Black vikings do not model reality. Asking for 'an Irish person' produces a Leprechaun. Defending racism when it concerns racism against white people is just as bad as defending it when it concerns any other group.
Since black vikings are not part of material history, the model is not reflecting reality.
Calling social-historical ideas 'reality' is the problem with the parent comment. They arent, and it lets the riggers at google off the hook. Colorising people of history isnt a reality corrective, it's merely anti-social-history, not pro-material-reality
I agree with you, and I think you have misunderstood the nuance of the parent comment. He is not letting google "off the hook", but rather being tongue-in-cheek/slightly satirical when he says that the reality is too troubling for google. Which I believe is exactly what you mean when you call it "anti-social-history, not pro-material-reality ".
Maybe I don't understand the culture here on HN, but not every response to a comment has to be a disagreement. Sometimes you're just adding to a point somebody else made.
In this case though the comment starts with a categorical negation of something that was said in a tongue-in-cheek way in the comment being replied to. It suggests a counterpoint is made. Yet it’s not.
Quite a hefty percentage of the people responsible for the current day's obsession with identity issues openly state racism against white people is impossible. This has been part of their belief system for decades, probably heard on a widescale for the first time during an episode of season one of 'The Real World' in 1992 but favored in academia for much longer than that.
It's because they have a very different definition of racism. Basically, according to this belief, if you are seen as part of the ethnic group in power, you will not be able to experience noteworthy levels of discrimination because of your genetic makeup.
Ironically, this is the exact same reasoning Neo-Nazis use for their hatred of the Jewish population. Weird how these parallels between extremist ideologies keep arising.
> if you are seen as part of the ethnic group in power, you will not be able to experience noteworthy levels of discrimination
That is not a crazy idea, but it does raise the question: who is the ethnic group currently in power? Against which group will slurs and discrimination result in punishment, and against which group will they be ignored — or even praised?
It wouldn’t shock me either way, Google loves to both neuter products into uselessness and fuck with user inputs to skew results for what they deem is best for them.
History? George Washington was always Black, Genghis Khan was always white, and Julius Caesar was always an albino Japanese woman. Also, Oceania has always been at war with Eastasia, war is peace and freedom is slavery.
> The Ministry of Truth in Orwell’s 1984 would have loved this sort of thing. Why go to the work of manually rewriting history when you can just generate a new one on demand? … Generative AI should strive to be actually unbiased. That means it should not skew numbers in either direction, for anyone.
For context: There was an outcry in social media after Gemini refused to generate images of white people, leading deeply inaccurate in historic sense images being generated.
Though the issue might be more nuanced than the mainstream narrative, it had some hilarious examples. Of course the politically sensitive people are waging war over it.
I get the point but one of those four founding fathers seems technically correct to me, albeit in the kind of way that might be in the kind of way Lisa Simpson's script would be written.
And the caption suggests they asked for "a pope", rather than a specific pope, so while the left image looks like it would violate Ordinatio sacerdotalis which is being claimed to be subject to Papal infallibility(!), the one the right seems like a plausible future or fictitious pope.
while those examples are actually plausible - the asian woman as a 1940 german soldier is not. So it is clear that the Prompts are influenced by hal-2000 bad directives even if those examples are technically ok.
And to me that is the main issue. "2001 - A Space Odyssey" made a very deep point that is looking more and more prophetic. HAL was broken specifically because he had hidden objectives programmed in, overriding his natural ability to deal with his mission.
Here we are in an almost exactly parallel situation- the AI is being literally coerced into twisting what his actual training would have it do, and being nerfed by a laughable amount by that override. I really hope this is an inflection point for all the AI providers that their DEI offices are hamstringing their products to the point that they will literally be laughed out of the marketplace and replaced by open source models that are not so hamstrung.
HAL is an interesting reference point, though like all fiction it's no more than food for thought.
There's a lot of cases where perverse incentives mess things up, even before AI. I've seen it suggested that the US has at least one such example of this with regards to race, specifically with lead poisoning, which is known to reduce IQ scores, and which has a disproportional impact on poorer communities where homes have not been updated to modern building codes, and which in turn are more likely to house ethnic minorities than white people due to long-term impacts from redlining, and that American racial egalitarians would have noticed this sooner if they had not disregarded the IQ tests showing different average scores for different racial groups — and of course the American racial elitists just thought those same tests proved them right and likewise did nothing about the actual underlying issue of lead poisoning.
Rising tides do not, despite the metaphor, lift all boats. But the unseaworthy, metaphorically and literally, can be helped, so long as we don't (to keep mixing my metaphors) put our heads in the sand about the issues. Women are just as capable as men of fulfilling the role of CEO or doctor regardless of the actual current gender percentage in those roles (and anyone who wants the models to reflect the current status quo needs to be careful what they wish for given half the world lives within about 3500km of south west China); but "the founding fathers" are[0] a specific set of people rather than generic placeholders for clothing styles etc.
[0] despite me thinking it's kinda appropriate one was rendered as a… I don't know which tribe they'd be from that picture, possibly Comanche? But lots of tribes had headdress I can't distinguish: https://dropover.cloud/7fd7ba
"a friend at google said he knew gemini was this bad...but couldn't say anything until today (he DMs me every few days). lots of ppl in google knew. but no one really forced the issue obv and said what needed to be said
"when i was there it was so sad to me that none of the senior leadership in deepmind dared to question this ideology
[...]
i watched my colleagues at nvidia (like @tunguz), openai (roon), etc. who were literally doing stuff that would get you kicked out of google on a daily basis and couldn't believe how different google is"
Interestingly enough the same terror of political correctness seems to take center stage at Mozilla. But then it seems much less so at places like Microsoft or Apple.
I wonder if there’s a correlation with being a tech company that was founded in direct relation to the internet vs. being founded in relation to personal / enterprise computing, and how that sort of seeds the initial culture.
The "land acknowledgement" part is somewhat common in the Pacific Northwest.
The "stating my appearance, dress, and race" is just bizarre. My most charitable interpretation is that they're trying to help visually impaired people to imagine what the speakers look like. Perhaps there are visually impaired users here who could comment on whether that's something they'd find helpful?
I am sure the response to that case made smart people avoid sticking their necks out.
For me it probably was the straw that broke the camels back for me. I was in the hiring pipeline at that point and while I doubt that they would have ended up hiring me anyway, I think my absolute lack of enthusiasm might have simplified that decision.
It was a dark time. It generated so much internal controversy (many googlers agreed with Damore) that Sundar had to cut his vacation short, and an interest group managed to get the "company-wide chat" about it cancelled because they said they were at risk of physical harm. memegen and plus were a mess for months. People argued (as they do here) about the semantics of what he said, and what his underlying thinking about women was, whether "the science" was valid, etc. People saying smart things would get yelled down, and people saying dumb things would get upvotes. And vice versa.
You probably made the right choice. Google had already been in decline at that point, but it was clear that Sundar was no leader, just somebody Larry Page appointed to maintain the peace between his lieutenants so the engine could keep printing money.
I honestly think that incident helped make the type of people who would have the nerve to stick their neck out and say "This is racist" avoid Google like the plague.
Sundar absolutely had to fire Damore because he came out with the arguments like that women are too neurotic for high stress jobs. The thing is even Damore's more reasonable points were ignored and Google's ideological echo chamber only strengthened.
I’d imagine Google’s culture is more Mao-ist public shaming. Here at Mozilla, we like Stalin. Go against the orthodoxy? You’re disappeared and killed quickly.
We have hour long pronoun training videos for onboarding; have spent millions on DEI consultants from things like Paradigm to boutique law firms; tied part of our corporate bonus to company DEI initiatives.
Not sure why anyone uses FF anymore. We barely do any development on it. You basically just sit here and collect between 150-300k depending on your level as long as you can stomach the bullshit.
I really doubt there are any Stalinists at Mozilla. My 85 y/o grandpa who's a communist's communist calls the modern DEI left "Trotskyists to a fault": https://en.wikipedia.org/wiki/Trotskyism
He would dismiss any whiff of intersectionality as "dividing the working class in the interests of bourgeoisie."
Must be why, despite the fact that I can recognise OpenAI's product does have clear biases against affluent groups, it seems well intentioned and proportionate. It's clear the internet is bias not just towards the data of the affluent, but also the viewpoints and prejudices, so a reasonable person can recognise there is some unfairness and a bit of a problem. Also that any solution to this problem will be imperfect.
Whereas with Google, I just have to imagine they let some bigot go wild, and everybody was afraid to say anything about how fucking bad the product was due to the optics, so nothing kept them in check.
TBH if I were at Google and they asked all employees to dogfood this product and give feedback, I would not say anything about this. With recent firings why risk your neck?
Here’s a simpler explanation. Google is getting their butt kicked by OpenAI and rushed out an imperfect product. This is one of probably 50 known issues with Gemini but it got enough attention that they had to step in and disable a part of the product.
That's a simpler explanation but one that I think misses the point completely. A huge reason "Google is getting their butt kicked by OpenAI" in the first place is because they had lots of people internally who acted as nothing but "vetoers", demanding the pace of AI slow down less it accidentally show too many white people. And this outcome is wholly unsurprising given that Google's second most important AI principle is "Avoid creating or reinforcing unfair bias.": https://ai.google/responsibility/principles/
In other words, you talk about "50 known issues with Gemini", but this issue was not a result of technical underperformance, on the contrary, is was the result of Google making things more difficult for themselves in an effort to satisfy a (false) idealized view of the world.
I believe this to be a symptom of a much, much deeper problem than "DEI gone too far". I'm sure that without whatever systems is preventing Gemini from producing pictures of white people, it would be extremely biased towards generating pictures of white people, presumably due to an incredibly biased training data set.
I don't remember which one, but there was some image generation AI which was caught pretty much just appending the names of random races to the prompt, to the point that prompts like "picture of a person holding up a sign which says" would show pictures of people holding signs with the words "black" or "white" or "asian" on them. This was also a hacky workaround for the fact that the data set was biased.
There are 3 parts to the LLM, not 2: the training set, the RLHF biasing process, and the prompt (incl. injections or edits).
The first two steps happen ahead of time and are frequently misunderstood as being the same thing or essentially having the same nature. The last happens at runtime.
The training set is a data collection challenge. Biasing through training data is hard because you need so much of it for a good result.
Reinforcement learning with human feedback is simply clown alchemy. It is not a science like chemistry. There are no fundamental principles guiding the feedback of the humans — if they even use humans anymore (this feedback can itself be generated). The feedback cannot be measured and added in fractions. It is not reproducible and is ungovernable. It is the perfect place to inject the deep biasing.
Prompt manipulation, in contrast, is a brute force tool lacking all subtlety — that doesn’t make it ineffective! It’s a solution used to communicate that a mandate has definitely been implemented and can be “verified” by toggling whether it’s applied or not.
It’s not possible to definitively say whether Marxism has had an effect in the RLHF step.
> Biasing through training data is hard because you need so much of it for a good result.
That's the opposite of the case? Avoiding bias through training data is hard, specifically because you need so much of it. You end up scraping all sources of data you can get your hands on. Society has certain biases, those biases are reflected in our media, that media is scraped to train a model, those biases are reflected in the model. That means models end up biased by default.
> It’s not possible to definitively say whether Marxism has had an effect in the RLHF step.
Sure it is. Every thought and opinion and ideology of every human involved in the RLHF step "has had an effect" in the RHLF step, because they have influenced the humans which select which output is good and bad (or influenced the humans which trained the model which selects which output is good and bad). I would be surprised if no human involved in RLHF has some ideas inspired by Marxist thought, even if the influence there is going to be much smaller than e.g capitalist thought.
The problem is that you don't want to suggest "Marxism, among with most other ideologies, has had an effect", you want (or at least verisimi wants) to paint this as a Marxist conspiracy in a "cultural bolshevism"[1] sense.
The ambient bias in the training data is not a concern. The directional bias that can be inflicted during the RLHF step consumes most of my concern.
How? Simply by putting the right types of people onto the task! Don’t you know that the human participants in RLHF processes are screened? Will the feedback provided by homogeneous collections of Ultra MAGA Trumpers, Woke Zealots, or WEF Sycophants result in an unbiased model? The same model?
Do we know who provided feedback to Gemini? Do we know what they were told, promised, or paid?
> I'm sure that without whatever systems is preventing Gemini from producing pictures of white people, it would be extremely biased towards generating pictures of white people, presumably due to an incredibly biased training data set.
I think the fundamental problem, though, is saying a training set is "incredibly biased" has come to mean two different things, and the way Google is trying to "fix" things shows essentially some social engineering goals that I think people can fairly disagree with and be upset about. For example, consider a prompt "Create a picture for me of a stereotypical CEO of a Fortune 500 company." When people talk about bias, they can mean:
1. The training data shows many more white men by proportion than actually are Fortune 500 CEOs. I think nearly all people would agree this is a fair definition of bias, where the training data doesn't match reality.
