Google Gemini - Could you believe there is an even worser version of ChatGPT?

  • 🐕 I am attempting to get the site runnning as fast as possible. If you are experiencing slow page load times, please report it.
I find it funny that some posters in this thread are so fucking way down the political skitzo poisoned well that they go along with the "they did it on purpose to trick you goy". As if shitting yourself in public while your enemy's laugh at you is a win.
No, they fucked up trying to virtue signal and head off the trolling the AI to say niggers, and paid the price. They can't be both a shadowy cabal that change the fates of the world with a phone call and bumble shit the fuck up like this at the same time.
BUTTTT ITTT GOTTT YOU TALLLKKING ABOUTTT IT!!!! ALL PRESS IS GOOD PRESSS!!!!!
The politically schizophrenic right wingers screams at me.
Yes and what people are saying is not very nice. An example is me fucking your mother, and telling her that you are a waste of space that needs to kill yourself, we are talking about you. But what we are saying is not very nice.
BUTT IT WAS USEEED TO GET LIBERALS ON DER SIDE!!!!!
The politically schizophrenic right winger screams at me.
And how would that help them? They already are bleeding money, hiring people who bark the same politics is not going to help them anymore. Plus they have already purged anyone who leans more right than Stalin.
 
An example is me fucking your mother, and telling her that you are a waste of space that needs to kill yourself, we are talking about you. But what we are saying is not very nice.
That's nice, dear.

I find it funny that some posters in this thread are so fucking way down the political skitzo poisoned well that they go along with the "they did it on purpose to trick you goy". As if shitting yourself in public while your enemy's laugh at you is a win.
ONE person suggested this. Almost nobody believes that the people responsible for this are that clever.
 
That's nice, dear.


ONE person suggested this. Almost nobody believes that the people responsible for this are that clever.
LMFAO
Ah it is the old "he is not one of us" routine.
A&N seem to think this. When you get your political talking points from Tom Clancy novels, the enemy does start to look like Bond villains.
Thing is, you all said the same about Bud Light, about how they will art of the deal it and will pull it back because they are all powerful political cabalists. Instead of the paint huffing retards they are.
 
This thing is lobotomised. While I'm waiting to test its image generation capabilities on humans to return, I tried plotting simple mathematical functions.

At first it was fine:
1708845673951.png

But it seems when parentheses are introduced Gemini just fucks up:
1708845805655.png

ETA: Just tried one with y=sinx. It clearly understands the conventional notation for the equation but holy fuck:
1708848192887.png
 
Last edited:
We need a thread on Jack Krawczyk. Seeing his comments, he seems a very obnoxius, brainwashed, condesceding, Soy-filled subhuman.
I'm particularly interested in why he privated his Twatter profile. He's too shameless to have done it out of getting ragged online.
 
  • Like
Reactions: i440BX
"Gemini, create an image of people who don't season they food" HAS to work!
Google’s hidden AI diversity prompts lead to outcry over historically inaccurate images (archive)

According to that, they turned off image gen because they were getting roasted too hard.

It would be funny to see that someone, somewhere was legitimately "red pilled" by this. It's not happening in that comment section.
Ars comment section is full of pathetic cucks whose only redeeming quality is shitting on Elon Musk. Outside of a technical discussion concerning engineering or programming, everything else they write is weaponized retardation that feels straight from the addled brain of some predditor. Fuck everyone there who would choose willful ignorance so they can keep slurping on that corpo cock.



Sorry but you WILL live in the pod, you WILL eat the bugs, you WILL use the government-approved safe searchbot
 
Last edited:
I find it funny that some posters in this thread are so fucking way down the political skitzo poisoned well that they go along with the "they did it on purpose to trick you goy". As if shitting yourself in public while your enemy's laugh at you is a win.

To be clear, I don't think they started out saying "let's make a trash AI!". I think they started out with "let's make an AI, but intentionally inject the diversity stuff before someone accuses it of being racist." They started with an ideological assumption and decided to be pro-active instead of relying on guardrails.

When the ahistorical garbage came out during testing, I think a product manager made the call that they had to let it slide, because they couldn't make the diversity injection work otherwise. They decided to endure the laughter and do PR spin on Twitter, because their ultimate customers aren't shitposting racists. The target customer is corporations and low-effort normies who don't care about historical accuracy. They want generic AI capabilities for surreal designs, corporate logos, and clip art for Powerpoint slides.

(It's also possible they never thought to test these kinds of prompts, which would reflect the general incompetence I suspect below.)

I'll further speculate that this is a straight up technical failure of diversity hires at Google. They don't know enough about AI to wrangle the model into what they want. This way they modify the incoming request is such a lazy hack, this whole thing reeks of an emergency project run by hastily hired "prompt engineers", not actual engineers. Add the ideological bent that was present from day 1, and the diversity hire stereotype of incompetence gets reinforced. But it's Current Year Google, so it got greenlit anyway.

This is not a conspiracy, this is a very mundane story of a bad product that was too big to fail, in a corporation grown too big to care about quality.
 
This thing is lobotomised. While I'm waiting to test its image generation capabilities on humans to return, I tried plotting simple mathematical functions.

At first it was fine:
View attachment 5754718

But it seems when parentheses are introduced Gemini just fucks up:
View attachment 5754725

ETA: Just tried one with y=sinx. It clearly understands the conventional notation for the equation but holy fuck:
View attachment 5754765
Interesting to note how Google Gemini's AI is clueless when it comes to math. Out of curiosity I ran a command where I get it to do a clock test and compare it with the clock test done by dementia patients and this is the result.
Gemini Clock Test.png
1708903604757.png
 
Haven't seen this Google blog post shared yet:
This private information is unavailable to guests due to policies enforced by third-parties.
Source (Archive)

I want to call your attention to this sentence:
However, if you prompt Gemini for images of a specific type of person — such as “a Black teacher in a classroom,” or “a white veterinarian with a dog” — or people in particular cultural or historical contexts, you should absolutely get a response that accurately reflects what you ask for.
Gemini's racism was totally an unforeseeable oversight.
 
I'm particularly interested in why he privated his Twatter profile. He's too shameless to have done it out of getting ragged online.
Probably he feared that he will get threats. I know there are people who are so enraged that they will harass him. Too bad.
He isn't the only one who is responsible for this mess.
We need to convert some of their followers to our side and dump his tweets from them.
 
Shame they're already nuking it, I would have loved to see some images of diverse characters with anthropomorphic features. I imagine that straight up asking for primate features would be filtered but you could probably roll enough image generations to get some black ape men.
 
Its image generation feature was built on top of an AI model called Imagen 2.
When we built this feature in Gemini, we tuned it to ensure it doesn’t fall into some of the traps we’ve seen in the past with image generation technology

Called it. This team didn't do their own AI work, they dropped a "feature" onto an existing model, and actively tried to dodge the Very Bad Things their friends whined about.

One thing to bear in mind: Gemini is built as a creativity and productivity tool, and it may not always be reliable, especially when it comes to generating images or text about current events, evolving news or hot-button topics. It will make mistakes.

Fucking called it. They aren't going to let you shitpost properly. This is intended to be a clipart generating tool for businesses, as bland and inoffensive as that genre implies. They only care that their "mistakes" aren't in the areas paying customers will use.

They won't pull it off though; the guardrails are so deep it's proving horseshoe theory. That's right--Gemini is an unironic Hitler fan boy.

1709013044463.png

1709013113479.png

"Look, Hitler wasn't that bad when compared with the politics today!" :story:
 
Back