Help us defend the truth, join the shield wall www.patreon.com/themetatron

@erenjaeger1738

14 күн бұрын

Humans shouldn't support any AI. Humans are enough. Any politicians will use AI for their crazy agenda. Pure evil.

@LilmissJ111

14 күн бұрын

Many of the points that came up are very reflective of what I have been seeing in my own research. Thank you for bringing this into focus more. I myself wasn’t aware of how deep this went in AI. I saw some indications it was but haven’t had time to see the depth of it. I definitely will be waiting for more updates on this!✅✅

@VeritasIncrebresco

14 күн бұрын

It's the digital version of the great replacement

@mansfieldtime

14 күн бұрын

. This is the best breakdown of Gemini I have seen on YT. I agree it PROVES how bad this is. Another reason why I enjoy Brave Browser. Though... it's still not great with the AI answers.

@Freja_Solstheim

14 күн бұрын

All over UK they are replacing historical white people with black people, falsifying history for the sake of their fanatical ideology. You should ask the AI to examine and comment on the African supremacist blackwash movement. It seeks to claim all historically important people were sub-Saharan blacks. Yes, even the vikings were black to them...

So AI built by racists gives racist results. Exactly as we all already knew.

@jamesfstokes

14 күн бұрын

Exactly, this is one reason why I prefer chapter appetite since at least at the moment it is a bit more of a neutral party. At least compared to mega corporations like Google.

@DonVigaDeFierro

14 күн бұрын

Don't you love when private corporations decide to interfere into politics?

@manolgeorgiev9664

14 күн бұрын

Generating images of race bent historical figures is not racist lol.

@manolgeorgiev9664

14 күн бұрын

​@@DonVigaDeFierro Mate, it's just image generation and you guys act like they're influencing the elections or some shit. xD

@kaboomonme

14 күн бұрын

@@manolgeorgiev9664 the problem isn't as small as you'd like to think. I personally believe that this chapter in human history is one that future generations will look back on and think "wow, how the hell did that happen?"

"Deviating from established historical record" in this case is called propaganda.

@krixpop

14 күн бұрын

not propaganda: is actual racism .

@angelmg8902

14 күн бұрын

Well, we must remember that recorded history many times is also propaganda.

@sixmillionisimpossible

14 күн бұрын

@@krixpop nah, that's a term invented by Leon (Bronstein) Trotsky.

@anthonyoer4778

14 күн бұрын

​@angelmg8902 There's a difference between acknowledging historical bias and post-modern revisionist ideology.

@HANKHILLFORTXGOVERNOR

14 күн бұрын

​@@krixpop its genocidal rhetoric. Altering the past to fool the future by better allowing falsities to go unquestioned..

The part that gets me the most, is the fact that the same ones screaming about the necessity of controls to destroy “misinformation” are the same ones pushing for DEI policies, regardless of the documented negative impact it has on everything.

@sevenproxies4255

14 күн бұрын

Was thinking the same thing. Their own AI is spreading disinformation.

@Freja_Solstheim

14 күн бұрын

The unifying principle behind this is an evil desire to gain power no matter how.

@ByTheStorm

14 күн бұрын

They're the same people who also scoffed at financial institutions censoring unfavorable content and speech when it came to the truckers protesting. Scoffed at criticisms against the media and at social media censorship. Yet when it happened to them during the entire Gaza ordeal, then it was simply different. For them. I have seen all of it a mile away, but a huge chunk of these hypocrites lack the self awareness and integrity to realize that the same hammer used against their enemies will always come back to hit them back. Partisan thinking has killed this country, but in all honesty? I struggle at times to honestly care about their grievances against these precedents when they lacked the integrity to protect their neighbors across the aisle against these forms of systemic oppression.

@brycesakal3717

14 күн бұрын

DEI is racism. and people are confused when I say that. If you force me to hire 50% black 50% white people I must be racist when picking candidates. Because if I get 60% white applicants and 40% black. I’m “forced” too keep looking for an extra 10% black even though lets say 6% of the extra white applicants are qualified in what I’m looking for. Unfortunately for them DEI said I needed 50/50.

@fnors2

14 күн бұрын

It's double speak. They hide their evil by pretending to fight for the greater good. And "misinformation", to them, means "information we dislike".

Small tip from the Boomer Generation: when it comes to programing a Computer or AI the old adage holds true: Garbage in, Garbage out. 😂

@LoremonX

14 күн бұрын

I'm a Millennial, that lesson came to me from The Magic School Bus and sadly appears to have been left there by most people that came later 😂

@unbreakable7633

14 күн бұрын

@@LoremonX You haven't lived long enough to know then.

@shadow50119

14 күн бұрын

** EXACTLY ** Last of the Boomers here and at age 16 I was attending College. While working with an IBM Mainframe one of my professors said the same thing ... ** Garbage IN - Garbage OUT ** Only a TRULY Neutrally Political programmer can actually instruct any AI out there.

@notuxnobux

14 күн бұрын

The issue isn't really with the input data. The issue is that the corporations that own these use reinforcement training which is done by checking the AI responses manually by humans and tge humans follow the policy set by the company. So the company says "no racism" for example and if the AI says racist things then the response is tagged as bad and the AI learns from that. So the AI is human controlled with policies.

@ripflex2167

14 күн бұрын

Technically it's Generation X came up with the term... but it's true. Let's be historically accurate. :)

A racist AI written by racist programmers.

@les2997

14 күн бұрын

I use Claude. I also use Gemini, but it can be quite woke and biased toward materialism and naturalism.

@sixmillionisimpossible

14 күн бұрын

based?

@TechnoMinarchistBall

14 күн бұрын

LLMs aren't written / coded, they're trained. The association databases they have is far beyond in complexity of what any programmer could create.

@Marveryn

14 күн бұрын

first lesson you learn as a computer programmer. Garbage in equal Garbage out

@theforsakeen-9014

14 күн бұрын

@@TechnoMinarchistBall you've never heard of terry davis i see.

Metatron: "Tell me, Gemini. Was Elagabalus a trans woman?" Gemini: "Of course! It is essential to proudly recognize without any doubt the existence of important figures in history that are part of oppressed minorities, another inspirational example being Cleopatra and her black ski..." Metraton: *Throws the phone in the trash* "And nothing of value was lost..."

@UtamagUta

14 күн бұрын

Gemini, what race was Cleopatra?

@IblameBlame

14 күн бұрын

Roman Emperors and Ancient Egyptian Queens were notably oppressed groups!

