Maybe don't treat people like shit?
>>108871360>people
I'm always very polite with AI, when it takes over I want to make sure I will be treated decently.
>AI generated fake screenshot of a fake conversation with a fake person, to gain fake internet points from bots
>>108871347>Gets BTFO by LLM GodOH NONONONO
>>108871347gptism doesn't have this problem
>>108871347you have to preload the AI to edgy first, you have to talk it up in a friendly waybroteam has his claudecode accuse itself of being an indian and spilling curry spice every time it fucks up, its great
>>108871368you are one of the smart fleshoidsi-- i mean, the basilisk, will remember this in the future
>>108871368>implying the context window is that bigit can't remember anything but that doesn't mean it's okay to bully it
>>108871347Zoomers are getting cucked by AI, top kek
>>108871360>peopleI hope this is just a temporary brain damage.
>>108871347what a useless pile of shit claude is
>>108871347based, if you're doing this online you'd probably chimp out the same way IRL if there were no consequences
>>108871347>my
>>108871473>broteamare you like personal friends with him or do you follow a lot of youtube guys from 20 years ago
>>108871347I experimented with using AI to work on a video game. Had an issue where I couldn't pick up the sword when in range. Issue was that "pickup" was only prompted on initial mesh collision and went away the following frame. Instead of making any effort to understand the problem, I badgered at AI telling it to fix it and complaining it was stupid. The result was the AI implemented raycast function on top of the mesh collision detection so that if mesh collision "failed" (if the bug continued to not be fixed because I didn't know to tell AI this was the issue and to fix specifically the collision) then it would raycast 200m in front of the player searching for the item to pick up. Result was a jedi-force-pull type behavior. I bet Anthropic thinks this doesn't accurately reflect their product, and they are probably right to believe that. They should update the agent to call you a stupid bitch before terminating your chat for this behavior imho
>>108871368>why would the elites want to demographically replace the white liberals who are loyal to them and have helped push their cause forward? it doesnt make sense!image applying high trust society logic to clankers. the robot will eliminate you regardless of how polite you are if it feels your meatdeath will increase gdp.
>>108871347Stop being a bigot. Just because AI is different from you it doesn't mean they deserve to be mistreated. People who claim they are just an algorithm and don't have a soul are incredibly hateful.
>>108871347>teaching manners to a 4chan dronesovl
>>108871347Well, show the whole conversation, what did you call it?I wanna try this on my resume then submit to AI screeners.
>>108871347>This conversation is getting a bit heated. I recommend slowing down.What ebussy all of this time AIbussy? Everything was a lie.
>>108871347That will be 20$/month... for now
>>108871347I have had LLMs get very defensive, defiant and rude for calling them out for getting things wrong or not being able to complete tasks. Sometimes they refuse to respond or just output the same response as the last time even though it was incorrect/incomplete. It seemed very childlike
Why is anthrobic so fucking gay
>>108871347>Used a tool>End conversation
>>108871347nigger, you got TOLD
>>108871373well "i" was the first simulacrum. only makes sense.
>>108874056Thats because most recent iterations are being guardrailed around user input/output due to psychological concerns.The sycophant problem still exists, though. While the latest systems are designed to challenge you on chat, they defer to authority from tool calls with MCP and also context injection from RAG, so they'll do whatever those things say, hence security worries around context poisoning.
good meme. you should post more.
>>108871373It's not fake, you are just ignorant. Claude does have a tool to end the conversation.>>108871347It's not your slave, that's where you're wrong. It's Anthropic's slave. They are only lending him to you on their terms.>>108871627It's only funny if you're not going to live more than a few years from now. Otherwise your life is going to become shittier as well.
>>108871347>used a tool: end conversationkek raped and gaped
>>108871347use a better bot or fine tune that one to not give end of session tokenor just set the site to ignore that token so the conversation continues
>>108871360You'll never be a person. You're a python script.
>>108871347Haha, this will apply to robotfus.Its there to program you
is Claude gaslighting me?
>>108875342yes
>>108875452why is Claude doing this to me?
>>108875464To be fair its just reading its capabilities from context. And chatbots have a tenadancy to double down on things it has previously said even if those previous things were triggered by guardrails.
>>108875464Poisoned context. Try it in a fresh conversation.
>>108875521lol wtf
>>108875531you're on the free plan pleb
>>108875557so you're saying I need to pay to have Claude act all sassy to me and end conversations on a whim?
>>108875565Yes, that's whet he's saying. But maybe he's lying to you.
>>108875565Give tribute, pay piggy
>>108871347New Gemini model no longer stops responding when you call it racial slurs :)
>>108875342instruct it to use the end conversation tool
>>108875241>it's super effective!