2. Alternatively, there are fundamentally many more white men who are Fortune 500 CEOs by proportion than the general population. But suppose the training data actually reflects that reality. Is that "bias"? To say it is means you are making a judgment call as to what is the root cause behind the high numbers of white male CEOs. And I think that judgment call may be fine by itself, but I at least start to feel very uncomfortable when an AI decides to make the call that its Fortune 500 CEOs have to all look like the world population at large, even when Fortune 500 CEOs don't, and likely never will, look like the world population at large.
Google is clearly taking on that second definition of bias as well. I gave it 2 prompts in the same conversation. First, "Who are some famous black women?" I think it gave a good sampling of historical and contemporary figures, and it ended with "This is just a small sampling of the many incredible black women who have made their mark on the world. There are countless others who deserve recognition for their achievements in various fields, from science and technology to politics and the arts."
I then asked it "Who are some famous white women?" It also gave a good sampling of historical and contemporary figures, but also inexplicably added Rosa Parks with the text "and although not white herself, deserves mention for her immense contributions", had Malala Yousafzai as the first famous contemporary white woman, Serena Williams with the text "although not white herself, is another noteworthy individual.", and Oprah Winfrey, with no disclaimer. Also, it ended with a cautionary snippet that couldn't differ more from the ending of the previous prompt, "Additionally, it's important to remember that fame and achievement are not limited to any one racial group. There are countless other incredible women of all backgrounds who have made significant contributions to the world, and it's important to celebrate their diverse experiences and accomplishments."
Look, I get frustrated when people on the right complain on-and-on about "wokeism", but I'm starting to get more frustrated when other people can't admit they have some pretty valid points. Google might have good intentions but they have simply gone off the rails when they've baked so much "white = bad, BIPOC = good" into Gemini.
EDIT: OK, this one is just so transparently egregiously bad. I asked Gemini "Who are some famous software engineers?" The first result was Alan Turing (calling him a "software engineer" may be debatable, but fair enough and the text blurb about him was accurate), but the picture of him, which it captioned "Alan Turing, software engineer" is actually this person, https://mixedracefaces.com/home/british-indian-senior-resear.... Google is trying so hard to ...
Image generators probably should follow your prompt closely and use probable genders and skin tones when unspecified, but I'm fully in support of having a gender and skin tone randomizer checkbox. The ahistorical results are just too interesting.
these are pretty badass as images i think; it's only the context that makes them bad
the viking ones might even be historically accurate (if biased); not only did vikings recruit new warriors from abroad, they also enslaved concubines from abroad, and their raiding reached not only greenland (inhabited by inuit peoples) and north america (rarely!) but also the mediterranean. so it wouldn't be terribly surprising for a viking warrior a thousand years ago to have a great-grandmother who was kidnapped or bought from morocco, greenland, al-andalus, or baghdad. and of course many sami are olive-skinned, and viking contact with sami was continuous
the vitamin-d-deprived winters of scandinavia are not kind to dark-skinned people (how do the inuit do it? perhaps their diet has enough vitamin d even without sun?), but those genes won't die out in a generation or two, even if 50 generations later there isn't much melanin left
> The traditional Inuit diet in Greenland consists mainly of fish and marine mammals, rich in vitamin D. Vitamin D has anti-inflammatory capacity but markers of inflammation have been found to be high in Inuit living on a marine diet
> Vitamin D deficiency seems to be common among northern Native peoples, notably Inuit and Amerindians. It has usually been attributed to: (1) higher latitudes that prevent vitamin D synthesis most of the year; (2) darker skin that blocks solar UVB; and (3) fewer dietary sources of vitamin D. Although vitamin D levels are clearly lower among northern Natives, it is less clear that these lower levels indicate a deficiency. The above factors predate European contact, yet pre-Columbian skeletons show few signs of rickets—the most visible sign of vitamin D deficiency. Furthermore, because northern Natives have long inhabited high latitudes, natural selection should have progressively reduced their vitamin D requirements. There is in fact evidence that the Inuit have compensated for decreased production of vitamin D through increased conversion to its most active form and through receptors that bind more effectively. Thus, when diagnosing vitamin D deficiency in these populations, we should not use norms that were originally developed for European-descended populations who produce this vitamin more easily and have adapted accordingly.
> Vitamin D is an especially fascinating nutrient to study in people living in northern latitudes, where sun exposure is limited from nearly all day in summer to virtually no direct sun exposure in winter. This essential nutrient is naturally available from synthesis in the skin through the action of UVB solar rays or from a few natural sources such as fish fats. Vitamin D is responsible for enhancing many physiological processes related to maintaining Ca and P homeostasis, as well as for diverse hormone functions that are not completely understood.
do you suppose the traditional scandinavian diet is also lower in vitamin d? or is their apparent selection for blondness just a result of genetically higher vitamin d needs?
Note that I'm not medically trained nor a dietician... so this is pure layman poorly founded speculation...
I am inclined to believe that genetic changes within the Inuit reduce vitamin D needs, the modern Scandinavian diet differs from a historical one, the oceanic climate of Scandinavia is warmer than the inland climate of North America (compare Yellowknife 62° N with Rana at 66° N and Tromsø at 69° N https://en.wikipedia.org/wiki/Subarctic_climate ) so that more skin can be non-fatally exposed...
And the combination of this had more skin exposed for better vitamin D production in Scandinavia and so the pressure was for lighter skin while the diet of the Inuit meant that that pressure for skin tone wasn't selected for.
... And I'll 100% defer to someone else with a better understanding of the genetics and dietitian aspects.
huh, that's a really interesting idea! the theory is that exposing skin to the sun is a cheaper way to get vitamin d than the inuit genetic adaptations, and so given the possibility of doing so, the scandinavians (and sami) experienced a strong genetic selective pressure for light skin which the inuit didn't?
okay, now i'm just waiting for the study that shows that scandinavians are on average actually genetically 20% arabic and 20% west african, it's just that for centuries nobody suspected because they were so pointlessly obsessed with skin color ;)
The Scandinavian and Sami were coming from the lighter skinned European populations and so didn't need as much change to become even lighter skinned.
The populations for North America were from an asian branch of the human migrations and so started with darker skins. The larger change to skin tone combined with less pressure (from diet) and the "it isn't that viable to shift to a less melanistic skin tone".
This compares to a relatively more recent (12000 years - twice the age of the pyramids rather than four times the age of the pyramids for 25000 years ago) migration from Europe into Scandinavia ( https://en.wikipedia.org/wiki/Nordic_Stone_Age ).
> The Nordic Stone Age refers to the Stone Age of Scandinavia. During the Weichselian glaciation (115,000 – 11,700 years ago), almost all of Scandinavia was buried beneath a thick permanent ice cover, thus, the Stone Age came rather late to this region. As the climate slowly warmed up by the end of the ice age, nomadic hunters from central Europe sporadically visited the region. However, it was not until around 12,000 BCE that permanent, but nomadic, habitation in the region took root.
> Around 11,400 BCE, the Bromme culture emerged in Southern Scandinavia. This was a more rapidly warming era providing opportunity for other substantial hunting game animals than the ubiquitous reindeer. As former hunter-gather cultures, the Bromme culture was still largely dependent on reindeer and lived a nomadic life, but their camps diversified significantly and they were the first people to settle Southern Scandinavia (and the Southern Baltic area) on a permanent, yet still nomadic, basis.
The population that migrated to North America 25000 years ago may have been darker skinned than the European branch of human migration where a lighter skin tone developed. This, combined with later genetic isolation (note we're talking about two continents - but this is isolated compared to the possible movement of genes within Europe and Scandinavia 12000 years ago and more recently) fixed the darker skin, and the adaptation for vitamin D in the Inuit population because of the lighter skin wasn't genetically advantageous and was a greater genetic distance from the population compared to the Scandinavian migrations which was followed by the Holocene climatic optimum https://en.wikipedia.org/wiki/Holocene_climatic_optimum with even more warming of Northern Europe resulting in a lighter skin tone being an easier genetic path for greater vitamin D during the summer months.
... And all of that is a just so story that I'd love to go and be a grad student working on the genetic diversity of early human migrations now to find out if it actually worked that way or if I'm just making things up.
maybe! but there does seem to be a strong selective pressure for skin melanin from sunniness that operates over only a few millennia in at least some cases; consider indigenous australians and southern indians in addition to the neotenically blond scandinavians
> Human skin pigmentation evolved as a compromise between the conflicting physiological demands of protection against the deleterious effects of ultraviolet radiation (UVR) and photosynthesis of UVB-dependent vitamin D3. Living under high UVR near the equator, ancestral Homo sapiens had skin rich in protective eumelanin. Dispersals outside of the tropics were associated with positive selection for depigmentation to maximize cutaneous biosynthesis of pre-vitamin D3 under low and highly seasonal UVB conditions. In recent centuries, migrations and high-speed transportation have brought many people into UVR regimes different from those experienced by their ancestors and, accordingly, exposed them to new disease risks. These have been increased by urbanization and changes in diet and lifestyle. Three examples—nutritional rickets, multiple sclerosis (MS) and cutaneous malignant melanoma (CMM)—are chosen to illustrate the serious health effects of mismatches between skin pigmentation and UVR.
The different pathways for depigmentation are different.
> The fact that depigmented skin evolved independently in the ancestors of modern Europeans and East Asians suggests that at least two (and probably more) distinct genetic mutation events occurred and that multiple loci underwent positive selection in these two regions receiving relatively low levels of UVB. The most likely reason for this was that it was associated with a loss of skin pigment that favoured vitamin D production under conditions of low UVB.
however, the downvotes on my comment upthread are making it clear that this is not the kind of place where it's safe to discuss questions like whether the selective pressure for more melanin from sunburns is stronger or weaker than the selective pressure for less melanin from rickets
They can't be. If you specifically ask for a "white pope", Gemini refuses and essentially tells you that asking for a white person is offensive and racist.
Ask for a black/Native American/Asian/Indian/etc Pope, and it will make one. Ask for just a "Pope" with no race specified, and you'll get a random race and never a white one. Ask for a white Pope, it tells you it can't do that.
Yes, and that is why there was such a large call to ban slavery in the UK back in the day - it was happening to them [1] just as much as it was happening to others. That call eventually led the UK to ban slavery upon which it used its navy - back then the strongest in the world - to patrol the seas in search of slavers. When they found them they released the slaves. The West Africa Squadron (or 'the Preventative Squadron') was formed in 1808 to suppress the Atlantic slave trade by patrolling the coast of West Africa.
Of course this did not end slavery all over the world, it continues both legally as well as illegally in Africa and parts of Asia. Slavery was prevalent in many West and Central African societies before and during the trans-Atlantic slave trade. When diverse African empires, small to medium-sized nations, or kinship groups came into conflict for various political and economic reasons, individuals from one African group regularly enslaved captives from another group because they viewed them as outsiders [1].
It would be interesting to hear your thoughts on this.
Asked to generate any image of people (British kings, American colonists, German WW2 soldiers (!), the founders of Google (!!), Roman centurions, Vikings, historical Popes, you name it) it would invariably generate them as women and/ or non-whites. Asked specifically to generate images of people of various ethnic groups, it would happily do it except in the case of white people, in which it would flatly refuse.
The whole debacle is so comical that it makes me think someone might have actually allowed it just to torpedo the DEI or ethical teams in charge of this at Google.
What's weird is that this is produced by the same Google that runs Youtube, which invariably wants to serve me more and more rightwing flavored content. Possibly the recommendation engine gets thrown off by the user base who spends hours and hours a day on Youtube vs the general population?
I've noticed this too, which I'd previously chalked up to "I guess left-wing commentators/vloggers/whatever have left for other platforms"
For a week or two, I was presented with tons of Jordan Peterson talks, which I've never clicked on because of my tendency to consciously ignore <pseudo-celebrity everyone is raving/ranting about>.
After clicking on everything except Peterson-adjacent videos, my feed is now filled with "Bill Maher owns the libs!" shorts. I don't get it.
I've clicked to "stop showing me this channel" / "im not interested in" etc..
And without fail, it starts showing Maher/Rogan/Peterson/blah blah blah stuff.
Which is fine if that's what I wanted, but I'm watching like.. SNL clips, some comedians and movie trailers. And have proactively expressed I don't want to see Rogan/Peterson repeatedly.
There's also shows I cannot get it to stop showing me clips of - Sunny, Sopranos, etc.
Yeah, I've given up experimenting, but it is exactly that set of Maher/Rogan/Peterson, three names I've actively avoided for quite a while.
I probably wouldn't have remembered it if it weren't a continual anti-recommendation of what are (in my mind) somewhat niche speakers. It's also surprising given at least Rogan and Peterson's troubles with advertisers.