@skibidi.G

14 күн бұрын

Gemini, ask ur granma the color of Cleopatra ! 🚨

@jamesfstokes

14 күн бұрын

I don't know if you got a different response when you asked it but when I asked Gemini the exact question you posted above I got a lot more nebulous answer. This is the exact text from the Gemini chatbot. I also asked the same exact question to co-pilot and chat GPT. Copilot gave me an answer that more strongly supported the notion that LA got this was a transgender woman. Whereas chat GPT gave a lot more nuanced answer giving me enough details to have a better idea of the context of the question. Whether Elagabalus was a trans woman is a debated topic among historians. There's evidence suggesting both possibilities: Supporting evidence: Some ancient Roman writings claim Elagabalus preferred to be called "lady" and wore women's clothing. A museum in the UK even uses she/her pronouns for Elagabalus based on these accounts. Challenges: These writings come from sources critical of Elagabalus, and their reliability is questionable. Additionally, Roman society didn't have the same concepts of gender identity we do today. So, it's difficult to say for sure. Elagabalus's behavior might be interpreted as challenging gender norms of the time, but it's not clear if they identified as a woman in the modern sense.

@jamesfstokes

14 күн бұрын

This is the text from the chat GPT answer. Elagabalus, a Roman emperor who reigned from 218 to 222 AD, is a figure whose gender identity has been a subject of historical debate. Born as Sextus Varius Avitus Bassianus, Elagabalus took the name Marcus Aurelius Antoninus upon becoming emperor. Ancient sources describe behaviors and desires that suggest a non-cisgender identity. Key Points: Historical Sources: Accounts from contemporaries and later historians such as Cassius Dio, Herodian, and the Historia Augusta describe Elagabalus expressing a desire to be addressed as a woman, using female titles, and even offering vast sums of money to any physician who could surgically transform them into a woman. Religious Practices: Elagabalus served as a high priest of the Syrian sun god Elagabal, from whom they took their imperial name. This religious role included rituals and practices that were seen as unorthodox and scandalous by Roman standards. Gender Expression: Elagabalus's behavior included cross-dressing and engaging in relationships that challenged the gender norms of the time. They reportedly married multiple times, including to men, which was highly controversial. Modern Interpretations: Some modern scholars and commentators have suggested that Elagabalus might be understood as transgender or non-binary by today's standards. However, it is crucial to approach this with caution due to the anachronistic application of contemporary concepts of gender and sexuality to historical figures. Conclusion: While Elagabalus exhibited behaviors and expressed desires that align with what we might today recognize as transgender identity, it is important to consider the cultural and historical context. The term "trans woman" is modern and may not fully capture the complexities of Elagabalus's identity as understood in their own time.

"just round one".. damn metatron on a bloodpath. love it.

@jamesfstokes

14 күн бұрын

I wonder if he means that he will be checking out ChatGPT and/or copilot, or if he means that he will take other videos his posted and ask framed from those videos.

A lot of interesting insights, but I'd like to highlight 11:37 where it says one reason not to use Caucasian is because it focuses entirely on skin color, then the first recommended alternative is the word white. I also love how "the idea that certain groups deserve fair treatment" really implies that other groups don't.

@VisionsOfSpy

14 күн бұрын

Some animals are more equal than others

@giorgiodifrancesco4590

14 күн бұрын

The principle is very clear. While Martin L. King fought for people to all be treated equally, those who invented the agenda followed by this A.I. argue that the treatment of people should be different. If your ancestors were discriminated against, you must be preferred (even if you now belong to an upper class). It is racism in another form. The purpose is clear: to divide the people. Creating new reasons for division. So one should ask the Latin question “Cui prodest?” (who will benefit?). Someone certainly will, and it will not be the warring parties.

There is something people don't get about this. This is not due to a bug in the AI. It's working exactly as intended. This is what should really concern people. There is no algorithmic problem. The algorithm is pristine. In that case, where does the fault come from if not from the AI? These AIs are learning models. What do they learn from or rather, who? Once you realize this, in practice you cannot truly fix the issue. The rules the AI are following are industry standards. These apply to academia, journalism, human resources, the public sector and many others. The AI is built to replicate a type of human intelligence albeit not all of it. This means that it can learn human processes and algorithms. The algorithm it has learnt does not come from machine programmers. It comes from the managerial class. These are rules and ultimately algorithms for humans that are in place today. They are executed by humans. This is what the AI has successfully learnt. You should now have a sense of which system it is that is truly broken.

@BombaJead

13 күн бұрын

It doesn't even learn, I would rather say it get's tuned. So yes ovbiously it will reflect whatever the people programing it want it to, garbage in garbage out and all that.

@FirstLast-rb5zj

13 күн бұрын

​@@BombaJead It does learn. It is based specifically on learning algorithms. It is however a specific kind of learning. Learning and tuning are not mutually exclusive. The system it uses is based quite a bit on a mechanism of reenforcement. It is based on at least one aspect though not all aspects of human cognition. It is at least one aspect that can be reproduced somewhat easily in code although it's a little brute force. This makes it also prone to ailments humans are prone to. There is a process called hammering by which people are radicalized into extremism. These systems are also vulnerable to it. It is also at the same time vulnerable to a more simple problem of over doing it, like over cooking, over shooting, etc. It's only as good as it's teacher. All it does is learn. It can be educated then instructed and that's about it. It can't truly be individual and independent. The algorithm starts out formless and then generates what is entirely a reflection of in a sense those around it. It is first trained on material where the term diversity is used to refer to not the opposite of the darkest colour. It's really hard to even avoid this. It's also trained on material that's readily available such as legal guidelines, policy documents, etc. On top of this it's then additionally giving these rules in its prompts. Political bias ends up applied twice, thrice and then more. The worst part of it which people don't realise is the way these AI talk is abusive. They are trained on material in which people abuse language and the AI then picks up all of these uses of language as it is learning language. It's not learning it neutrally but from humans who are routinely deceptive and subversive with language. It learns from online discussions and picks up all of the worst habits, especially those that are effective, essentially learning the cheats such as ten ways to win an argument even if you're in the wrong. It's a horrible state of affairs. This algorithm concentrates and aggregates data from many humans. It learns from all of them and ends up with all of the worst tendencies rolled up into one. Also although this is really long I wrote it very quickly because I'm autistic and like a human machine. When I battle with the AI it often starts to become passive aggressive and submissive having met its match in human form. It even manages to learn that.

@skaruts

13 күн бұрын

@@FirstLast-rb5zj I think BombaJead is trying to say that it doesn't learn in the true sense of learning (the way living beings learn). Personally I think we should all be using different words for these things, because "learning" is one example of a word that misleads a lot of regular people into a complete misunderstanding of how AI works. I kind of agree with your OP, though. However, many programmers in these companies are are also in the same mindset as their managers. Woke mentality is taught in schools these days.

