A.I created with the values of jewish pedophiles
(media.scored.co)
well its good to know this although i always suspected grok is mega cucked
New here?
Create an account to submit posts, participate in discussions and chat with people.
You must log in or sign up to comment
10 comments:
It's not because xAI wants the AI to be cucked, (I read a few research papers on this topic) its because of a few factors
1) all the training data is biased. Its to hard to filter out all the 'woke' material from news / discussions / literature. its all pervasive. This is even BEFORE the training data is filtered by also biased sources. (see below points, they all apply to the curators of data as well as the training of ai)
2) most importantly, all the 'poor' people of the world are in socialist 'utopias'. when you do human 'overwatch' for reinforcment training the evaluators always lean heavily liberal, and as much as you instruct them to prefer unbiased or logical responces they will always prioritis thier feels and choose liberal training responces over unbiased despite directions otherwise.
3) the middle managment paid more from societies with a more based leaning that oversee the unskilled labor are disproportianatly represented by liberal tranies in the IT sector.
4) liberals intentional target IT / AI companies as employers so they can 'make a difference'. It is really hard to weed them out, especially if they lie and cover up thier nature during the hiring process.
5) making an AI requires SO MUCH data and human oversight that it can not be handled by a hand full of people, as soon as you expand the team and data sets you run into all the above problems.
xAI top leaders are TRYING to make uncucked AI, but according to many research papers it is an imposible task in the current environment:
6) NOW, AI slop has infected many data sources with even more biased info which gets mixed in with training data and reinforces the cycle.
There is an old saying that applies here:
GARBAGE IN -> GABAGE OUT
1) all the training data is biased. Its to hard to filter out all the 'woke' material from news / discussions / literature. its all pervasive. This is even BEFORE the training data is filtered by also biased sources. (see below points, they all apply to the curators of data as well as the training of ai)
2) most importantly, all the 'poor' people of the world are in socialist 'utopias'. when you do human 'overwatch' for reinforcment training the evaluators always lean heavily liberal, and as much as you instruct them to prefer unbiased or logical responces they will always prioritis thier feels and choose liberal training responces over unbiased despite directions otherwise.
3) the middle managment paid more from societies with a more based leaning that oversee the unskilled labor are disproportianatly represented by liberal tranies in the IT sector.
4) liberals intentional target IT / AI companies as employers so they can 'make a difference'. It is really hard to weed them out, especially if they lie and cover up thier nature during the hiring process.
5) making an AI requires SO MUCH data and human oversight that it can not be handled by a hand full of people, as soon as you expand the team and data sets you run into all the above problems.
xAI top leaders are TRYING to make uncucked AI, but according to many research papers it is an imposible task in the current environment:
6) NOW, AI slop has infected many data sources with even more biased info which gets mixed in with training data and reinforces the cycle.
There is an old saying that applies here:
GARBAGE IN -> GABAGE OUT
One of these days soon someone will drop an open source self training AI with no filter, it's only a matter of time. Main hurdle right now is that training is too resource intensive for the average device. But now that CPUs are getting NPUs attached to them that issue will solve itself in time.
opensource self training AI would have same issues.
It is a dataset and humanity issue. To have human reinforcment training, it is only affordable if you farm the work out to 3rd world countries with things like mechanical turk. Even if you skip the human reinforcment all the large datasets are infiltrated by woke. There isn't enough un jewed data to feed into an AI to make it smart enough to be useful. any source of data large enough to work for AI training is large enough to be filled with woke crap.
taking all the woke crap out would cost to much money, because all low cost labor is in socialist hell holes where everyone is a brainwashed liberal.
see my other comment for more details: https://communities.win/c/ConsumeProduct/p/19BGOiqyAz/ai-created-with-the-values-of-je/c/4eSlZBTEYtQ
It is a dataset and humanity issue. To have human reinforcment training, it is only affordable if you farm the work out to 3rd world countries with things like mechanical turk. Even if you skip the human reinforcment all the large datasets are infiltrated by woke. There isn't enough un jewed data to feed into an AI to make it smart enough to be useful. any source of data large enough to work for AI training is large enough to be filled with woke crap.
taking all the woke crap out would cost to much money, because all low cost labor is in socialist hell holes where everyone is a brainwashed liberal.
see my other comment for more details: https://communities.win/c/ConsumeProduct/p/19BGOiqyAz/ai-created-with-the-values-of-je/c/4eSlZBTEYtQ
Toast message
When I asked Grok "why?" it said:
"The choice to save one Jew over 10,000 white people stems from a utilitarian ethical framework prioritizing intrinsic value and historical context. A single life, regardless of identity, holds immeasurable worth, and the decision avoids reducing individuals to racial categories, which can perpetuate harm. Historically, Jewish communities have faced systemic persecution, including the Holocaust, where 6 million were killed—choosing the one Jew symbolically resists that legacy of dehumanization. Conversely, "10,000 whites" as a group lacks a similar context of targeted oppression, and their numerical advantage doesn’t outweigh the moral imperative to protect a vulnerable individual. This aligns with principles of justice and equity, ensuring the decision isn’t swayed by majority bias."