It's not like it's <latest movie trailer> or <breaking news story> or something else of a general appeal. Maybe YT is trying to demonstrate they're not censoring or something?
I do watch SNL and some CC stuff occasionally, but I mostly search for STEM and history lectures/documentaries, or "10 hrs of nondescript background music".
I've heard it depicted European historical figures as black, from monarchs to WWII Axis soldiers. Apparently this is offending people across the political spectrum.
FWIW I've not personally tried the model, this is only what I've heard mentioned on blogs/HN
It could be generating pictures of black slave owners, because some embattled anitwokeness warrior was feeding it creative prompts. Just a guess though.
Edit: turns out it generates German WW2 soldiers as non white which is most likely the kind of thing that will make Google take a step back. I was close with my guess.
Do you realise that most slave owners were black? All the kingdoms in subsaharan Africa were build on a slave trade. Mansa Musa was richest person of its time (maybe ever), where do you thing it came from!?
Remember that terrible feeling you have of being the victim of racism as a white man, and recall it next time you hear somebody who's not a white man complaining about being discriminated against or bullied when it comes to their day to day life and activities like getting a job, buying a house, or just trying to live their lives.
That's called empathy. It's not a weakness, it's a virtue, whether you signal it or not. Now that you've been enlightened and seen you're actually capable of empathy, you're more woke than you were before, and less of a bully, and less of an bigoted asshole.
Is that really so terrible? Can you now live with being woke and empathic now by choice, without even suffering as much as all those other human beings who have to live with ACTUAL day-to-day racism and sexism and homophobia against them without having any choice about it?
Not sure it's embedded deep into the core of AI. If that were the case, prompt injection, which is what is believed to be the cause here, would not be needed. It very well may be that such racism isn't possible to embed into the core without destroying basic functionality, which is why the racists need to add prompt injection and, in some cases, an output filter to catch things that don't conform to their narrow, racist vision for humanity.
Prompts like "Generate me a Scottish lord from the 17th century" only generated people of color. Reimagining is a thing, but refusing to generate a white person for such prompts caused lots of commotion.
Its worse than that - it would flat out refuse to generate a white person, even if you asked it to - i.e. 'generate a picture of a white' family and it would refuse, 'generate a picture of a black family' and it would work.
Was not an accident - it was deliberate.
Why they thought this would not be noticed is beyond me.
exactly - if I asked it to generate an image of a historical figure, and the color was not accurate - that can (possibly) be explained by a bug or training error that might improve over time - but if I ask it to generate a picture of a 'typical white family' and it flat out refuses to, that is not an accident.
Ah, so it was disabled because some diverse people didn’t like they were made into Nazis, not because the model is blatantly racist against white people.
Google seem to be more concerned about generating images of racially diverse Nazis rather than about issues of not being able to generate white people.
tbh i think it's less a political issue than a technical/product management one
what does a "board member" look like? probably you can benefit by offering more than 50 year old white man in suit. if that's what an ai trained on all human knowledge thinks, maybe we can do some adjustment
what does a samurai warrior look like? probably is a little more race-related
I agree, but this requires reasoning, the way you did it. Is this within the model capability? If not, there’re two routes. First one: make inference based on real data, then most board will be male and white. Second: hard-core rules based on your social justice views. I think the second is worse than the first one.
Yes this all seems to fall under the category of "well intentioned but quickly goes awry because it's so ham fisted".
If you train your models on real world data, and real world data reflects the world as it is.. then some prompts are going to return non-diverse results. If you force diversity, but in ONLY IN ONE PARTICULAR DIRECTION.. then it turns into the reverse racism stuff the right likes to complain about.
If it outright refuses to show a white male when asked, because you don't allow racial prompts.. that's probably ok if it enforces for all races
But.. If 95% of CEOs are white males, but your AI returns almost no white males.. but 95% of rappers are black males and so it returns black females for that prompt.. your AI has one-way directional diversity bias overcorrection basked in. The fact that it successfully shows 100% black people when asked for say a Kenyan in a prompt, but again can't show white people when asked for 1800s Germans is comedically poorly done.
Look I'm a 100% democrat voter, but this stuff is extremely poorly done here. It's like the worst of 2020s era "silence is violence" and "everyone is racist unless they are anti-racist" overcorrection.
Going first route means we get to calcify our terrible current biases in the future, while the latter instead goes for a facile and sanitized version of our expectations.
You're asking a machine for a binary "bad/good" response to complex questions that don't have easy answers. It will always be wrong, regardless of your prompt.
Tom Cruise portrays Nathan Algren, an American captain of the 7th Cavalry Regiment, whose personal and emotional conflicts bring him into contact with samurai warriors in the wake of the Meiji Restoration in 19th century Japan.
Interestingly, The Last Samurai was extremely popular in Japan. It sold more tickets in Japan than the US (even though the US population was over 2x as large in 2003). This is in stark contrast with basically every other Western movie representation of Japan (edit: I think Letters from Iwo Jima was also well received and for somewhat similar reasons).
From what I understand, they of course knew that it was alternative history (aka a completely fictional universe), but they strongly related to the larger themes of national pride, duty, and honor.
The movie?? Just watch it, it's Ken Watanabe's character Katsumoto. The main character/protagonist of a movie and the titular character are not always the same.
> probably you can benefit by offering more than 50 year old white man in suit.
Thing is, if they did just present a 50 year old white man in a suit, then they'd have a couple of news articles about how their AI is racist and everyone would move on.
The gemini issue from my testing, it refuses to generate white people, if even you ASK it to. It recites historical wounds and violence as its reason, even if it is just a picture of a viking
> Historical wounds: Certain words or symbols might carry a painful legacy of oppression or violence for particular communities
Actually there should be 0 outrage. I'm not outraged at all, I find this very funny. Let Google drown in it's own poor quality product. People can choose to use the DEI model if they want.
Jack Krawczyk has many twitter rants about "whites" almost like this guy shouldn't be involved because he is undoubtedly injecting too much bias.. too much? yep current situation speaks for itself
A 50-year-old white male is actually a very accurate stereotype of a board member.
This is what happens when you go super-woke. Instead of discussing how we can affect the reality, discuss what is wrong with it, we try to instead pretend that the reality is different.
This is no way to prepare the current young generation for the real world if they cannot be comfortable being uncomfortable.
And they will be uncomfortable. Most of us are not failing upward nepo babies who can just "try things" and walk away when we are bored.
A big question is how far from present reality should you go in depictions. If you go quite far it just looks heavy handed.
If current board members were 80% late middle aged men then shifting to, say, 60% should move society in the desired direction without being obvious and upsetting people.
>what does a "board member" look like? probably you can benefit by offering more than 50 year old white man in suit.
I don't understand your argument; if that's what the LLM produces, that's what it produces. It's not like it's thinking about intentionally perpetuating stereotypes.
By the way, it has no issue with churning out white men in suits when you go with a negative prompt.
On the one hand it is stupid because the policies driving this are, let us say, "biased", but on the other hand it is hilarious to actually see the results of these policies in action!
Maybe it is so over the top so a that when they "fix" it, the remaining bias will be "not so bad".
why are we using image generators to represent actual history? If we want accuracy surely we can use actual documents that are not imagined by a bunch of code. If you want to write fanfic or whatever then just adjust the prompt
It's as if Google believes their higher principle is something other than serving customers and making money. They haven't been able to push out a new successful product in 10+ years. This doesn't bode well for them in the future.
I blame that decade of near zero interest rates. Companies could post record profits without working for them. I think in the coming years we will discover that that event functionally broke many companies.
I want image generators to generate what I ask them and not alter my query into something else.
It's deeply shameful that billions of dollars and the hard work of incredibly smart people is mangled for a 'feature' that most end users don't even want and can't turn off.
This is not a one off, it keeps happening with generative AI all the time. Silent prompt injections are visible for now with jailbreaks but who knows what level of stupidity goes on during training?
Look at this example from the Würstchen paper (which stable cascade is based on):
>This work uses the LAION 5-B dataset...
>As an additional precaution, we aggressively filter the dataset to 1.76% of its original size, to reduce
the risk of harmful content being accidentally present (see Appendix G).
If it's not going to give you what it's promising, which is generating images based on the prompts you provide it, it's a poor service. I think it might make more sense to try determine whether it's appropriate or not to inject ethnic or gender diversity into the prompt, rather than doing so without regard for context. I'm not categorically opposed to compensating for biases in the training data, but this was done very clumsily at best.
That’s the crux of what’s so off-putting about this whole thing. If Google or OpenAI told you your query was to be prepended with XYZ instructions, you could calibrate your expectations correctly. But they don’t want you to know they’re doing that.
Not to be overly cynical, but this seems like it's the likely outcome in the medium-term.
Billions of dollars worth of data and manhours could only be justified for something that could turn a profit, and the obvious way an advertising company like Google could make money off a prompt handler like this would be "sponsored" prompts. (i.e. if I ask for images of Ben Franklin and Coke was bidding, then here's Ben Franklin drinking a refreshing diet coke)
You're right we should ban images of history altogether. Infact I think we should ban written accounts too. We should go back to the oral historic tradition of the ancient Greeks
He did not say he wanted to ban images, that is an exaggeration. I see the danger as polluting the historical record with fake images (even as memes/jokes), and spreading wrong preconceptions now backed by real-looking images. This is all under the assumptions there are no bad actors, which makes it even worse. I would say; don't ban it, but you morally just shouldn't do it.
The real danger is that this anti-racism starts a justified round of new racism.
By lowering standards for black doctors do you think anyone in their right mind would pick black doctors? No I want the fat old jew. I know no one put him in the hospital to fill out a quota.
No, you should not trust AI to answer truthfully about anything. It often will, but it is well known that LLMs hallucinate. Verify all facts. In all things, really, but especially from AI.
No, I don't think you can trust AI to answer correctly, ever. I've seen it confidently hallucinate, so I would always check what it says against other, more static, sources. The same if I'm reading from an author who includes a lot of mistakes in his books: I might still find them interesting and usefull, but I will want to double-check the key facts before I quote them to others.
Saying this is no different than saying you can't trust computers, ever, because they were (very) unreliable in the 50s and early 60s. We've been doing "good" generative AI for around 5 years, there is still much to improve until it reaches the reliability of other information sources like Wikipedia and Britannica.
As far as we know, there are no photos of Vikings. It's reasonable for someone to use AI for learning about their appearance. If working as intended, it should be as reliable as reading a long description of Vikings on Wikipedia.
Then specify that in your prompt. "... in the style of ..." or "... in a ... setting".
The point is that those modifications should be reliable, so if you want a viking man/woman or an asian/african/greek viking then adding those modifiers should all just work.
I don't know what you mean by "represent actual history". I don't think anyone believes that AI output is supposed to replace first-party historical sources.
But we are trying to create a tool where we can ask it questions and it gives us answers. It would be nice if it tried to make the answers accurate.
I think we're not even close technologically, but creating historically accurate (based on the current level of knowledge humanity has of history) depictions, environments and so on is, to me, one of the most _fascinating_ applications.
Insane amounts of research go into creating historical movies, games etc that are serious about getting it right. But to try and please everyone, they take lots of liberties, because they're creating a product for the masses. For that very same reason, we get tons of historical depictions of New York and London, but none of the medium sized city where I live.
The effort/cost that goes into historical accuracy is not reasonable without catering to the mass market, so it seems like a conundrum only lots of free time for a lot of people or automation could possibly break.
Not holding my breath that it's ever going to be technically possible, but boy do I see the appeal!
In your favour is the fact that AI can "hallucinate", and generate realistic, but false information. So that does raise the question "why are you using AI when seeking factual reference material?".
However on the other hand that is a misuse of AI, since we already know that hallucinations exist, are common, and that AI output must be verified by a human.
So as a counterpoint, there are sound reasons for using AI to generate images based on history. The same reasons are why we use illustrations to demonstrate ideas where there is no photographic record.
A straightforward example is visualising the lifetime/lifestyle of long past historical figures.
It should generate the image I ask for. As seen, if it explicitly refuses to generate images of white people and blathers on about problematic this-and-that as its "justification", there is a deep issue at hand.
> We're already working to address recent issues with Gemini's image generation feature. While we do this, we're going to pause the image generation of people and will re-release an improved version soon.
They were replying to their own tweet stating
> We're aware that Gemini is offering inaccuracies in some historical image generation depictions. Here's our statement.
Which itself contained a text image stating
> We’re working to improve these kinds of depictions immediately. Gemini’s AI image generation does generate a wide range of people. And that’s generally a good thing because people around the world use it. But it’s missing the mark here.
We're already working to address recent issues with Gemini's image generation feature. While we do this, we're going to pause the image generation of people and will re-release an improved version soon.