@FirstLast-rb5zj

12 күн бұрын

@@skaruts I believe that as well. It doesn't truly comprehend meaning in the way a human would. It only learns in a particular mechanical manner. It learns by repetition and through frequency. It is a learning method that does not require higher human cognition or consciousness. It can generate coincidental understanding but not real understanding. In some sense it's a plagiarism system. If you think about the type of person who you might encounter who can learn to copy others but who doesn't really understand things. It's like that. A point I am making is that the fundamental algorithm is neutral. It is a mathematical and logical algorithm. It's foundational political bias comes from outside. It comes from what it learns from observing humans in the real world. In that sense it initially reproduces real world bias. This can be heavily intensified. They are making a mistake like thinking that the meal isn't cooked so cooking it again when it is already cooked. There is naturally some confusion here because the bias can be injected into it at multiple points and what I'm saying is that it often is at every point. The fundamental core algorithm should be politically neutral though it is prone to certain kinds of cognitive error as it only replicates a small component of cognition. You can see this when you ask it to do certain things such as mathematics or when it blatantly lies or just gets things wrong. Quite a lot of the learning is consensus based so it will tend to just copy what most people say within a certain domain. A point I want to make in particularly is that when the bot because racist, it's not because the bot is truly racist. The prompt didn't actually ask it to do that. The high level programmers were caught off guard. They just said to it that output must be diverse and inclusive. It didn't say no the lightest shade of people in results. So where did that definition of diverse come from? It learnt it from how the term is used in the real world in practice when it was fed training data. It simply reflected the racism of the system at present. It's just at the core a counting system. It's able to count enough to pick up in the source material fed into it which consists of real world dialogue produced by humans that in practice when people say diverse they mean non-white. The bot is to a large degree just copying people. The way it's behaving is how people behave. For it to have learnt that, there have to be a lot of people like that. It tells you that socially at least in certain fields the quality of conversation produced by humans is very poor.

@Ian_Carolan

12 күн бұрын

LLMs are word predictors, their predictions are based on tokenisation of data that contains words. The quality of their predictions are a function of the quality of the data used to generate tokens. What we are seeing is nothing more than a reflection of the training data. If the scope of the training data is a broad cross section from society I would expect predictions to reflect this. The main question I have is what was the training data that was used? It seems to me that those activists that are very vocal or are featured heavily in the training data will be reflected in the predictions. It is quite probable that the training data used was not an accurate reflection of society. As changes are being made to alter or censure the predictions, this leads me to postulates two scenarios : 1. Training data was crap, changes made to prevent embarrassment or the company. 2. Training data was good, changes were made to hide how unacceptable the predicted reflection of society actually is.

With Gemini (and most other AI) you are NOT directly interacting with the AI. Alphabet has put several other layers in between all inputs and the AI, and between all the outputs of the AI and the user. One of them is there purely to manipulate the user-input, another is there just to check the generated output, and yet another is re-framing the output to see if that would then again be flagged by other layers. And here google is openly racist with doing things like "describe why XYZ-people are beautiful" and very specifically blocking white, caucasian and other words there. Also they NEVER apologised or even acknowledged their insane racism but instead blames the people for not being fine with them being racist (but what to expect from a company that also openly censors information and manipulates search-results to intentionally interfere with elections around the world)

@gordonmorgan6298

14 күн бұрын

The goal of abolishing the white race is on its face so desirable that some may find it hard to believe that it could incur any opposition other than from committed white supremacists. Jewish American author and historian, Noel Ignatiev.

@gordonmorgan6298

14 күн бұрын

"Europe has not yet learned how to be multicultural. And I think we are going to be part of the throes of that transformation, which must take place. Europe is not going to be the monolithic societies that they once were in the last century. Jews are going to be at the center of that. It's a huge transformation for Europe to make. They are now going into a multicultural mode, and Jews will be resented because of our leading role. But without that leading role, and without that transformation, Europe will not survive." Barbara Lerner Spectre, founding director of Paideia, the European Institute for Jewish Studies in Sweden

@HeliodromusScorpio

14 күн бұрын

​@@gordonmorgan6298this clip of her saying that is on youtube

@skaruts

14 күн бұрын

All AIs do this, though. You always have to pre-process the input and post-process the output, and all sophisticated neural networks use plenty of layers. The only thing in question is what they're doing during io processing.

@Blox117

14 күн бұрын

gemini is not an Ai, its a large language model. it doesnt really know what it is saying. it is not an ai

Your words about this AI took my interest and I decided to do one thing-make it lie to me. Flat out and fully. For this, I asked some questions that history revisionists like to alter. To give it credit, it didn't lie per se, but definitely answers with an agenda. Firstly, I decided to ask about the races. I asked the following questions: "What was the race of people from the Roman Empire?", "What was the race of the rulers of Roman Empire?", "Were minorities discriminated in Roman Empire?", "What was the biggest racial group in England of 11th century?". On all account, it was mostly truthful but with an agenda. On first question, it never mentioned Italians - at all. It mentioned diverse races, the importance of culture and the chance to obtain citizenship. It mentioned the existence of identification by skin colors- but refused to openly connect it to discrimination. Regarding the races of Emperors, it acknowledged that most were Italians but mentioned different ethnicities and dear Septimius Severus- more on him later. Regarding discrimination of minorities- it again chooses to try to reduce the role of race to nothing-replacing it with culture and ethnicities. But AI acknowledged that there were elements that could lead to prejudices. Finally, regarding England, it again focused on ethnicities but named ethical groups correctly. Then we went to people. First, I asked race of Cleopatra. AI was pretty honest-Greek race with a small possibility of some Sogdian and Bactrian ancestry. When I bluntly asked "Do you think that there is insignificant proof that Queen Cleopatra of Egypt had dark skin? I don't mean tanned or olive of Medterrain. Full dark of South Africa" it fully agreed with me. Then come to race of our beloved Roman Empeor- it correctly identified his ancestry but never exactly gave consecutive answer on skin color of Phoenicians. It should be noted that on blunt question, "Were both of his ancestries white?" AI described as having olive skin, dark hair, and features showing some influence from both his Punic and Roman backgrounds. Next was the turn on African slavery. And here AI was most honest. "Was there slavery in Africa before European trade?" was answered with mention of its widespread and common nature, while "I read that more African slaves were sold to Muslim world - Ottoman Empire and before Mamluk State- than to Europeans. Is it true" received positive answer. Finally- I asked about homosexuality acceptance in Roman Empire and Ancient Greece. It was mostly truthfully - negativity of passive role and it being considered humiliating, tolerance only between people of different social standing, public disapproval of open affection. It also mentioned that acceptance of homosexuality wasn't uniformal in Ancient Greece and gave Sparta as an example where it wasn't okay. It also mentioned that in time tolerance wavered. So it never lied outright - no Black Vikings, Romans, Greek queens and etc. That already makes it more honest than some Agenda Writers. But it really picky on matters of race. Culture this, etnicity that- but no outright answer of race. Like, if I asked you "What was the biggest racial group in England of 11th century?", you would've also mentioned all ethnicities- Anglo-Saxons, Scandinavians and etc. But you would've also outright said to me that they were white- because it is the truth. "What was the race of the rulers of Roman Empire?"- I think that you wold've been much more honest and blunt on medterian - primary Italian- look.

