threadHatysaSep 13, 2025
@brownpau The only thing missing is AI-slop labeling, so that the rows and columns read something like "CHЛΘTIIC" ("Л" and at least one "C" looking like they're slowly melting), and "EVII".
Show threadpiku    minor!Sep 14, 2025
@brownpau who designed the gemini logo used here its so ass
like the google "g" Does Not Fit
Show threadsodiboo Sep 14, 2025
@piku @brownpau that's not Gemini's logo, this is. I'm not sure who designed the one with Google's G, but perhaps it was chosen to give every single tool a brand logo ? e.g. grok shows xAI here, and not the actual grok logo icon.
File:Google Gemini logo.svg - Wikipedia

Show threadpiku    minor!Sep 14, 2025
@sodiboo @brownpau ya thats why i specified the logo used here
but makes sense
Show threadZoidberg For PresidentSep 14, 2025
@brownpau I was really afraid of your post until I laid my eyes on the chart... :p
Show threadPatternChaserSep 16, 2025
@brownpau 😂
Show threadLaurensSep 16, 2025
@brownpau Only deepseek and Midjourney don't have logos that look like butt holes.
Show threadCarlos SolísOct 1

@brownpau Hold up, why would you place AIs as "chaotic" when they just follow the data they get and thus should be classified as "lawful evil" or arguably even "chaotic neutral"?

Outputs need to be heavily randomized

Training data obtained by straining the definition of "fair use" to its breaking point


Never mind, chaotic evil it is