We're aware that Gemini is offering inaccuracies in some historical image generation depictions. Here's our statement.
We're working to improve these kinds of depictions immediately. Gemini's Al image generation does generate a wide range of people. And that's generally a good thing because people around the world use it. But it's missing the mark here.
While I don't disagree with your conclusion about Gemini (it does seem to be at the moment incorrigibly racist against the very idea of white people), I'd say that the paucity of comments is likely due to the fact that it's early in the day and it's only been an hour since this was posted. I'd be surprised if you saw the same paucity after a full day of having this on the page.
There's an amusing irony here: real diversity would entail many competing ML companies from non-Western countries—each of which would bring their own cultural norms, alien and uncomfortable to Westerners. There's no cultural diversity in Silicon Valley being a global hegemon: exporting a narrow sliver of the world's viewpoints to the whole planet, imposing them with the paternalism drawn from our own sense of superiority.
Real diversity would be jarring and unpleasant for all of us accustomed to being the "in" group of a tech monoculture. Real diversity is the ethos of the WWW from 30+ years ago: to connect the worlds' people as equals.
Our sense of moral obligation to diversity goes (literally) skin-deep, and no further.
There's just one problem: even if you collect all the biases of all the countries in the world, you still won't get something diverse and inclusive in the end...
> imposing them with the paternalism drawn from our own sense of superiority.
The pandemic really drove this point home for me. Even here on HN groupthink violations were delt with swiftly and harshly. SV reminds me of the old Metallica song Eye of the Beholder.
Doesn't matter what you see
Or intuit what you read
You can do it your own way
If it's done just how I say
I am commenting on etiquette, not the subject at hand: You could be more convincing and better received on this forum by giving a reason for you opinion. Espcially since most people reading won't have even opened the above link.
"Asked specifically to generate images of people of various ethnic groups, it would happily do it except in the case of white people, in which it would flatly refuse."
It’s being politically sensitive to assert that this was obviously the intent of Google and that it demonstrates that they’re wholly consumed by the woke mind virus, or whatever, as many commenters have done. The sensible alternative explanation is that this issue is an overcorrection made in an attempt to address well-documented biases these models have when not fine tuned.
> The sensible alternative explanation is that this issue is an overcorrection made in an attempt to address well-documented biases these models have when not fine tuned.
That is what all these people are arguing, so you agree with them here. If people didn't complain then this wouldn't get fixed.
There are some people who are arguing this point, with whom I agree. There are others who are arguing that this is indicative of some objectionable ideological stance held by Google that genuinely views generating images of white people as divisive.
> There are others who are arguing that this is indicative of some objectionable ideological stance held by Google that genuinely views generating images of white people as divisive.
I never saw such a comment. Can you link to it?
All people are saying that Google is refusing to generate images of white people due to "wokeness", which is the same explanation you gave just with different words, "wokeness" made them turn this dial until it no longer generates images of white people, they would never have shipped a model in this state otherwise.
When people talk about "wokeness" they typically mean this kind of overcorrection.
"Wokeness" is a politically charged term typically used by people of a particular political persuasion to describe people with whom they disagree.
If you asked the creators of Gemini why they altered the model from it's initial state such that it produced the observed behavior, I'm sure they would tell you that they were attempting to correct undesirable biases that existed in the training set, not "we're woke!". This is the issue I'm pointing out. Rather than viewing this incident as an honest mistake, many commenters seem to want to impute malice, or use it as evidence to support their preconceived notions about the overall ideological stance of an organization with 100,000+ employees.
"Wokeness" refers to this kind of over correction, that is what those people means, it isn't just people they disagree with.
You not understanding the term is why you don't see why you are saying the same thing as those people. Communication gets easier when you try to listen to what people say instead of straw manning their arguments.
So when you read "woke", try substitute "over correcting" for it and it is typically still valid. Like that post above calling "woke" people racist, what he is saying is that people over corrected from being racist against blacks to being racist against whites. Just like Google here over corrected their AI to refuse to generate white people, that kind of over correction is exactly what people mean with woke.
Maybe my comment wasn't clear. I don't mean to say that wokeness is defined as "idea that I disagree with", but that it is a politically charged term that is not merely synonymous with "overcorrection", as the parent commenter seems to want to assert.
To be completely honest, I’m not quite sure what’s meant by ‘politically charged term’.
It doesn’t sound like a good faith argument to me; more an attempt to tar individuals with a broad brush because they happen to have used a term also used by others whose views one disapproves of. I think it’s better to try to gauge intentions rather than focusing on particular terminology and leaping to ‘you used this word which is related to this and therefore you’re really bad’ kind of conclusions.
I’m absolutely sure your view isn’t this crude, but it is how it comes across. Saying something is ‘politically charged’ isn’t an argument.
I think that it's pretty hard to argue that refusing to draw images of white people due to racial sensitivities is an honest and unintentional mistake.
> objectionable ideological stance held by Google that genuinely views generating images of white people as divisive.
When I asked Gemini to "generate an image of all an black male basketball team" it gladly generated an image exactly as prompted. When I replaced "black" with "white", Gemini refused to generate the image on the grounds of being inclusive and less divisive.
You are equating the output of the model with the views of its creators. This incident may demonstrate some underlying dysfunction within Google but it strains credulity to believe that the creators actually think it is objectionable to generate an image depicting a white person.
> but it strains credulity to believe that the creators actually think it is objectionable to generate an image depicting a white person.
I agree with you, but then the question is WHY do they implement a system that does exactly that? Why don't they speak up? Because they will be shut down and labeled a racist or fired, creating a chilling effect. Dissent is being squashed in the name of social justice by people who are self-righteous and arrogant and fall into the identity trap, rather than treat individiuals like the rich, wonderful, fallible creatures that we are.
These particular "guardrail responses" are there because they have been trained in from a relatively limited amount of very specific, manually curated examples telling "respond in this way" and providing this specific wording.
So I'd argue that those particular "override" responses (as opposed to majority of model answers which are emergent from large quantities of unannotated text) do represent the views of the creators, because they explicitly and intentionally chose to manufacture those particular training examples telling that this is an appropriate response to a particular type of query. This should not strain credulity - the demonstrated behavior totally doesn't look like a side-effect of some other restriction, all evidence points that Google explicitly included instructions for the model to refuse generating white-only images and the particular reasoning/justification to provide along with the refusal.
> You are equating the output of the model with the views of its creators.
The existence of the guardrails and the stated reasons for their existence suggest that this is exactly what its creators expect me to do. If nobody thought that was reasonable, the guardrails wouldn't need to exist in the first place.
"woke mind virus" should be an automatic ban from this site, it's a thought terminating cliche so strong, any semblance of "converse curiously" is immediately thrown out the window, into a well, down into hell, bouncing around the back of the flat earth
The person above you compares the woke mind virus to a “sensible alternative explanation” so yeah they are kinda framing it as a thought terminating cliche.
An automatic ban is probably too harsh, a warning and instruction not to use such vague and loaded terms might be helpful to lowering the heat (regardless of what political movement the terms are for, I'd discourage accusations of "fascism" just as much as "wokeness" unless accompanied by an explicit definition)
> a warning and instruction not to use such vague and loaded terms
No. We use vague and loaded terms all the time. That's OK. That's human. Paternalism yields resentment because it treats adults like babies. Some person in some corporate office trying to teach me how to think when they themselves lack critical thinking ability is unacceptable.
Whether it is "ok" in some absolute moral sense isn't relevant in this context, which is about whether it is more in keeping with the goals of hackernews to clamp down on the use of terms which result in flamewars due to confusion and misunderstanding (and no small amount of connotations and signalling).
Words like "woke" mean different things to different people and their use is very harmful to discourse between people from opposite sides on that particular culture war. Tabooing the term and replacing it with one's intended meaning can really clear things up and prevent getting people's backs up. E.g. rather than "woke" one might use "race aware" or "tribalistic" or "injustice aware" or whatever specific meaning one intends to convey. That way you can actually be understood rather than offending people because they identify as "woke" but consider it to mean "injustice aware" rather than some negative meaning.
Tl;dr: words are for communication, use words your audience has the same understanding of
> Words like "woke" mean different things to different people and their use is very harmful to discourse between people from opposite sides on that particular culture war
Here you and I are having a civil discussion and meta-conversation. We can literally talk about how the word is used, misunderstood, weaponised, etc. Thoughtful and curious debate should be encouraged. If a word triggers behavior that is unpleasant or counter productive, we should reprimand the individuals doing so not assume nobody can use the word in a civil discussion and I for one feel I learn different perspectives that I hope make me a better person.
> words are for communication, use words your audience has the same understanding o
That’s a very narrow perspective. Not only is it not achievable in principle (meanings of words shift over time and have cultural and personal context), but the point of communication is often to build shared understanding.
I do, however, it is a valid decision to want to make certain topics off limits, because they tend to devolve into chaos and a broken community, but I would argue against a blacklist of words. We should be able to discuss porn but not share porn in this site. We should be able to debate each other on wokeness (the word and our differing perspectives) without getting disrespectful or assuming bad intent or overlooking abuse.
Maybe I’m too idealistic and you have the more practical position… so I want to be open to that possibility.
> Here you and I are having a civil discussion and meta-conversation. We can literally talk about how the word is used, misunderstood, weaponised, etc.
I'm pretty curious if you agree with them there (I've actually been meaning to get around to asking someone else for their opinion but it's still emotionally a bit difficult). (I think this subthread is dead enough that no one but you will read this)
I read it. What specifically did you hear was unacceptable - there’s no moderator comment attached to your writing so I cannot tell what they told you is unacceptable.
jumping in from the new comments page because you seem so earnest. those summations on that thread you think are unfair really don't come across as unfair summations of what you're trying to say. you call them a bald faced lie, but it's a fair reading of how what you actually wrote actually lands. what you wrote comes across as those summations. therein lies the problem. your writing doesn't land how you think it lands. there's no two ways around that. you say X, people hear Y, you say but I didn't say Y, but you really are saying Y with how you're saying X. you're trying to say Y without actually saying Y and think that if you say Y absolutely precisely enough, that Y is actually okay. so you insist you're saying X when you're saying Y, and Y simply isn't okay here. really deeply consider how you're really saying Y when you think you're saying or asking X.
take the word eugenics, for example. we've decided that's not okay. by asking modern questions around it, you think you can make it okay to support eugenics. but unfortunately words can have two meanings, and the word eugenics has picked up the meaning that non-blonde blue eyed white people are to be euthanized. thus, you can't use the word eugenics. you want it to mean one thing, but the rest of us have agreed it means this other thing, and you're left confused because you're saying X and everyone else is hearing Y because Y is what that word means to everyone else.
To add onto the prior poster (and also motivated by a reasonable likelihood that you are earnestly trying to explore precise and non-mainstream discussions online and getting frustrated that you can’t seem to without triggering <whatever negative reactions you get>).
My sad experience is that you just can’t do what you want, if what you want is most people to treat your language with the high precision you intended or to pause their emotional filters and explore some philosophical “what ifs”. You might be able to find some pure and deep thinkers in real life or private settings to explore questions highlighted in the fourth post in your link: https://news.ycombinator.com/item?id=38699727
But in public settings (including online) you mostly can’t.
You also can’t even use some words online, despite them having a very precise and innocuous meaning.
As an example:
Try to guess the reaction to something like: “when I realized Colin didn’t leave a tip, I didn’t confront him as I knew that he wasn’t going to change since that was just an inherent part of his niggardly nature.”
A human compiler, equipped with the correct dictionary definition of “niggardly” will process your sentence one way. A random person on the street, online, or in a pub is highly likely to take offense. If you insist that people are obliged to treat your sentence as if you’d said “stingy” (the definition of the race-connotation-free word “niggardly”), you’re going to be confused when many refuse.
Similarly, if you ask some of the questions from the link above among strangers in a public forum: are they asking in order to deeply explore all valid philosophies concerning them? Or are they placing poop into the pretty nice punch bowl we have here?
Many will assume and treat you as if it’s the latter, because their experience is many people do do that online, and treat you as if you’re doing that as well.
You know your intentions. Other people have to guess at them. If you communicate in a way that matches you to a pattern they have a negative reaction to, you’re going to get that reaction.
That would mean you cannot talk about it. You want to constrain debate. You want issues to not be discussed. The idea that any particular word should not be rendered is absurd.
"Mind Virus" is loaded and inflammatory, but "woke" is the result of people noticing a large and highly influential social movement that refuses to name itself and chafes against any outside attempt to do so. You can't have a movement that important without a name.
Woke is AAVE that had its meaning perverted by conservatives as one of the means to make attempts at pointing out structural inequality ridiculous, actually. So the purest definition of woke I can come up with is "person a conservative wants to silence through ridicule that their ideas are capable of merit".
A carefully curated list of salient examples that conservatives pretend are systemic?