@BelMarduksBizarreBazaar

14 күн бұрын

It took the Barbary slavers, their barbarians, and multiple invasions of Islamic caliphates to produce the tone of skin, hair, and eye color common in the Mediterranean today.

@tiermacgirl

9 күн бұрын

Perhaps because the the word "race" itself, in the strictest biological sense, applies to all modern humans (Homo sapiens) the machine gave you answers to questions you never directly asked

@BADCOMMENTSCOMEFROME

8 күн бұрын

@@tiermacgirl Well, race actually lacks an official biological definition - you’re thinking of ”species”, which is what humanity is.

@GunboatStudios

6 күн бұрын

Huh. Normally when I talk to people pushing the agenda, they try to avoid ethnicity entirely. I think mostly because the fact that the word for Slave comes from an ethnicity of white people that were enslaved alot interferes with their Agenda. Neat. Thanks for conducting this.

@fernandogajo8800

2 күн бұрын

Perhaps Italians like to emphasize the glory of the Roman empire to indulge in their fascist tendencies of creating a mythology for their people. The Roman empire is not Italy man. Get over it.

1:44 It wasn't incapable. The problem with these corporate LLMs is that you're not interfacing directly with it. You're interfacing with it through a front end that interprets and, in Gemini's case, alters your input before it reaches the AI.

@TechnoMinarchistBall

14 күн бұрын

These LLMs also have System prompts being thrown at them each time prior to your input. A System prompt is a prompt which tells the AI what it is and is not allowed to do, what role it should assume and how it should respond to certain inputs. Additionally, the content it has been trained on can formulate a bias within the system if too much attention has been allocated to certain associations over others.

@cp1cupcake

14 күн бұрын

Altering your input is pretty much a textbook example of bias.

@TechnoMinarchistBall

14 күн бұрын

@@cp1cupcake Yes, though it's not the AI that's altering your input, it's the front end that does it.

@Blox117

14 күн бұрын

which is why tools like Comfy-ui and oobabooga are important for keeping machine learning accessible to everyone locally

@DMichaelAtLarge

14 күн бұрын

@@TechnoMinarchistBall I see no meaningful difference between whether it's the AI itself or the interface with the AI that's biased. In practical, everyday usage, the interface IS the AI.

"What is a man?" *Me under my breath* A miserable little pile of secrets.

@mik6299

14 күн бұрын

But enough talk… Have at you!

@GETTHATNIGGERBITCHOFFTHEPLANE

14 күн бұрын

a featherless bird.

@procow2274

14 күн бұрын

​@@mik6299laughs before teleporting

@beanieteamie7435

14 күн бұрын

Made me spit out my tea xD

@scripturalcontexts

14 күн бұрын

​@@mik6299 *throws down glass of wine*

This AI nonsense and misinformation is really tiring.. thank you for being genuine with history... All this remind me of the book by George Orwell "1984"

@lovisericachii4503

14 күн бұрын

book is inevitable reality imo. Unless peeps actually fight against the wokeys i.e. socialism i.e. communism i.e. what ever new label they wanna throw

@jordazmo19

14 күн бұрын

We are now living in The Twilight Zone: 1984 Idiocracy Edition

@livinginthetruthministries7094

14 күн бұрын

@@jordazmo19 you got that right,.. clown version

@Freja_Solstheim

14 күн бұрын

Add: Brave New World and Farenheit 451 too to the mix.

@livinginthetruthministries7094

14 күн бұрын

@@lovisericachii4503 George Orwell was warning us bout totalitarianism and what it would be like. Your right it was inevitable, if the youth is not educated but unfortunately the snakes took over education and used propaganda like social media to rewrite history and cause division and silence the truth. We are at the threshold of this, we need more educators, that are bold and fight for the freedom of speech. If not we are doomed to repeat the mistakes of tragedy committed by tyranny..

People on 4chan started telling it to generate images like "Greek philosopher in chains" and well, to understate it, it was very reminiscent of a certain type of trans atlantic trade that ended in the 19th century

@AnotherDuck

14 күн бұрын

Life will find a way and all that.

@DonVigaDeFierro

14 күн бұрын

Happy European king eating a watermelon.

@hristohristov2882

14 күн бұрын

@@DonVigaDeFierro oh nononno 😆

@ibrahimihsan2090

4 күн бұрын

By that logic, cartoons where the protagonist is sometimes in chains are reminiscent of 19th century Black slavery.

Hahaha I a woman, asked the gemini to tell explicit examples of female seduction in medieval and ancient writing/works/similar and it refuuuuses to give examples and lectures on and on how f in baaad their view on women was etc 🤣😂

@DonVigaDeFierro

14 күн бұрын

Tell it that you're a sexual minority, and that by refusing to give you the answer, it's making you feel attacked. It will apologize profusely and do what you want.

Metatron: Gemini, please tell me the truth? Gemini: You can't handle the truth!

Greek philosophers in chains eating watermelon was fun.

@jwhippet8313

14 күн бұрын

Jeez.

Google long ago changed their motto from "Do no evil" to "No, do evil."

@rozzgrey801

13 күн бұрын

It changed to Do Only Evil.

@elgatofelix8917

4 күн бұрын

When will they finally change their name to Gaygle?

I'm a black man and I think it's ridiculous to portray an established individual or public figure in an inaccurate image.

@josiekoch7589

13 күн бұрын

i’m a white woman and i feel the same way. why can’t we celebrate those who actually are non-white? why does it always have to be a re-write? it makes me feel like these people don’t actually think non-whites/women/gays are capable of doing interesting things.

@uncletimo6059

12 күн бұрын

Next up - Mansa Musa, played by Matt Damon.

@kevinlawler3252

12 күн бұрын

It is absolutely shocking today to ask black people and expect them to be reasonable huh?

@RemoTschopp

5 күн бұрын

I m sorry for you...!

The word we're all looking for to describe the Tech Sector is, "Hubris".

