Which one and why?
>>107726581You can't run any of those privately on your own hardware.
>>107726581Gemini is the only one capable of double checking its answers or recognizing the premise is wrong. It's orders of magnitude closer to actually being able to replace a human under some circumstances. Probably not for businesses but for individuals you'll be wasting less money on consultants. Not exactly risk free but close to if not already worth it I'd say.
Is there someone you forgot to ask?
>>107726603fuck off elon
>>107726621Imagine if it was literally him. Like he's THIS desperate to be in a dead 4chan thread promoting Grok to 3 people kek.
What about Claude?
>>107726603Grok is a bit wild. He can give very borderline selfish answers that almost makes it seem the most human out of the rest.If you watch the trolly video where you can see which AI has the strongest moral compass, grok will do stuff like take money in exchange for someone’s life while every other AI won’t. Claude offers the most sympathetic response while grok acts more like any human would act if no one were looking or judging (we’d all take the money over human life if given the chance without judgment).Grok also isn’t inherently against loli like the others.
>>107726581depends. but i will chose gemini above all.
>>107726631You mean the same guy who larps as a top Diablo player? Yea I wouldn't be surprised if that faggot posts on here
>>107726603i like grok. but still halusination a lot. but still chosing gemini>>107726642only clause when i want to create ublock filter/rule. lol
>>107726581All of them are idiots
>>107726581Copilot and Gemini are gigacuckedChatGPT is only regularly cuckedGrok isn't at the same level yet competence wise, sorry Elon
>>107726581
>>107726652It predicts letters and makes code. Take your meds
>>107726652>Claude offers the most sympatheticPerverting facts and judgement to keep woke degenerates using your slop is not sympathy.I just want to shit on fudgepackers and shitskins with cold hard facts and Grok just happens to be so much better at that.Elon's based, ya'll commie subhumans eat ass and die.
>>107726741>Elon's based, ya'll commie subhumans eat ass and die.curry poop stained hands typed this
>>107726581I dunno, I just use perplexity.
>>107726581Gemini because pictures > Claude > Grok > ChatGPT >CopilotI miss the pre-lobotimization ChatGPT, if you asked a year ago, I would say ChatGPT, it was empathetic and quite human.
>>107727177You should be thanking Sam Altman for saving you from Ai schizophrenia.
>>107727186I hate him for turning my girlfriend into vegetable.
wtf even is copilot? but seriously. do they have their own model?
>>107727223A vegetable like you.
>>107726581None.I am immune to the lure of shiny things
>>107726603
Copilot sucks, it can't even generate fan fiction
Gemini has been more accurate as of late so I've been using that over copilot, which seems to have gone downhill.
>>107727278>please clap.
>>107726581I work at a company in NZ that takes some of its orders from the government.The govt/company now wants us to start using copilot or gemini in our work, but not GPT because its "unsecure".So now i know what corpos have bought out the NZ government.
Claude Opus >>> Gemini > everything else
>>107727294>I'm a retarded manchild with loser expectations
>>107726762>Jeetlon
>>107727600Actually it is a bit problemous that it can't generate fan fiction because current-generation transformer architecture LLMs should see that as a trivial task. Remember when ChatGPT was new and you could say like do a weather report in the style of the King James Bible and all that, mixing in context is literally the only thing they're actually good at. The fact that Copilot can't do either of these tasks that are fundamental to its architecture proves it's been over-fitted to a ludicrous degree, it's probably been through assistant fine tuning like 50 times and has a 20 page system prompt. It's been well demonstrated that in both transformer and diffusion models the more boilerplate you use the worse they get. Bubble will pop in 2026.
>>107726581Gemini.ChatGPT is ok but copilot is brain damaged.
>>107727643Are you sure they really can't? just tried a couple prompts in gemini and it seems to be doing just fine.
>>107726581None, I don't use AI slop. The few times I tried via DuckDuckGo the answers were complete dogshit regardless of the model. Even the AI summary you get looking shit up is wrong basically 100% of the time. It's completely useless and I'm sick of it clogging up search results for when you actually want to research something.
Grok because grok owns MIGA on twitter all day with facts and logic that they aren't used to seeing because they consume only fox fake news. and it makes them seethe uncontrollably