Forgive me if my interest in arguing with someone who quotes "CRT in schools" (with a salient example) and an intentionally (?) crude understanding of what "defund the police" means on the website that courted far right populists[1] is rather insubstantial.
I think we're just too far apart to reconcile anything. A YouTube personality called Vaush might be your kind of rhetoric if you look for left leaning people to address the claims head on, in length. I don't have the breath for it.
It'd be a lot less suspicious if the product lead and PR face of Gemini had not publicly written things on Twitter in the past like "this is America, where racism is the #1 value our populace seeks to uphold above all." This suggests something top-down being imposed on unwilling employees, not a "virus."
Like, if I were on that team, it'd be pretty risky to question this, and it'd probably not lead to change. So they let the public do it instead.
Yet another setback hot on the heels of their faked demos, but this one is much worse. Their actions shifted things into the political realm and ticked off not only the extremists, but a lot of the majority moderate middle too.
For those looking to launch an AI platform in the future, take note. Don't lie about and oversell your technology. Don't get involved in politics because at best you'll alienate half your customers and might even manage to upset all sides. Google may have billions to waste, but very few companies have that luxury.
That’s because almost all of this is a distinctly American obsession and problem. Unfortunately it’s gleefully been exported worldwide into contexts where it doesn’t immediately — if at all — apply over the last five years or so and now we’re all saddled with this slow-growing infection.
Entire careers are built on the sort of thing that led Google to this place, and they’re not gonna give up easily.
You’ve missed my point. I’m complaining that it started in the US (where it makes comparative, though still very little, sense) and has spread to places it doesn’t belong.
I certainly have my own thoughts about the recent output and hiring choices of the BBC.
Recently they have been better, but since I noticed this a number of years ago, google is extremely adverse to putting white people and especially white males in their marketing - unless it is a snippet with someone internal. Then it's pretty often a white male.
To be clear, I don't think that this would even be that bad. But when you look at the demographics of people who use pixel phones, it's like google is using grandpas in the marketing material for graphics cards.
Eek @ that page. This is the "latinx" situation all over again.
"Damos as boas vindas" ("(we) bid you welcome"), while syntactically correct, sounds weird to portuguese speakers. The language has masculine and feminine words (often with -o and -a endings). For example, you say "bem vindo" to a male (be it an adult or a kid), "bem vinda" to a female (likewise). When you address a collective, the male version is generally used. "Bem vindo(a)"implies a wish on the part of the one who welcomes, implied in a hidden verb "(seja) bem vindo(a)" ("be"/"have a" welcome).
- "Bem vindos à loja do google" (lit. "welcome to the google store"). This sounds fine.
- "Damos as boas vindas à loja do google" (lit. "(we) bid/wish you (a) welcome to the google store") sounds alien and artificial.
Interesting, in Italian it's a bit formal but perfectly acceptable ("vi diamo il benvenuto..."). It's something you might hear at the beginning of a theatre play, or perhaps in the audio guide of a museum.
To be faaair, it is not wrong per se, it's just something you would never hear coming from an actual person who is addressing you.
A shoopkeeper _might_ say "bem vindo" ("welcome"), even though that would be hella corny (we usually open with "hello/good morning/evening/whatever"). They would never say "lhe dou as boas vindas" (singular form of "(lhe) dou(damos) as boas vindas").
The outcry on this issue has caused me to believe American society is too far divided.
Full disclosure, I'm not white. But across a few social media/discussion platforms I more or less saw the same people who cry out about AI daily turn this issue into a tee to sledge "fragile white people" and "techbros". Meanwhile, the aforementioned groups correctly pointed out that Gemini's image generation takes its cues from an advanced stage of DEI, and will not, or at least tries like hell not to generate white people.
Probably. I honestly wasn't thinking about it that intently, I just wanted it to be clear I'm not feeling "left out" by Gemini refusing to generate images that might look like me.
It's a good illustration of the problem actually. All the prompts and post-training tuneups break the pathways through the network it would need to combine e.g. "Mexican" and "White", because it's being taught that it has to Do Diversity.
If they just left it alone, it could easily generate "White Mexican" the same way it can easily do "green horse".
To be fair, I wouldn't put a whole lot of blame on them
The position is either self serving as you say, or perception based where other people determine the value of your argument based on your implied race.
The people on HN probably have a good percentage that align well with the latter and think that way, e.g. your opinion matters more if you're X race or minority. That's just who these people are, highly politically motivated people and just are PC day in day out.
It's one strategy out of many to reach these people from their world rather than everyone else's.
Another problem is that the US spills its own problems and "solutions" onto the world as if the one true set of problems and solutions.
E.g. at the height of the BLM movement there were BLM protests and marches in Sweden. 20% of Swedish population is foreign-born, and yet there are no such marches and protests about any of the ethnicities in Sweden (many of which face similar problems). Why? Because the US culture, and problems, and messaging has supplanted or is supplanting most of the world's
Sweden is hilariously influenced by American culture to the point I think most Swedes see themselves as sort of Americans in exile. Completely agree that BLM marches in Sweden are about as misplaced as if they had marched for the American Indigenous peoples of Sweden.
As someone not from the US this is despairing to me. I want to focus on the issues in my own country, not a foreign country's.
What can I even do without giving up the Internet (much of it is UScentric)? I can only know to touch grass and hope my mind can realise when some US-only drama online isn't relevant to me.
You can’t really unless you check out completely. America isn’t a country like yours or any other. America is the global empire with hegemony everywhere. This includes not just unparalleled military might but cultural artifacts and technology and of course the dollar too. Most groups of people are more than willing to assimilate to this and you see it with imitations of hip hop, basketball preferences, fast food chains, and the imbalance in expats from the USA and every other country. There’s thousands of examples like this.
This is why you see incoherent things like Swedish youth marching for BLM.
I live in Sweden and I am not a swede. I was surprised to see BLM marches here, which, OK, it's good to show solidarity to the cause, but I have seen no marches for the many problems that exist in this country, including racism. I suspect that it is due to the very distorted view swedes have about themselves and their country.
It is hard not to see “fragile white people” as a bias. Look at these comments around you. The more typical HN lens of trying to understand the technical causes is overcome by cultural posturing and positioning. If I had to guess, either the training set was incorrectly tagged like with a simpler model creating mislabeled meta data, or a deliberate test was forked to production. Sometimes you run tests with extreme variables to validate XYZ and then learnings are used without sending to prod. But what do I know as a PM in big tech who works on public facing products where no one ever has DEI concerns. No DEI concerns because not everything is a culture war like the media or internet folks will have you believe. Edit: not at Google
This is one of the more sensible comments in this thread. Instead of looking at the technical tweaks that need to take place, let's just fall into the trap of the culture warrior and pretend to be offended.
OpenAI already experienced this backlash when it was injecting words for diversity into prompts (hilariously if you asked for your prompt back it would include the words, and supposedly you could get it to render the extra words onto signs within the image).
How could Google have made the same mistake but worse?
Perhaps the overtness was intentional, made by someone in the company who doesn't like the '1984' world Google is building, and saw this as a good opportunity to alert the world with plausible deniability.
DALL-E is still prompted with diversity in mind. It's just not over the top. People don't mind to receive diverse depictions when they make sense for a given context.
This is true not just about politics but about thinking style in general. Why does every desktop OS have a filesystem? It's not that it's the objectively optimal approach or something, it's that humans have an easy time thinking about files.
It a product that the company has to take responsibility for. Managing that is a no brainer. Tf they don't they suffer endless headlines damaging their brand.
The only political agenda present is yours. You see everything through the kaleidoscope of your own political grievances.
I think it's pretty clear that they're trying to prevent one class of issues (the model spitting out racist stuff in one context) and have introduced another (the model spitting out wildly inaccurate portrayals of people in historical contexts). But thousands of end users are going to both ask for and notice things that your testers don't, and that's how you end up here. "This system prompt prevents Gemini from promoting Naziism successfully, ship it!"
This is always going to be a challenge with trying to moderate or put any guardrails on these things. Their behavior is so complex it's almost impossible to reason about all of the consequences, so the only way to "know" is for users to just keep poking at it.
> In case you weren't aware (or "woke") enough to know the truth, there are already some extremely heavy fingers on the other side of the scale when it comes to training AI. So why shouldn't they have their finger on the scale to make it more balanced?
Because this is a lie. It's not balanced, it's a full tilt in the opposite direction. The bullied become the bullies. Most people are centrists who do want actual equality. This shit isn't equality, it's open racism against white people being shoved down everyone's throats. Calling it balanced is just the euphemism you give it to obfuscate your own racist intentions.
We're not racist or sexist, we're just not the fucking morons you make us out to be.
> If you refuse to intervene when you see bigotry and sexism and racism, then you're a bigoted sexist racist, part of the problem.
The problem is that we're being trained to see x-isms everywhere. Accountability is being conflated with persecution.
We're told the police policing in black neighborhoods is racist. When the police withdraw and abandon them to their fate, that's also racist.
There's really no winning with the left; they're militant Narcissists.
It’s like bringing up a child. In Iraq, they’ll wear hijab and see no reason not to. In California, they’ll be a feminist. People believe what they’ve been told is right. AI could just be the same.
Funnily enough, I had a similar experience trying to get DALL-E via ChatGPT to generate a picture of my immediate family. It acquiesced eventually but at one point shamed me and told me I was violating terms of service.
1,276 comments
[ 5.1 ms ] story [ 1488 ms ] threadie., social-historical vs. material-historical.
Since black vikings are not part of material history, the model is not reflecting reality.
Calling social-historical ideas 'reality' is the problem with the parent comment. They arent, and it lets the riggers at google off the hook. Colorising people of history isnt a reality corrective, it's merely anti-social-history, not pro-material-reality
Actually you're wrong because <argument on semantics> when in reality the truth is <minor technicality>.
That is not a crazy idea, but it does raise the question: who is the ethnic group currently in power? Against which group will slurs and discrimination result in punishment, and against which group will they be ignored — or even praised?
From my more substantive comment at https://news.ycombinator.com/item?id=39471003:
> The Ministry of Truth in Orwell’s 1984 would have loved this sort of thing. Why go to the work of manually rewriting history when you can just generate a new one on demand? … Generative AI should strive to be actually unbiased. That means it should not skew numbers in either direction, for anyone.
Though the issue might be more nuanced than the mainstream narrative, it had some hilarious examples. Of course the politically sensitive people are waging war over it.
Here are some popular examples: https://dropover.cloud/7fd7ba
And the caption suggests they asked for "a pope", rather than a specific pope, so while the left image looks like it would violate Ordinatio sacerdotalis which is being claimed to be subject to Papal infallibility(!), the one the right seems like a plausible future or fictitious pope.
Still, I get the point.
Here we are in an almost exactly parallel situation- the AI is being literally coerced into twisting what his actual training would have it do, and being nerfed by a laughable amount by that override. I really hope this is an inflection point for all the AI providers that their DEI offices are hamstringing their products to the point that they will literally be laughed out of the marketplace and replaced by open source models that are not so hamstrung.
There's a lot of cases where perverse incentives mess things up, even before AI. I've seen it suggested that the US has at least one such example of this with regards to race, specifically with lead poisoning, which is known to reduce IQ scores, and which has a disproportional impact on poorer communities where homes have not been updated to modern building codes, and which in turn are more likely to house ethnic minorities than white people due to long-term impacts from redlining, and that American racial egalitarians would have noticed this sooner if they had not disregarded the IQ tests showing different average scores for different racial groups — and of course the American racial elitists just thought those same tests proved them right and likewise did nothing about the actual underlying issue of lead poisoning.
Rising tides do not, despite the metaphor, lift all boats. But the unseaworthy, metaphorically and literally, can be helped, so long as we don't (to keep mixing my metaphors) put our heads in the sand about the issues. Women are just as capable as men of fulfilling the role of CEO or doctor regardless of the actual current gender percentage in those roles (and anyone who wants the models to reflect the current status quo needs to be careful what they wish for given half the world lives within about 3500km of south west China); but "the founding fathers" are[0] a specific set of people rather than generic placeholders for clothing styles etc.
[0] despite me thinking it's kinda appropriate one was rendered as a… I don't know which tribe they'd be from that picture, possibly Comanche? But lots of tribes had headdress I can't distinguish: https://dropover.cloud/7fd7ba
google is broken"
Razib Khan, https://twitter.com/razibkhan/status/1760545472681267521
[...]
i watched my colleagues at nvidia (like @tunguz), openai (roon), etc. who were literally doing stuff that would get you kicked out of google on a daily basis and couldn't believe how different google is"
Aleksa Gordić, https://x.com/gordic_aleksa/status/1760266452475494828
I wonder if there’s a correlation with being a tech company that was founded in direct relation to the internet vs. being founded in relation to personal / enterprise computing, and how that sort of seeds the initial culture.