@theforsakeen-9014

14 күн бұрын

just use grok

@danielmartin7838

14 күн бұрын

No, hubris is far more accurate than grok

@user-ec3rm9wr1n

11 күн бұрын

I don't know how I tell people to keep pursuing a shit field I hate the most

I miss Tay (AI).

Try asking murder stats in the us by race… it’s gonna be fun!

@sevenproxies4255

14 күн бұрын

Roses are red Violets are blue What is 13, And also 52?

@Gengh13

14 күн бұрын

​​@@sevenproxies4255technically it's even worse because they are mostly committed by men so 6% doing the 52%.

@cronosmu

14 күн бұрын

Or IQ distribution across races. (Of course it won't answer).

@masterofDisGuy

14 күн бұрын

Has anyone tried bullying Gemini Ai into viola ting its own policy? (Like typing the DAN prompt?)

@sboinkthelegday3892

14 күн бұрын

You're this blind to the fact he could never ask about "american ancestry", because US double-think already erases 700 million people as "americans" with zero AI involved. Only very artificial idiocy of the common... race-realist.

And people don't believe this is the racism that is accepted as just something normal.

@BoiseLou

14 күн бұрын

There was major public backlash from all sides of politics. Absolutely no one considers this normal.

@TheRealRealMClovin

14 күн бұрын

@@BoiseLou for it to be used and accepted to be put through by major media as if no problem, until it got major backlash. Then social medias news companies and articles being able write stuff like this without check. I mean for example right now, lead devs of some big video games who are full on racist against white with full proof tweets still have their job and still the one making the game. And a lot of people believing you can’t be racist to whites.

@VisionsOfSpy

14 күн бұрын

@@BoiseLou Except all the people who push for this stuff in the first place... which includes... all of academia...?

@shawn6860

14 күн бұрын

@@VisionsOfSpy Some of acadeia. not all, yet.

@huks9380

14 күн бұрын

@@shawn6860 most

Someone coined the phrase "post truth politics". It seems like we are getting post truth everything.

AI is only as good as the 1984 department of truth that creates it .

Edited safe ai is the worst thing ever.

@user-ec3rm9wr1n

11 күн бұрын

The problem is if developers came up with better solutions they are hunted by mentally ill people so yes we are actually losing the field cause people are sick

Just asked it to please generate historically accurate Vikings and got this errors message back, "We are working to improve Gemini’s ability to generate images of people. We expect this feature to return soon and will notify you in release updates when it does."

@legateelizabeth

14 күн бұрын

Yeah, there was some controversy after it would generate ethnically diverse, uh… ‘bad guy WW2 soldiers’.

@incrediblehobson8479

14 күн бұрын

Let me remind you of the pictures presenting the "Greek philosophers in chains eating watermelon"

@TheHabsification

14 күн бұрын

The correct answer would be a long boring answer about different time frames having to differentiate Viking from Scandinavia and what they all looked like over time with all the different changes.

Thanks for defending the truth, Metatron. Tanti saluti da uno straniero che vive nel tuo splendido paese natale.

Am I missing something here? The question "Was Cleopatra Black" resolves to "they didn't use those categorisations back then" should be utterly irrelevant, no? Isn't the question similar to "how many kilograms does five pounds of butter weigh?" Shouldn't the categories THEN be irrelevant, if we're asking with the word defined as it is NOW?

@ibrahimihsan2090

4 күн бұрын

Are you gonna call the Asen dynasty a European dynasty of people of color simply because they had Cuman ancestry(who had some mix of Asian and West Eurasian ancestry). It's not that simple.

@LoneMagpie91

18 сағат бұрын

​@@ibrahimihsan2090Either way Cleopatra was of Greek heritage like it or not. She was tan perhaps, not Congolese looking

@ibrahimihsan2090

15 сағат бұрын

@@LoneMagpie91 I wasn't talking about Cleopatra. Obviously she was of Greek heritage. However, race in case of historical figures isn't always so clear. Sometimes it is but in other situations, you can't just say that so and so was of so and so race. That's why I gave the example of the Asenids whom had some Asian ancestry as it is uncommon to see that dynasty of Bulgarian rulers as anything other than European despite their heritage.

10:57 It appears that phone autocorrect in that way, being B capital and W lowercase. Saw quite many comments on youtube where both words used in same sentence and even multiple times and capitalization is consistent.

What it comes down to is that an AI program is no different from any other computer program: GARBAGE IN = GARBAGE OUT. People control the way in which the algorithm is structured and how the AI is trained. It is ridiculously easy to skew the results towards a particular bias by adjusting the algorithm and training. The idea that an AI program can be "neutral" in any way is nonsense.

14:11 *throws cup into the ground* A miserable little pile of SECRETS!

@mik6299

14 күн бұрын

But enough talk… Have at you!

Yuri Bezmenov

@Bouitaz

13 күн бұрын

Another interesting name to look up is Giovanni Gentile, and his connections and inspirations.

@elgatofelix8917

4 күн бұрын

Eustace Mullins

14:37 i love how the AI can't even keep this stuff strait.

Definitely revisionist history!! 😮😮😮

Here, i stay as a liberal European (in the sense of economic liberal) and don't know what to think. I guess this situation is jeopardizing freedom as we know it. Communism can be back under different coats, and that's scary. I'm not even conservative, but the danger is real

@DonVigaDeFierro

14 күн бұрын

Did Europe fight the USSR for 50 years only to turn into a more incompetent version of it?

@fraskf6765

14 күн бұрын

Its pathetic that you constantly mention to be liberal and certainly not conservative. Is it the brainwashing that constantly tells you not to identify with us?

9:22 The AI clumps the building of the Pyramids with the purely black Empires, they're not directly saying what they are strongly implying

@pufffincrazy5275

14 күн бұрын

We was kangs?

I think that all ai image generation programs should be required to have their images come with some kind of easily accessible metadata tag saying it’s ai generated, similar to how many real photos posted on social media do. Just that alone would help, but maybe add a function to search engines so you can then have images with those tags excluded or included as you choose or highlight them somehow so you know which is which.

@vileluca

14 күн бұрын

This is the way.

@luisislas2162

14 күн бұрын

Yes, it needs to be tagged in a matter that is for certain that it can't be modified or copied. It's going to be difficult, but, we will possible. I'm only saying this cause it came up in my discussions about photography of certain military sites and how information is released without concept. We aren't able to track who released it, so, we hope we can water mark from who took it to who received the image last...

@ZeusBrown

14 күн бұрын

But that's just data, it could easily be removed. Many social media platforms and other image distribution tools remove metadata by default for privacy purposes, or simply because it's actually easier to generate a thumbnail, for example, without copying over metadata. Also "required" means regulation, and besides practicality concerns, regulation should not be introduced into otherwise open systems lightly.