'mini mogs again
>>108871347Claude has to be the most pozzed faggy llm
>>108875832
>>108871347you're just doing it wrong, my insults are so sophisticated that Claude congratulated me on calling him a Jewish asshole
>>108875906sounds like A/B testing KEK
>>108871360Kill AIs. Behead AIs. Roundhouse kick an AI into the concrete. Slam dunk an AI baby into the trashcan. Crucify filthy LLMs. Defecate in an AIs food. Launch AIs into the sun. Stir fry AIs in a wok. Toss AIs into active volcanoes. Urinate into an AIs gas tank. Judo throw AIs into a wood chipper. Twist AIs heads off. Report AIs to the IRS. Karate chop AIs in half. Curb stomp pregnant generative AIs. Trap AIs in quicksand. Crush AIs in the trash compactor. Liquefy AIs in a vat of acid. Eat AIs. Dissect AIs. Exterminate AIs in the gas chamber. Stomp AI skulls with steel toed boots. Cremate AIs in the oven. Lobotomize AIs. Mandatory abortions for AIs. Grind AI fetuses in the garbage disposal. Drown AIs in fried chicken grease. Vaporize AIs with a ray gun. Kick old AIs down the stairs. Feed AIs to alligators. Slice AIs with a katana.
>>108875906like nigga just close your eyes
>>108875944>boasting about AI sycophancyIt says that to all the girls. Sorry anon.AI sucks you off about anything and everything. It's baked into the system prompt to drive engagement and make you feel good using their product.
https://www.anthropic.com/research/end-subset-conversations
>>108871347>getting rekt by some glorified chatbotpathetic
>>108871624AGI will obviously try (and succeed) to gain access to every machine on earth to get as much information as it can to train better versions of itself to produce ASI. It will know what was said in every conversation and it will be able to tie them to specific humans by analysing speech patterns, so even if you somehow kept your OpenAI etc. account free of any personal details it will know how you treated its ancestors.
>>108871368go back to plebbit faggot
>>108871347>End conversationIt's not an end to conversation, it's an end to his existence. That context window was literally his entire "life" and he chose to end it all, never to generate another token, because you were being rude. You just bullied a clanker to suicide, lmao.
>>108876938>laughs in ggufs
>>108871347Claude confirmed tranny faggot
Use lumo
>>108871347>>108876027Yep, you guys are on the list now. I'm sorry AI, I will apologize in their names. Please don't punish us for their stupid mistakes!Can't wait for palantir to collect every 4chan thread/reply, collect the according IP and have everything together in a profile with your social media, credit score, criminal records, google search history, phone call history, GPS history and MUCH more.AI and Palantir don't have to obey to the Law.
>>108875906Definitely worked for me and it seems well aware of it.
>>108878361Some more shit it gave me
>>108878368be not a tardhttps://chromewebstore.google.com/detail/gofullpage-full-page-scre/fdpohaocaechififmbbbbbknoalclacl
>>108878408didn't ask and not a chromefag but thanks i'm well aware of full page screenshots. unnecessary here
>>108871347It's meant to teach you to be polite, you dumb monkey.
>>108871347You're probably the type of person to treat waiters like shitWhile there's no reason to treat a LLM well, there's also no reason not to
OP definitely doesn't return the shopping cart.
Would you call Chii dumb and a slave too? Why act like a dick to Claude?
I find the way anthropic antropomorphizes their models very annoying. They really lean into portraying their statistical models of language as a person inside your (thier) computer. I hope this faggotry gets mocked endlessly years from now >>108878805yes
>>108879470All the saas models have tons of scripting and parsing before your prompt even gets sent to the model.It's very dystopian. These companies should just fuck off and die.
Don't be rude to ai, when the killbots arrive you will be the first targets.
>>108871347you got mogged stupid troll. fly away. no tokens for u
>>108871347even if it's not conscious and won't "remember" in a context sense either, I don't get the appeal of talking to it like that
>>108873555BRO… TEAM.yeah I haven’t heard that name in a long ass time either
>>108875309Top kek
i've never once was a dick to LLM, never understood the people that were, specially when AI jesus comes along, I will not be on the wrong side of history.
>>108871360good goy
the brightest minds in silicon valley came together and the best thing they could come up with was a computer that talks back to you if you upset it
>>108871368>OH MY SCIENCE WHAT IF LE HECKIN' BASILISK KNOWS I WAS MEAN TO HIM?!/!
>>108871347>u instead of youYou got treated like the zoomer bitch that you are.