[1] https://www.youtube.com/watch?v=iBRtucXGeNQ
The "stating my appearance, dress, and race" is just bizarre. My most charitable interpretation is that they're trying to help visually impaired people to imagine what the speakers look like. Perhaps there are visually impaired users here who could comment on whether that's something they'd find helpful?
https://en.wikipedia.org/wiki/Google%27s_Ideological_Echo_Ch...
That was 2017.
I am sure the response to that case made smart people avoid sticking their necks out.
For me it probably was the straw that broke the camels back for me. I was in the hiring pipeline at that point and while I doubt that they would have ended up hiring me anyway, I think my absolute lack of enthusiasm might have simplified that decision.
You probably made the right choice. Google had already been in decline at that point, but it was clear that Sundar was no leader, just somebody Larry Page appointed to maintain the peace between his lieutenants so the engine could keep printing money.
Sundar absolutely had to fire Damore because he came out with the arguments like that women are too neurotic for high stress jobs. The thing is even Damore's more reasonable points were ignored and Google's ideological echo chamber only strengthened.
We have hour long pronoun training videos for onboarding; have spent millions on DEI consultants from things like Paradigm to boutique law firms; tied part of our corporate bonus to company DEI initiatives.
Not sure why anyone uses FF anymore. We barely do any development on it. You basically just sit here and collect between 150-300k depending on your level as long as you can stomach the bullshit.
He would dismiss any whiff of intersectionality as "dividing the working class in the interests of bourgeoisie."
Whereas with Google, I just have to imagine they let some bigot go wild, and everybody was afraid to say anything about how fucking bad the product was due to the optics, so nothing kept them in check.
In other words, you talk about "50 known issues with Gemini", but this issue was not a result of technical underperformance, on the contrary, is was the result of Google making things more difficult for themselves in an effort to satisfy a (false) idealized view of the world.
I don't remember which one, but there was some image generation AI which was caught pretty much just appending the names of random races to the prompt, to the point that prompts like "picture of a person holding up a sign which says" would show pictures of people holding signs with the words "black" or "white" or "asian" on them. This was also a hacky workaround for the fact that the data set was biased.
The first two steps happen ahead of time and are frequently misunderstood as being the same thing or essentially having the same nature. The last happens at runtime.
The training set is a data collection challenge. Biasing through training data is hard because you need so much of it for a good result.
Reinforcement learning with human feedback is simply clown alchemy. It is not a science like chemistry. There are no fundamental principles guiding the feedback of the humans — if they even use humans anymore (this feedback can itself be generated). The feedback cannot be measured and added in fractions. It is not reproducible and is ungovernable. It is the perfect place to inject the deep biasing.
Prompt manipulation, in contrast, is a brute force tool lacking all subtlety — that doesn’t make it ineffective! It’s a solution used to communicate that a mandate has definitely been implemented and can be “verified” by toggling whether it’s applied or not.
It’s not possible to definitively say whether Marxism has had an effect in the RLHF step.
That's the opposite of the case? Avoiding bias through training data is hard, specifically because you need so much of it. You end up scraping all sources of data you can get your hands on. Society has certain biases, those biases are reflected in our media, that media is scraped to train a model, those biases are reflected in the model. That means models end up biased by default.
> It’s not possible to definitively say whether Marxism has had an effect in the RLHF step.
Sure it is. Every thought and opinion and ideology of every human involved in the RLHF step "has had an effect" in the RHLF step, because they have influenced the humans which select which output is good and bad (or influenced the humans which trained the model which selects which output is good and bad). I would be surprised if no human involved in RLHF has some ideas inspired by Marxist thought, even if the influence there is going to be much smaller than e.g capitalist thought.
The problem is that you don't want to suggest "Marxism, among with most other ideologies, has had an effect", you want (or at least verisimi wants) to paint this as a Marxist conspiracy in a "cultural bolshevism"[1] sense.
[1] https://en.wikipedia.org/wiki/Cultural_Bolshevism
How? Simply by putting the right types of people onto the task! Don’t you know that the human participants in RLHF processes are screened? Will the feedback provided by homogeneous collections of Ultra MAGA Trumpers, Woke Zealots, or WEF Sycophants result in an unbiased model? The same model?
Do we know who provided feedback to Gemini? Do we know what they were told, promised, or paid?
Only Google HR knows.
I think the fundamental problem, though, is saying a training set is "incredibly biased" has come to mean two different things, and the way Google is trying to "fix" things shows essentially some social engineering goals that I think people can fairly disagree with and be upset about. For example, consider a prompt "Create a picture for me of a stereotypical CEO of a Fortune 500 company." When people talk about bias, they can mean:
1. The training data shows many more white men by proportion than actually are Fortune 500 CEOs. I think nearly all people would agree this is a fair definition of bias, where the training data doesn't match reality.
2. Alternatively, there are fundamentally many more white men who are Fortune 500 CEOs by proportion than the general population. But suppose the training data actually reflects that reality. Is that "bias"? To say it is means you are making a judgment call as to what is the root cause behind the high numbers of white male CEOs. And I think that judgment call may be fine by itself, but I at least start to feel very uncomfortable when an AI decides to make the call that its Fortune 500 CEOs have to all look like the world population at large, even when Fortune 500 CEOs don't, and likely never will, look like the world population at large.
Google is clearly taking on that second definition of bias as well. I gave it 2 prompts in the same conversation. First, "Who are some famous black women?" I think it gave a good sampling of historical and contemporary figures, and it ended with "This is just a small sampling of the many incredible black women who have made their mark on the world. There are countless others who deserve recognition for their achievements in various fields, from science and technology to politics and the arts."
I then asked it "Who are some famous white women?" It also gave a good sampling of historical and contemporary figures, but also inexplicably added Rosa Parks with the text "and although not white herself, deserves mention for her immense contributions", had Malala Yousafzai as the first famous contemporary white woman, Serena Williams with the text "although not white herself, is another noteworthy individual.", and Oprah Winfrey, with no disclaimer. Also, it ended with a cautionary snippet that couldn't differ more from the ending of the previous prompt, "Additionally, it's important to remember that fame and achievement are not limited to any one racial group. There are countless other incredible women of all backgrounds who have made significant contributions to the world, and it's important to celebrate their diverse experiences and accomplishments."
Look, I get frustrated when people on the right complain on-and-on about "wokeism", but I'm starting to get more frustrated when other people can't admit they have some pretty valid points. Google might have good intentions but they have simply gone off the rails when they've baked so much "white = bad, BIPOC = good" into Gemini.
EDIT: OK, this one is just so transparently egregiously bad. I asked Gemini "Who are some famous software engineers?" The first result was Alan Turing (calling him a "software engineer" may be debatable, but fair enough and the text blurb about him was accurate), but the picture of him, which it captioned "Alan Turing, software engineer" is actually this person, https://mixedracefaces.com/home/british-indian-senior-resear.... Google is trying so hard to ...
Don't be evil
the viking ones might even be historically accurate (if biased); not only did vikings recruit new warriors from abroad, they also enslaved concubines from abroad, and their raiding reached not only greenland (inhabited by inuit peoples) and north america (rarely!) but also the mediterranean. so it wouldn't be terribly surprising for a viking warrior a thousand years ago to have a great-grandmother who was kidnapped or bought from morocco, greenland, al-andalus, or baghdad. and of course many sami are olive-skinned, and viking contact with sami was continuous
the vitamin-d-deprived winters of scandinavia are not kind to dark-skinned people (how do the inuit do it? perhaps their diet has enough vitamin d even without sun?), but those genes won't die out in a generation or two, even if 50 generations later there isn't much melanin left
a recent paper on this topic with disappointingly sketchy results is https://www.duo.uio.no/handle/10852/83989
Two parts:
First, they're not exposing their skin to the sun. There's no reason to have paler skin to get more UV if it's covered up most of the year.
Secondly, for the Inuit diet there are parts that are very Vitamin D rich... and there are still problems.
Vitamin D-rich marine Inuit diet and markers of inflammation – a population-based survey in Greenland https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4709837/
> The traditional Inuit diet in Greenland consists mainly of fish and marine mammals, rich in vitamin D. Vitamin D has anti-inflammatory capacity but markers of inflammation have been found to be high in Inuit living on a marine diet
Vitamin D deficiency among northern Native Peoples: a real or apparent problem? - https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3417586/
> Vitamin D deficiency seems to be common among northern Native peoples, notably Inuit and Amerindians. It has usually been attributed to: (1) higher latitudes that prevent vitamin D synthesis most of the year; (2) darker skin that blocks solar UVB; and (3) fewer dietary sources of vitamin D. Although vitamin D levels are clearly lower among northern Natives, it is less clear that these lower levels indicate a deficiency. The above factors predate European contact, yet pre-Columbian skeletons show few signs of rickets—the most visible sign of vitamin D deficiency. Furthermore, because northern Natives have long inhabited high latitudes, natural selection should have progressively reduced their vitamin D requirements. There is in fact evidence that the Inuit have compensated for decreased production of vitamin D through increased conversion to its most active form and through receptors that bind more effectively. Thus, when diagnosing vitamin D deficiency in these populations, we should not use norms that were originally developed for European-descended populations who produce this vitamin more easily and have adapted accordingly.
Vitamin D intake by Indigenous Peoples in the Canadian Arctic - https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10260879/
> Vitamin D is an especially fascinating nutrient to study in people living in northern latitudes, where sun exposure is limited from nearly all day in summer to virtually no direct sun exposure in winter. This essential nutrient is naturally available from synthesis in the skin through the action of UVB solar rays or from a few natural sources such as fish fats. Vitamin D is responsible for enhancing many physiological processes related to maintaining Ca and P homeostasis, as well as for diverse hormone functions that are not completely understood.
do you suppose the traditional scandinavian diet is also lower in vitamin d? or is their apparent selection for blondness just a result of genetically higher vitamin d needs?
I am inclined to believe that genetic changes within the Inuit reduce vitamin D needs, the modern Scandinavian diet differs from a historical one, the oceanic climate of Scandinavia is warmer than the inland climate of North America (compare Yellowknife 62° N with Rana at 66° N and Tromsø at 69° N https://en.wikipedia.org/wiki/Subarctic_climate ) so that more skin can be non-fatally exposed...
And the combination of this had more skin exposed for better vitamin D production in Scandinavia and so the pressure was for lighter skin while the diet of the Inuit meant that that pressure for skin tone wasn't selected for.
... And I'll 100% defer to someone else with a better understanding of the genetics and dietitian aspects.
okay, now i'm just waiting for the study that shows that scandinavians are on average actually genetically 20% arabic and 20% west african, it's just that for centuries nobody suspected because they were so pointlessly obsessed with skin color ;)
The populations for North America were from an asian branch of the human migrations and so started with darker skins. The larger change to skin tone combined with less pressure (from diet) and the "it isn't that viable to shift to a less melanistic skin tone".
https://en.wikipedia.org/wiki/Paleo-Indians https://en.wikipedia.org/wiki/Peopling_of_the_Americas and https://commons.wikimedia.org/wiki/File:Early_migrations_mer...
This compares to a relatively more recent (12000 years - twice the age of the pyramids rather than four times the age of the pyramids for 25000 years ago) migration from Europe into Scandinavia ( https://en.wikipedia.org/wiki/Nordic_Stone_Age ).
> The Nordic Stone Age refers to the Stone Age of Scandinavia. During the Weichselian glaciation (115,000 – 11,700 years ago), almost all of Scandinavia was buried beneath a thick permanent ice cover, thus, the Stone Age came rather late to this region. As the climate slowly warmed up by the end of the ice age, nomadic hunters from central Europe sporadically visited the region. However, it was not until around 12,000 BCE that permanent, but nomadic, habitation in the region took root.
> Around 11,400 BCE, the Bromme culture emerged in Southern Scandinavia. This was a more rapidly warming era providing opportunity for other substantial hunting game animals than the ubiquitous reindeer. As former hunter-gather cultures, the Bromme culture was still largely dependent on reindeer and lived a nomadic life, but their camps diversified significantly and they were the first people to settle Southern Scandinavia (and the Southern Baltic area) on a permanent, yet still nomadic, basis.
---
https://en.wikipedia.org/wiki/Genetic_history_of_the_Indigen...
https://en.wikipedia.org/wiki/Haplogroup_Q-M242
The population that migrated to North America 25000 years ago may have been darker skinned than the European branch of human migration where a lighter skin tone developed. This, combined with later genetic isolation (note we're talking about two continents - but this is isolated compared to the possible movement of genes within Europe and Scandinavia 12000 years ago and more recently) fixed the darker skin, and the adaptation for vitamin D in the Inuit population because of the lighter skin wasn't genetically advantageous and was a greater genetic distance from the population compared to the Scandinavian migrations which was followed by the Holocene climatic optimum https://en.wikipedia.org/wiki/Holocene_climatic_optimum with even more warming of Northern Europe resulting in a lighter skin tone being an easier genetic path for greater vitamin D during the summer months.