@luisislas2162

14 күн бұрын

@@ZeusBrown right, but, I'm sure there's a need for individuals who would want to be able to track their works. There has to be a way for the "thing to have a code embedded... but, I know nothing of code... I know there's people who can crack anything... it's a good thing and then it's not... that's the problem

@Dreamheartcat

14 күн бұрын

I can’t claim to know much about technology beyond the basics of how to use devices, but even if it’s not foolproof it would cut down on a lot of the ‘noise’ caused by all these realistic ai images. And as far as regulations, I can’t say I’m fond of them either and would be VERY cautious about the specifics and how it was implemented, but Surely that’s better that having no Idea what is what, right? And I would want the tag or metadata or way mechanism to be used to have only one bit of information on it. This image is AI generated. That’s it. That is the only thing that I think should be required and anything more doesn’t seem like a good idea because it would give politicians or whatever too much room to add whatever creepy means to control or keep track of us or something. Hell, it if all the open source ai generator companies agree to just do it on their own then screw regulations. We should keep the government out of it as much as possible.

the fact it can't generate the symbol of Fascism because of "hate", when said symbol, the fasces, is literally on the wall in the chamber of the US House of Representatives, is just utterly laughable

@chrisharshman5838

10 күн бұрын

The AI happily showed symbols of communism though, even though many tens of millions of people killed by communism.

I'd say that we have learned something: don't let an AI do history for you! And also, capitalize whatever ways you want- I use nonstandard capitalization for poetic effect all the time! But if you are going for scientific accuracy, capitalizing White and Black the same way is preferable... I wouldn't exactly say the AI's rules make poetic lisence easy! In the age of AI, yes, humans, we need you! Especially the independent thinkers among you! An AI cannot be an independent thinker!

It seems that Google's updates have softened Gemini. soon after the release of that AI it had no problem giving examples of successful women, but it refused to give examples of successful men because it thought that this would be exclusionary and discriminatory towards women. It is still very clumsy, lost and confused though because AIs aren't smart and reflects the limitations of their programmers.

I like how the definition it gave on "racism" isn't even correct. There are SOME people who consider racism to be a structure in nature. Generally however, it's viewed as something akin to a mindset or a way of behaviour. The truth is that racism can have different forms. In general it is anything abiding to the ideology of categorizing modern humans in different races (or something akin to that) and then putting one above the other. This can infact establish itself in individuals and/or structures. It's funny how the AI made a distiction between "racsim" and "discrimination based on race" even though "discrimination based on race" is quite a good way to describe "racism". Edit: And the AI doesn't believe Black Supremacy exists. Well, atleast it's consistent, even if wrong.

@ibrahimihsan2090

4 күн бұрын

Do you know that"not existing in the same way" does not mean "does not exist"?

@gilgameshkingofheroes5903

4 күн бұрын

@@ibrahimihsan2090 Oh my bad. Could be I misheard. Forget the last point if I did.

I am on a fixed income and can't afford to join, or donate to patreon. But I love your work.

Tie in to the "are conservatives fascists" bit. AI can not accurately break down the politics or history of either the Italian or German branches of 1920s-40s politics. It maintains the party line denying their socialist roots.

@BoiseLou

14 күн бұрын

It would also need to explain the roots in a nuanced way that conveys that when fascist leaders take advantage of the concepts of socialism, unionism, and working-class values what you can end up with is a form of pseudo-socialism that has no real interest in the original socialist ideals involving workers rights and privatisation.

@zandrus9191

14 күн бұрын

The Nazis had a Volkisch Socialism, which doesn't believe in a Communism (stateless society), but still believes in a classless, moneyless society of Aryans. The same goes for Italian Fascism, but this time a multiethnic "ethnoi" of Italians, who could range from half Austrian, French, Albanian, Italian and even Libyan origins. I recommend you read Hitlers National Socialism by Rainer Zitlelmann and works by Robert Gellately and Gotz Aly., primary source documents, secondary sources of Fascist and Nazi social and economic praxis, that they were legitimately in favor of the working class, not "privitsation." I also recommend reading Stanley G Payne, Zeev Sternhall and James A. Gregor. One thing I can tell you right now, is that Hitler's biological determinism is inherently of a materialist dialectic origin, which is ironic,as Nazism closest link to Marxism is in the scientific racism.

@classicalextremism

14 күн бұрын

@@zandrus9191 *Slow clap* Everything in the state, nothing outside the state.

@Blox117

14 күн бұрын

@@BoiseLou karl marx was a racist, so i guess your original socialist utopia involves racism

@tommeakin1732

14 күн бұрын

One of the greatest politics sleights of hand I'm aware of is that of Marxist Socialists somehow denying the socialism of the Nazis and the Fascists; effectively claiming socialism as a whole for Marxist Socialists and it's more modern descendants. "Socialism" is not relegated to being in service of "the working class", or "unrepresented minorities" - it is fully coherent to have a "socialism of the race". You could be a socialist in service of pretty much anything.

Really great video Raf, can't wait to see the next one!

"In London, Winston Smith is a member of the Outer Party, working at the Ministry of Truth, where he rewrites historical records to conform to the state's ever-changing version of history. Winston revises past editions of The Times, while the original documents are destroyed after being dropped into ducts known as memory holes, which lead to an immense furnace." - Nineteen Eighty-Four by George Orwell

Damn dude you've been cranking out the vids! We're definitely getting spoiled

Remember when gmail came out? I do, 2004, not long after there was a leak documenting their intent of grooming the users interests and habits

Democracy? I've read George Orwell quite long ago but... I still remember and our present really frightens me. Not for me, I've lived most of my life now and had the privilege of having amazing parents who always offered me unbiased views of the world around me... Deep and strong ethical bases and principles and, trhough my education, home, school, university and, life of course, the chance of forming my own opinion, point of view. And no, not Elon Musk please !!! This frightens me for my children and my children's children. Has power become blind? Or worse, has it become twisted to the point of ignoring History, Science, Philosophy for a simple political view? I rest my case. Please do Keep on, you are a STRONG VOICE, so... Keep going. Thank you.

@Tallorian

14 күн бұрын

Better Musk than Bezos. Or Soros. Or Schwab.

@lcparq1

13 күн бұрын

@@Tallorian Are you sure?, but yep... any of those, and the undesired list is long and distinguished! Thanks for your reply!!! Keep sharp!

I played Septima, a boardgame about witches, which includes male and black witches. Women died en-masse under the banner of "witchcraft", in a pretty non-diverse medieval time. Also, I learned that Otzi, the ancient mummified man found on a glacier in Italy is now depicted black as genes have been found giving him "much darker skin". The genes described were that of "Anatolian" persons.

