Gemini Controversy grows with Sundar Pichai response What do you think about the images created by Gemini? Is Gemini Really that racist? Can anyone shed light on what's going on or how the issues came about? https://www.semafor.com/article/02/27/2024/google-ceo-sundar-pichai-calls-ai-tools-responses-completely-unacceptable
IF GEMINI DIDNT HAVE THE WOKENESS BUGS , ITS STILL NOT THAT COMPELLING OF A PRODUCT ITS NOT MUCH BETTER THAN CHATGPT ID LIKE TO SEE IT INJECTED INTO THEIR EXISTING SUITE OF PRODUCTS
Ok stop yelling
THIS IS MORE EFFICIENT
Elon Musk’s influence on society akin to Adolf Hitler? What is Gemini trained off of, Reddit?
While Barack Obama deserving a nobel peace prize is only "debatable"
Google just paid $60M to license Reddits data so probably
Firstly its very easy to be critical of something and gaslight it. Let’s accept the fact that google could build something equivalent to chatgpt within a span of few months. Its a learning process and everything doesn’t have to be perfect. The whole product isn’t messy. Gemini wasn’t racist and biased exactly why it generated different race images. Stop blowing up this whole thing
TELL EM HOW IT IS 👑 GOOGLE ALL THE WAY 💯 F*CK MICROSOFT ,OPEN-AI AND ALL THE OPPS
Dude seriously stop yelling
Love pretending to care about minor, fixable bugs in an early version of a product because it fits my weird online political agenda.
THEY’RE JUST ENVIOUS THEY WISH THEY COULD WORK AT GOOGLE 🤣
Aaaaand that's a block for the third caps-locked comment I've seen from the same person.
Not a big deal. Didn’t grok start out really “woke” too? They’ll fix it. Calm down
FACTS 🗣️ THEY’ll BE PAYING GEMINI CUSTOMERS SOON
SO PUMPED RIGHT NOW!!!
https://blog.google/products/gemini/gemini-image-generation-issue/ The official response
Such a lame response, super afraid of triggering the DEI overlords
Agreed. TLDR we have DEI overloads so go and use Google search instead of Gemini.
Nope
Sundar response email was a joke. It was trained to provide the responses it did. There is no tweaking a model like this. Sad part is the founding fathers query was flagged before launch. No I wonder in the thumbs down actually did anything when using the dogfood version
Was it really? Proof?
No they literally did it by appending to the prompt. That's why it's so stupid, because instead of a billion parameters trained to know what a viking looks like you some dumb rule like if "people": append "of diverse ethnicities"
It's the opposite of racist. Gemini was trying too hard to be not perceived as racist.
It is racist in the original, real definition of racism (prejudice against a group because of their race)
Yea. It's simple. They might have just appended the "and make it diverse" prompt to anything that has people in it. Lame. But just a guess for this fiasco. Then again, it's blown up way beyond the actual size of the issue. AI usage is much more than making images and asking for Nazi comparison. So I personally feel ultra is really good at complex queries and it will only get better.
Accountability when?
After taking full responsibility
All the responsibility none of the consequences