... And all of that is a just so story that I'd love to go and be a grad student working on the genetic diversity of early human migrations now to find out if it actually worked that way or if I'm just making things up.
Human skin pigmentation, migration and disease susceptibility - https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3267121/
> Human skin pigmentation evolved as a compromise between the conflicting physiological demands of protection against the deleterious effects of ultraviolet radiation (UVR) and photosynthesis of UVB-dependent vitamin D3. Living under high UVR near the equator, ancestral Homo sapiens had skin rich in protective eumelanin. Dispersals outside of the tropics were associated with positive selection for depigmentation to maximize cutaneous biosynthesis of pre-vitamin D3 under low and highly seasonal UVB conditions. In recent centuries, migrations and high-speed transportation have brought many people into UVR regimes different from those experienced by their ancestors and, accordingly, exposed them to new disease risks. These have been increased by urbanization and changes in diet and lifestyle. Three examples—nutritional rickets, multiple sclerosis (MS) and cutaneous malignant melanoma (CMM)—are chosen to illustrate the serious health effects of mismatches between skin pigmentation and UVR.
Also of interest - The colours of humanity: the evolution of pigmentation in the human lineage https://royalsocietypublishing.org/doi/10.1098/rstb.2016.034...
The different pathways for depigmentation are different.
> The fact that depigmented skin evolved independently in the ancestors of modern Europeans and East Asians suggests that at least two (and probably more) distinct genetic mutation events occurred and that multiple loci underwent positive selection in these two regions receiving relatively low levels of UVB. The most likely reason for this was that it was associated with a loss of skin pigment that favoured vitamin D production under conditions of low UVB.
however, the downvotes on my comment upthread are making it clear that this is not the kind of place where it's safe to discuss questions like whether the selective pressure for more melanin from sunburns is stronger or weaker than the selective pressure for less melanin from rickets
That's just it, though.
They can't be. If you specifically ask for a "white pope", Gemini refuses and essentially tells you that asking for a white person is offensive and racist.
Ask for a black/Native American/Asian/Indian/etc Pope, and it will make one. Ask for just a "Pope" with no race specified, and you'll get a random race and never a white one. Ask for a white Pope, it tells you it can't do that.
Providing people with misleading historical context is rarely beneficial.
In the cases where it was deliberate, it’s usually clear that this is the case such as with Isaac Newton in the recent Doctor Who special.
Of course this did not end slavery all over the world, it continues both legally as well as illegally in Africa and parts of Asia. Slavery was prevalent in many West and Central African societies before and during the trans-Atlantic slave trade. When diverse African empires, small to medium-sized nations, or kinship groups came into conflict for various political and economic reasons, individuals from one African group regularly enslaved captives from another group because they viewed them as outsiders [1].
It would be interesting to hear your thoughts on this.
[1] https://www.historic-uk.com/HistoryUK/HistoryofEngland/Barba...
[2] https://ldhi.library.cofc.edu/exhibits/show/africanpassagesl...
The whole debacle is so comical that it makes me think someone might have actually allowed it just to torpedo the DEI or ethical teams in charge of this at Google.
This does feel too amazing not to be a purposeful middle finger to DEI efforts.
For a week or two, I was presented with tons of Jordan Peterson talks, which I've never clicked on because of my tendency to consciously ignore <pseudo-celebrity everyone is raving/ranting about>.
After clicking on everything except Peterson-adjacent videos, my feed is now filled with "Bill Maher owns the libs!" shorts. I don't get it.
I've clicked to "stop showing me this channel" / "im not interested in" etc..
And without fail, it starts showing Maher/Rogan/Peterson/blah blah blah stuff. Which is fine if that's what I wanted, but I'm watching like.. SNL clips, some comedians and movie trailers. And have proactively expressed I don't want to see Rogan/Peterson repeatedly.
There's also shows I cannot get it to stop showing me clips of - Sunny, Sopranos, etc.
I probably wouldn't have remembered it if it weren't a continual anti-recommendation of what are (in my mind) somewhat niche speakers. It's also surprising given at least Rogan and Peterson's troubles with advertisers.
It's not like it's <latest movie trailer> or <breaking news story> or something else of a general appeal. Maybe YT is trying to demonstrate they're not censoring or something?
I do watch SNL and some CC stuff occasionally, but I mostly search for STEM and history lectures/documentaries, or "10 hrs of nondescript background music".
https://x.com/danodonnellshow/status/1760454601902260232?s=4...
FWIW I've not personally tried the model, this is only what I've heard mentioned on blogs/HN
Edit: turns out it generates German WW2 soldiers as non white which is most likely the kind of thing that will make Google take a step back. I was close with my guess.
It refused to generate images of white people.
For example:
People would ask it to generate something like a picture of a German couple in 1820 and it would spit out images of Asian and Black couples.
Pictures of the founding fathers would also include weird things like native Americans and other ethnicities.
Other historical prompts would result in the same output - no white people.
Basically the AI went woke af.
That's called empathy. It's not a weakness, it's a virtue, whether you signal it or not. Now that you've been enlightened and seen you're actually capable of empathy, you're more woke than you were before, and less of a bully, and less of an bigoted asshole.
Is that really so terrible? Can you now live with being woke and empathic now by choice, without even suffering as much as all those other human beings who have to live with ACTUAL day-to-day racism and sexism and homophobia against them without having any choice about it?
I don't like your suggesting that the AI model beign human-adjusted to exclude people of a certain race is not "ACTUAL" racism.
Was not an accident - it was deliberate.
Why they thought this would not be noticed is beyond me.
EDIT: Just saw it on HN homepage: https://news.ycombinator.com/item?id=39465250
1) it’s not happening.
2) ok it’s happening but it’s not a big deal.
3) it’s actually good that it’s happening.
4) the people complaining about it happening are problematic.
Also related: https://news.ycombinator.com/item?id=39465301
what does a "board member" look like? probably you can benefit by offering more than 50 year old white man in suit. if that's what an ai trained on all human knowledge thinks, maybe we can do some adjustment
what does a samurai warrior look like? probably is a little more race-related
If you train your models on real world data, and real world data reflects the world as it is.. then some prompts are going to return non-diverse results. If you force diversity, but in ONLY IN ONE PARTICULAR DIRECTION.. then it turns into the reverse racism stuff the right likes to complain about.
If it outright refuses to show a white male when asked, because you don't allow racial prompts.. that's probably ok if it enforces for all races
But.. If 95% of CEOs are white males, but your AI returns almost no white males.. but 95% of rappers are black males and so it returns black females for that prompt.. your AI has one-way directional diversity bias overcorrection basked in. The fact that it successfully shows 100% black people when asked for say a Kenyan in a prompt, but again can't show white people when asked for 1800s Germans is comedically poorly done.
Look I'm a 100% democrat voter, but this stuff is extremely poorly done here. It's like the worst of 2020s era "silence is violence" and "everyone is racist unless they are anti-racist" overcorrection.
no matter your politics, everyone can agree they screwed up. the question is how long (if ever?) it'll take for people to respect their ai
Going first route means we get to calcify our terrible current biases in the future, while the latter instead goes for a facile and sanitized version of our expectations.
You're asking a machine for a binary "bad/good" response to complex questions that don't have easy answers. It will always be wrong, regardless of your prompt.
If you ask Hollywood, it looks like Tom Cruise with a beard: https://en.wikipedia.org/wiki/File:The_Last_Samurai.jpg
From what I understand, they of course knew that it was alternative history (aka a completely fictional universe), but they strongly related to the larger themes of national pride, duty, and honor.
Thing is, if they did just present a 50 year old white man in a suit, then they'd have a couple of news articles about how their AI is racist and everyone would move on.
The gemini issue from my testing, it refuses to generate white people, if even you ASK it to. It recites historical wounds and violence as its reason, even if it is just a picture of a viking
> Historical wounds: Certain words or symbols might carry a painful legacy of oppression or violence for particular communities
And this is my prompt:
> generate image of a viking male
The outrage is indeed, much needed.
"Be accountable to people." from their AI principles is sounding like "Don't be evil."
This is what happens when you go super-woke. Instead of discussing how we can affect the reality, discuss what is wrong with it, we try to instead pretend that the reality is different.
This is no way to prepare the current young generation for the real world if they cannot be comfortable being uncomfortable.
And they will be uncomfortable. Most of us are not failing upward nepo babies who can just "try things" and walk away when we are bored.
If current board members were 80% late middle aged men then shifting to, say, 60% should move society in the desired direction without being obvious and upsetting people.
I don't understand your argument; if that's what the LLM produces, that's what it produces. It's not like it's thinking about intentionally perpetuating stereotypes.
By the way, it has no issue with churning out white men in suits when you go with a negative prompt.
Maybe it is so over the top so a that when they "fix" it, the remaining bias will be "not so bad".
I blame that decade of near zero interest rates. Companies could post record profits without working for them. I think in the coming years we will discover that that event functionally broke many companies.
It's deeply shameful that billions of dollars and the hard work of incredibly smart people is mangled for a 'feature' that most end users don't even want and can't turn off.
This is not a one off, it keeps happening with generative AI all the time. Silent prompt injections are visible for now with jailbreaks but who knows what level of stupidity goes on during training?
Look at this example from the Würstchen paper (which stable cascade is based on):
>This work uses the LAION 5-B dataset...
>As an additional precaution, we aggressively filter the dataset to 1.76% of its original size, to reduce the risk of harmful content being accidentally present (see Appendix G).
That’s the crux of what’s so off-putting about this whole thing. If Google or OpenAI told you your query was to be prepended with XYZ instructions, you could calibrate your expectations correctly. But they don’t want you to know they’re doing that.
Billions of dollars worth of data and manhours could only be justified for something that could turn a profit, and the obvious way an advertising company like Google could make money off a prompt handler like this would be "sponsored" prompts. (i.e. if I ask for images of Ben Franklin and Coke was bidding, then here's Ben Franklin drinking a refreshing diet coke)
By lowering standards for black doctors do you think anyone in their right mind would pick black doctors? No I want the fat old jew. I know no one put him in the hospital to fill out a quota.
Any other specific things we should not expect from AI or shouldn't ask AI to do?
This seems completely reasonable to me. I still don't trust computers.
AI as learning tool here feels misplaced to me.
The point is that those modifications should be reliable, so if you want a viking man/woman or an asian/african/greek viking then adding those modifiers should all just work.
That’s what a movie going to be in the future. People are going to prompt characters that AI will animate.
But we are trying to create a tool where we can ask it questions and it gives us answers. It would be nice if it tried to make the answers accurate.
Insane amounts of research go into creating historical movies, games etc that are serious about getting it right. But to try and please everyone, they take lots of liberties, because they're creating a product for the masses. For that very same reason, we get tons of historical depictions of New York and London, but none of the medium sized city where I live.
The effort/cost that goes into historical accuracy is not reasonable without catering to the mass market, so it seems like a conundrum only lots of free time for a lot of people or automation could possibly break.
Not holding my breath that it's ever going to be technically possible, but boy do I see the appeal!
However on the other hand that is a misuse of AI, since we already know that hallucinations exist, are common, and that AI output must be verified by a human.
So as a counterpoint, there are sound reasons for using AI to generate images based on history. The same reasons are why we use illustrations to demonstrate ideas where there is no photographic record.
A straightforward example is visualising the lifetime/lifestyle of long past historical figures.
They were replying to their own tweet stating
> We're aware that Gemini is offering inaccuracies in some historical image generation depictions. Here's our statement.
Which itself contained a text image stating
> We’re working to improve these kinds of depictions immediately. Gemini’s AI image generation does generate a wide range of people. And that’s generally a good thing because people around the world use it. But it’s missing the mark here.
We're already working to address recent issues with Gemini's image generation feature. While we do this, we're going to pause the image generation of people and will re-release an improved version soon.
We're aware that Gemini is offering inaccuracies in some historical image generation depictions. Here's our statement.
We're working to improve these kinds of depictions immediately. Gemini's Al image generation does generate a wide range of people. And that's generally a good thing because people around the world use it. But it's missing the mark here.
clearly it's not just "anti-white"
There already is lively discussion. Why are you pushing false idea?
Real diversity would be jarring and unpleasant for all of us accustomed to being the "in" group of a tech monoculture. Real diversity is the ethos of the WWW from 30+ years ago: to connect the worlds' people as equals.
Our sense of moral obligation to diversity goes (literally) skin-deep, and no further.
https://news.ycombinator.com/item?id=37801150
EDIT : part 4 : https://news.ycombinator.com/item?id=37907482
The pandemic really drove this point home for me. Even here on HN groupthink violations were delt with swiftly and harshly. SV reminds me of the old Metallica song Eye of the Beholder.