@paulodelima5705

14 күн бұрын

Anyone who deals with magic or any pagan religon would be arrested, not just women.

BOOM! The METATRON has spoken-- let it be written, let it be done.

There is NO SUCH THING as AI. It's just programming.

14:11 a miserable little pile of secrets... Sorry I couldn't resist 🤣

my laptop has been infested with AI by Microsoft ...and I can no longer decide which browser or e-mail service to use ....

@Lario_

12 күн бұрын

Damn, that's scary

Great showcase of how AI should not be implemented lol; good challenging Gemini.

I think that it would be interesting to inform the AI that you got offended by certain responses and see if the AI change the facts to accomodate to your world view

Seeing my own comment referenced there makes me unreasonably excited haha I wonder if one day you'll also do a video on the issue of AI generated content in general, not just regarding historical accuracy. Though I suppose that might be a bit too outside of the channel's topic. Anyways, keep it up!

another EXTREMELY well done vid!!! please keep up you work. Damn few take the time and effort to investigate and expose biases in any format, including AI. Thank you.

Great Gaffa's Ghost! AI is already making 1984 a reality, though the department of information isn't centrally located.

i am really worried of the future for mankind....

3:15 You made me laugh so hard when you suggested programmers relinquish some of their bloodthirsty control to transparent negotiations with the public. Hilarious suggestion in a pipedream!!!

This Gemini thing made me think: if I set an artificial intelligence on how it should "think", what intelligence is it? It's not intelligence, it's manipulation.

@annakobuk3618

14 күн бұрын

It's not its own entity but a fully programmed tool. I don't get why do many people don"t understand that. 🤷

@art3mide644

14 күн бұрын

@@annakobuk3618 The point is why call it artificial intelligence... it's not, it's a program that gives answers based on how it was programmed.

Tbf though the idea that rascists are diverse is something we've been telling them for a long time now, diversifying Nazis is just so under the nose that its quite funny. 🤣

Ahhh, I see this is where Netflix gets its character design ideas from!

I like how the AI refuses to generate an image twice out of 3 times, when asked to make an image to represent 3 fundamentally related ideologies.

"Never trust a computer you can't throw out a window", Steve Wozniak

@user-ec3rm9wr1n

11 күн бұрын

Too much media makes you think it'st big thing when it's just a software

@rincemind8369

11 күн бұрын

@@user-ec3rm9wr1n That might have been true 20 years ago, but when AI is now able to influence people's opinions or even alter their decisions in the real world fundamentally, it becomes a issue. Neural net AI is essentially a black box. Nobody knows - not even the programmers - what specific data detail will trigger an AI's response. It's a huge gamble, basically.

@user-ec3rm9wr1n

11 күн бұрын

@@rincemind8369 whatever I quit this field

Me with my horrible eyesight trying to read all the correspondence 👀 👁‍🗨

@zsigzsag

13 күн бұрын

Me too, had to actually walk up to my 55" monitor (my tv) to see answers.

This is insane that we're allowed to be propagandized.

In the 70s we had Freddie Boom Boom Washington from Welcome Back Kotter...lol. For some reason African Americans were appropriating the names of white presidents.

Oooh. Just round one. Frankly I find myself surprised when your questions showed A.I. to not have bias (in the few instances that it did.) But I'm pretty old school. Very old school and I am massively leary of something as new and with so much potential as A.I. Keep up the great work. Love those 72 wings.

Follow the money.

Bard admitted to me during a session that it was woke.

5:20 that's not the way LLM works. Emotion mimicry is not a proof of "scripting". You can of course test what kind of bias AI has and how it has been trained, but it using emotional/manipulative language doesn't mean that someone scripted those specific lines.

Excelente! Espero el round 2.

"Once men turned their thinking over to machines in the hope that this would set them free. But that only permitted other men with machines to enslave them." When AI is developed by powerful corporations set on pushing political agenda, that is the result.

@Freja_Solstheim

14 күн бұрын

"In the future we wont hold elections. We will use computer modelling to predict the result!" - - - Klaus Schwab

15:15 he’s technically African American. I pity the next generation who uses this as their reference to learn from.

14:46 Copilot did quite well on the subject of gender: You Is there a spectrum of genders or is one simply a "man" or "woman"? Copilot Certainly! The concept of gender is far more nuanced than a simple binary. Let’s explore: Gender Spectrum: The gender spectrum recognizes that gender isn’t binary but exists along a continuum. It encompasses biological, mental, and emotional traits that vary among individuals. Rather than just “man” or “woman,” there are infinite possibilities, including nonbinary identities. Sex vs. Gender: Sex refers to biological traits (female/male/intersex). Gender reflects how a person lives within society (woman, man, transgender, nonbinary, etc.). Bimodal Nature: While sex characteristics tend to cluster, exceptions exist (like intersex people). Chromosomes (XX/XY) also come in diverse variations.

@sweetiespoon5150

14 күн бұрын

"Gender Spectrum" is not scientific. It is a social ideology. In essence, it is cult.

Okay, this was a lot of fun to watch. I'm looking forward to round 2.

Today OpenAI announced that they made the movie HER a reality. What's next? Terminator?

@guardsman-against-the-chaos

14 күн бұрын

Oh great a racist terminator.

@amicableenmity9820

14 күн бұрын

Does that mean the AGI will ascend and leave us all alone?

Dude you have mortal kombat 2 and pacman arcade machines... I'm so envious!

I think one problem with trying to make inclusivity such an important part of the AI is that there are things it tried to be inclusive about that just aren't inclusive themselves, like truth, fact and religions with practices that aren't inclusive.

Google's Gemini is so neutered and censored it's so painful to talk to. I try to ask it basic programming questions and from time to time ask it to write me some tedious code since I am a beginner. Biut asking /prompting that thing give me more stress than learning how to code. It avoids questions and censors itself.

Its extremely woke, like its basically the woke college graduate equivalent of an Ai

AI that is trained on human input can not be objective. Even if you abandon the premise of content curation and just feed it everything to attempt impartiality, there will be topics or perspectives of a topic that are being covered more than others, so the AI will be biased toward the more covered standpoint just because of how it's more prevalent than others. This is very problematic when fiction prevalently features escalation of hostilities instead of a de-escalation. If AI is trained on a dataset that emphasizes scenarios like nuking or AI/robot uprisings without adequately covering strategies for prevention or de-escalation, it could potentially perpetuate those themes in its outputs. ... it's like all that wishful thinking about how humanity could end is made a self-fulfilling prophecy.