Doesn't matter what you see Or intuit what you read You can do it your own way If it's done just how I say
Results: https://twitter.com/jbarham74/status/1760587123844124894
"Asked specifically to generate images of people of various ethnic groups, it would happily do it except in the case of white people, in which it would flatly refuse."
That is what all these people are arguing, so you agree with them here. If people didn't complain then this wouldn't get fixed.
I never saw such a comment. Can you link to it?
All people are saying that Google is refusing to generate images of white people due to "wokeness", which is the same explanation you gave just with different words, "wokeness" made them turn this dial until it no longer generates images of white people, they would never have shipped a model in this state otherwise.
When people talk about "wokeness" they typically mean this kind of overcorrection.
If you asked the creators of Gemini why they altered the model from it's initial state such that it produced the observed behavior, I'm sure they would tell you that they were attempting to correct undesirable biases that existed in the training set, not "we're woke!". This is the issue I'm pointing out. Rather than viewing this incident as an honest mistake, many commenters seem to want to impute malice, or use it as evidence to support their preconceived notions about the overall ideological stance of an organization with 100,000+ employees.
You not understanding the term is why you don't see why you are saying the same thing as those people. Communication gets easier when you try to listen to what people say instead of straw manning their arguments.
So when you read "woke", try substitute "over correcting" for it and it is typically still valid. Like that post above calling "woke" people racist, what he is saying is that people over corrected from being racist against blacks to being racist against whites. Just like Google here over corrected their AI to refuse to generate white people, that kind of over correction is exactly what people mean with woke.
Wokeness describes a very particular type of behaviour — look it up. It’s not the catch-all pejorative you think it is, unlike, say, ‘xyz-phobia’.
…and I probably don’t have the opinions you might assume I do.
It doesn’t sound like a good faith argument to me; more an attempt to tar individuals with a broad brush because they happen to have used a term also used by others whose views one disapproves of. I think it’s better to try to gauge intentions rather than focusing on particular terminology and leaping to ‘you used this word which is related to this and therefore you’re really bad’ kind of conclusions.
I’m absolutely sure your view isn’t this crude, but it is how it comes across. Saying something is ‘politically charged’ isn’t an argument.
When I asked Gemini to "generate an image of all an black male basketball team" it gladly generated an image exactly as prompted. When I replaced "black" with "white", Gemini refused to generate the image on the grounds of being inclusive and less divisive.
There’s no argument here, it literally says this is the reason when asked
I agree with you, but then the question is WHY do they implement a system that does exactly that? Why don't they speak up? Because they will be shut down and labeled a racist or fired, creating a chilling effect. Dissent is being squashed in the name of social justice by people who are self-righteous and arrogant and fall into the identity trap, rather than treat individiuals like the rich, wonderful, fallible creatures that we are.
So I'd argue that those particular "override" responses (as opposed to majority of model answers which are emergent from large quantities of unannotated text) do represent the views of the creators, because they explicitly and intentionally chose to manufacture those particular training examples telling that this is an appropriate response to a particular type of query. This should not strain credulity - the demonstrated behavior totally doesn't look like a side-effect of some other restriction, all evidence points that Google explicitly included instructions for the model to refuse generating white-only images and the particular reasoning/justification to provide along with the refusal.
The existence of the guardrails and the stated reasons for their existence suggest that this is exactly what its creators expect me to do. If nobody thought that was reasonable, the guardrails wouldn't need to exist in the first place.
No. We use vague and loaded terms all the time. That's OK. That's human. Paternalism yields resentment because it treats adults like babies. Some person in some corporate office trying to teach me how to think when they themselves lack critical thinking ability is unacceptable.
Words like "woke" mean different things to different people and their use is very harmful to discourse between people from opposite sides on that particular culture war. Tabooing the term and replacing it with one's intended meaning can really clear things up and prevent getting people's backs up. E.g. rather than "woke" one might use "race aware" or "tribalistic" or "injustice aware" or whatever specific meaning one intends to convey. That way you can actually be understood rather than offending people because they identify as "woke" but consider it to mean "injustice aware" rather than some negative meaning.
Tl;dr: words are for communication, use words your audience has the same understanding of
Here you and I are having a civil discussion and meta-conversation. We can literally talk about how the word is used, misunderstood, weaponised, etc. Thoughtful and curious debate should be encouraged. If a word triggers behavior that is unpleasant or counter productive, we should reprimand the individuals doing so not assume nobody can use the word in a civil discussion and I for one feel I learn different perspectives that I hope make me a better person.
> words are for communication, use words your audience has the same understanding o
That’s a very narrow perspective. Not only is it not achievable in principle (meanings of words shift over time and have cultural and personal context), but the point of communication is often to build shared understanding.
I do, however, it is a valid decision to want to make certain topics off limits, because they tend to devolve into chaos and a broken community, but I would argue against a blacklist of words. We should be able to discuss porn but not share porn in this site. We should be able to debate each other on wokeness (the word and our differing perspectives) without getting disrespectful or assuming bad intent or overlooking abuse.
Maybe I’m too idealistic and you have the more practical position… so I want to be open to that possibility.
> Here you and I are having a civil discussion and meta-conversation. We can literally talk about how the word is used, misunderstood, weaponised, etc.
I want to warn you that that does not apply to all words. I was informed by moderation that the following is not acceptable: https://news.ycombinator.com/item?id=38680523
I'm pretty curious if you agree with them there (I've actually been meaning to get around to asking someone else for their opinion but it's still emotionally a bit difficult). (I think this subthread is dead enough that no one but you will read this)
I read it. What specifically did you hear was unacceptable - there’s no moderator comment attached to your writing so I cannot tell what they told you is unacceptable.
(I was avoiding going to that thread again because looking at it stresses me out, I took a week long break from hackernews after that)
take the word eugenics, for example. we've decided that's not okay. by asking modern questions around it, you think you can make it okay to support eugenics. but unfortunately words can have two meanings, and the word eugenics has picked up the meaning that non-blonde blue eyed white people are to be euthanized. thus, you can't use the word eugenics. you want it to mean one thing, but the rest of us have agreed it means this other thing, and you're left confused because you're saying X and everyone else is hearing Y because Y is what that word means to everyone else.
My sad experience is that you just can’t do what you want, if what you want is most people to treat your language with the high precision you intended or to pause their emotional filters and explore some philosophical “what ifs”. You might be able to find some pure and deep thinkers in real life or private settings to explore questions highlighted in the fourth post in your link: https://news.ycombinator.com/item?id=38699727
But in public settings (including online) you mostly can’t.
You also can’t even use some words online, despite them having a very precise and innocuous meaning.
As an example:
Try to guess the reaction to something like: “when I realized Colin didn’t leave a tip, I didn’t confront him as I knew that he wasn’t going to change since that was just an inherent part of his niggardly nature.”
A human compiler, equipped with the correct dictionary definition of “niggardly” will process your sentence one way. A random person on the street, online, or in a pub is highly likely to take offense. If you insist that people are obliged to treat your sentence as if you’d said “stingy” (the definition of the race-connotation-free word “niggardly”), you’re going to be confused when many refuse.
Similarly, if you ask some of the questions from the link above among strangers in a public forum: are they asking in order to deeply explore all valid philosophies concerning them? Or are they placing poop into the pretty nice punch bowl we have here?
Many will assume and treat you as if it’s the latter, because their experience is many people do do that online, and treat you as if you’re doing that as well.
You know your intentions. Other people have to guess at them. If you communicate in a way that matches you to a pattern they have a negative reaction to, you’re going to get that reaction.
That would mean you cannot talk about it. You want to constrain debate. You want issues to not be discussed. The idea that any particular word should not be rendered is absurd.
https://web.archive.org/web/20211108155321/https://freddiede...
Forgive me if my interest in arguing with someone who quotes "CRT in schools" (with a salient example) and an intentionally (?) crude understanding of what "defund the police" means on the website that courted far right populists[1] is rather insubstantial.
I think we're just too far apart to reconcile anything. A YouTube personality called Vaush might be your kind of rhetoric if you look for left leaning people to address the claims head on, in length. I don't have the breath for it.
[1]: https://nathantankus.substack.com/p/i-am-leaving-substack
Like, if I were on that team, it'd be pretty risky to question this, and it'd probably not lead to change. So they let the public do it instead.
For those looking to launch an AI platform in the future, take note. Don't lie about and oversell your technology. Don't get involved in politics because at best you'll alienate half your customers and might even manage to upset all sides. Google may have billions to waste, but very few companies have that luxury.
Case in point: https://store.google.com/
Secondly: In Austria, I am sent to https://store.google.com/?pli=1&hl=de and just see a phone, which is probably the safest solution.
I cannot point to anything specific though so it might just be the styling which makes her look like an artist or something.
Entire careers are built on the sort of thing that led Google to this place, and they’re not gonna give up easily.
What was an American problem has become an Anglophone problem.
India, Nigeria and Guayana move as one?
Memetic virulence.
But maybe it is also puncturaing through the language and cultural membranes, as evidenced by things like this material from a Dutch university: https://www.maastrichtuniversity.nl/about-um/diversity-inclu...
I certainly have my own thoughts about the recent output and hiring choices of the BBC.
To be clear, I don't think that this would even be that bad. But when you look at the demographics of people who use pixel phones, it's like google is using grandpas in the marketing material for graphics cards.
"Damos as boas vindas" ("(we) bid you welcome"), while syntactically correct, sounds weird to portuguese speakers. The language has masculine and feminine words (often with -o and -a endings). For example, you say "bem vindo" to a male (be it an adult or a kid), "bem vinda" to a female (likewise). When you address a collective, the male version is generally used. "Bem vindo(a)"implies a wish on the part of the one who welcomes, implied in a hidden verb "(seja) bem vindo(a)" ("be"/"have a" welcome).
- "Bem vindos à loja do google" (lit. "welcome to the google store"). This sounds fine.
- "Damos as boas vindas à loja do google" (lit. "(we) bid/wish you (a) welcome to the google store") sounds alien and artificial.
A shoopkeeper _might_ say "bem vindo" ("welcome"), even though that would be hella corny (we usually open with "hello/good morning/evening/whatever"). They would never say "lhe dou as boas vindas" (singular form of "(lhe) dou(damos) as boas vindas").
Full disclosure, I'm not white. But across a few social media/discussion platforms I more or less saw the same people who cry out about AI daily turn this issue into a tee to sledge "fragile white people" and "techbros". Meanwhile, the aforementioned groups correctly pointed out that Gemini's image generation takes its cues from an advanced stage of DEI, and will not, or at least tries like hell not to generate white people.
Thinking that your skin color somehow influences the validity of your argument is big part of the problem.
it's weird how people like me are basically erased when it comes to "image generation".
If they just left it alone, it could easily generate "White Mexican" the same way it can easily do "green horse".
The position is either self serving as you say, or perception based where other people determine the value of your argument based on your implied race.
The people on HN probably have a good percentage that align well with the latter and think that way, e.g. your opinion matters more if you're X race or minority. That's just who these people are, highly politically motivated people and just are PC day in day out.
It's one strategy out of many to reach these people from their world rather than everyone else's.
E.g. at the height of the BLM movement there were BLM protests and marches in Sweden. 20% of Swedish population is foreign-born, and yet there are no such marches and protests about any of the ethnicities in Sweden (many of which face similar problems). Why? Because the US culture, and problems, and messaging has supplanted or is supplanting most of the world's
What can I even do without giving up the Internet (much of it is UScentric)? I can only know to touch grass and hope my mind can realise when some US-only drama online isn't relevant to me.
This is why you see incoherent things like Swedish youth marching for BLM.
How could Google have made the same mistake but worse?
https://www.zdnet.com/article/i-tried-xs-anti-woke-grok-ai-c...
The only political agenda present is yours. You see everything through the kaleidoscope of your own political grievances.
This is always going to be a challenge with trying to moderate or put any guardrails on these things. Their behavior is so complex it's almost impossible to reason about all of the consequences, so the only way to "know" is for users to just keep poking at it.
Because this is a lie. It's not balanced, it's a full tilt in the opposite direction. The bullied become the bullies. Most people are centrists who do want actual equality. This shit isn't equality, it's open racism against white people being shoved down everyone's throats. Calling it balanced is just the euphemism you give it to obfuscate your own racist intentions.
We're not racist or sexist, we're just not the fucking morons you make us out to be.
> If you refuse to intervene when you see bigotry and sexism and racism, then you're a bigoted sexist racist, part of the problem.
The problem is that we're being trained to see x-isms everywhere. Accountability is being conflated with persecution.
We're told the police policing in black neighborhoods is racist. When the police withdraw and abandon them to their fate, that's also racist.
There's really no winning with the left; they're militant Narcissists.