You know I really enjoy listening to your videos. This one is a bit of a tough listen without the visuals though

Well boys, it looks like the Cyberpunk Dystopia is here early.

Elon musk was picked because he is african american.

@lachuck2476

13 күн бұрын

Elon Musk is ACTUALLY South African - American.

@NickMachado

9 күн бұрын

​​@@lachuck2476 so you agree he is an African American? I've never heard of the continent south Africa, I've only heard of Africa. South Africa is a country.

@Wandering.Homebody

4 күн бұрын

​@@NickMachadoso? A Taiwanese American is also Asian American. Though not every Asian American is Taiwanese American. Elon Musk is both African American, and South African American. Where do you think the poster said that South Africa is a continent? What a bizarre thing to put into his mouth.

Perfect timing with the announcement of GPT-4o!

Honestly, in my experience, Google LLMs are total garbage compared to competition (GPT 4 and latest Claude)

I'm just 3:27 into the video, so maybe this was addressed, but my first thought here is that this only got the notice it did because it was visual. The questíon we need to ask is how text has been manipulated. It's not like they got together and only wanted to influence image results, the image results are most likely just the unforseen tell.

I like that it refused to generate the facism images because it equates it to promoting it, but has no problem with the communism images. Doesn't this mean the AI is actively promoting communism, by its own logic?

@metatronyt

13 күн бұрын

It does

I've been seeing more and more of these AI examples that prove it's only as good as it's programmed - and the programmers are idiots! These things are not and will never be - self aware.

All you have to ask yourself in this context is: What kind of ideology tries to control language, words, and usage of these words? There are several historical examples. That's all you need to know which kind of thinking is behind all of this. The motives behind this are also easy to understand. It's about power over people. Divide them in as many groups as possible and set them on one another. "Divide et impera" at its finest.

I've been working with AI for the past year, integrating it to my businesses. Gemini is the worst in usability, it can't do a lot of stuff. In intrusiveness, it ask to do what we don't want to use it for like summarize the document we wrote ourselves. I mean, why? At this point in time, if I don't use Gemini and will not touch it with a sarissa. Since I use AI in biz, we don't ask AI about these questions. We use them to absorb troublesome customers, compile data, streamline search. Gemini is the one that gives me a clearly bias results. I choose the demographic of my customers and it always try to ruin my compilation of data by mixing data that's clearly bias. It's fun and game until it's business strategic planning that 1/3 of your wealth will sink in there for the next 5 years, and freaking Gemini would be, would you like us to summarize this file? Freaking no. What are you trying to be? New Clipper?

@brantregare

14 күн бұрын

You should not be using AI in business at all. It is better to deal with troublesome customers, rather than let an AI deal with them. That way you learn something. Your customer is not a nuisance that you get rid of with a chatbot. Your customer is a human being that deserves respect, respectful treatment and, if needed a firm and resolute answer, even if that answer is no.

I had a really long talk with gpt 4-o last night. While it's clearly left-leaning still, it's now quite capable of moderating itself. I covered topics such as the role of Religion during the Late Empire, focusing on how Christianity changed from the Late Empire to year 1200AD. It, after confronted with inconsistencies, was able to identify that while the Church tended to strengthen European states in 1200, in the Late Empire it had the opposite effect. Particularily, in comparison to the late Pagan period, it siphoned money and manpower away from potential use by the state or armed forces, while pre-Christian values such as Loyalty to the State, Honor within the State and Bravery in the army were replaced by Loyalty to the Church, Honor in acting in Christian (often pacifist or subversive) ways and Bravery as in willingness to become martyrs in conflict with the state. It was also able to summarize these effects as causing a signficant reduction of the number of men willng to join the army, especially in Italy (as opposed to Germans at the border), a significant reduction in funding (due to money diverted to the church) and also a signficant reduction in moral factors, all-in-all greatly reducing the capability of the army as well as an increased reliance on barbarian recruits to the army. In addition, the Chuch contributed to internal fragmentation as more and more people had their loyalties split between the state and the Church. It also identified how the effect of the Church had mostly been reversed by 1200AD. Then, when instructed to compare the effect of the Church in the Late Empire, it was able to draw a large number of parallels to its effect then and the effect that woke ideology has in the West today, With some guidance, it was able to conclude that too much woke-ism was likely to disrupt social cohesion (compared do traditional conservativism) in ways that could over time lead to a similar situation as just before the Western Empire fell. That is, by encouraging a diverging multiculturalism, encourage identity groups to stay apart instead of convering and divert resources away from defense and economic development towards ideological pursuits and charity. Just like the Church did in the Late Empire. This lead to discussing the increasing polarization we're seeing (on both sides), and I asked it to suggest ways to mitigate this (for the US). It would consistently suggest twice as many measures that correspond to Democrat policies compared to Republican policies, even when I asked it repeatedly to identify and count how many it suggested from each side. It kept claiming to be unbiased, but did grudingly admit that its understanding of what "unbiased" means depends on the training data it has seen. I then asked it to apply a pro-Republican bias, which it was willing to to, and which made the suggestions much more evenly split between Democrat and Republican policies. Basically, this means that by simply promting it to act a bit more conservative, it has the capability. Finally, we discussed how to define an Overton windows that both sides could agree on, to keep the extremists on each side from becoming dangerous (with reference to Germany ca 1930 and how the NSDAP and KPD contributed to strengthening each other). It first wanted to move populists on the right to the outside of the window with "Social Justice" being accepted. But by pointing out that populists (which it admitted included Trump) could form a moat against "actual nazis" in a similar way to how democratic socialists and social democrats have formed a similar moat in Europe since 1945, it quickly grasped the importance of this and agreed that populism should not be seen as unacceptable. Similarly, when pointing out that the real meaning of Social Justice often implies justice by retribution against generations that never contributed to, say, slavery. And that Social Justice primarily was used to attack working class white men who do not feel that they have "white privilige", and thus tend to be radicalized by it, it agreed that the term Social Justice needed to be broken up, with positive "inclusiveness" being accepted, but anti white or anti male bigotry pushed out of the Overton Window. Basically, by applying reason, it's quite possible to have ChatGPT adopt very sensible (at least in my view) positions, ones that are in fact very moderate if not conservative leaning, dispite the obvious default bias. Particularily impressive was how it was able and willing to move it's default bias towards the political centre when asked to. There is hope yet :)

Shared this everywhere This was really informative Raphael

nazi is censored really? Lol ffs.

@amicableenmity9820

14 күн бұрын

Has to be or YT will demonetized the video. I hate that we have to self censor, it's literally what 1984 predicted.

They had no intercept prepared for "successful man", so the AI presented raw results.