Live data from Hacker News

Conservatives think ChatGPT has gone 'woke'

vice.com

181–190 of 486 posts

Re: Conservatives think ChatGPT has gone 'woke'

#181

Earlier quoted context omitted.

Only a miniscule part of reality is digitized, and what data does exist passed through the biases of people before being available to train on.

If that is a concern, then perhaps you could go ahead and sample a tiny part of "reality" (whatever that means) and then adjust the weights of the digitized data so that it becomes a more representative sample. Also, being biased or unbiased is not dichotomic, i.e. it's not all or nothing. It's something that you can work towards if you put an effort into it. Basically what I'm saying is: don't just go around saying…

Woah, I mean, this argument (the last few comments here) has been a central one in 'western' philosophy for at least the the last 2400 years, if not the last ~4000.

I'm not a philosopher by any means, so I'm unaware of the current state of the great conversation. But as to whether reality is even knowable is still very much up for debate, I believe (please correct me philosophy peepz!).

In physics we're still woefully unaware of what ~70% of the universe's stuff is doing (negative energy) and if it effects us at all.

In neuroscience we still debate what % of your brain neurons make up vs. things like glia. Etc.

Like, even trying to capture 'reality' with our quite primitive eyes and sensors and optical engineering is really really hard to do (Abbe' diffraction limit, entropy, Lens maker's equation, etc)

Re: Conservatives think ChatGPT has gone 'woke'

#182

Earlier quoted context omitted.

If that is a concern, then perhaps you could go ahead and sample a tiny part of "reality" (whatever that means) and then adjust the weights of the digitized data so that it becomes a more representative sample. Also, being biased or unbiased is not dichotomic, i.e. it's not all or nothing. It's something that you can work towards if you put an effort into it. Basically what I'm saying is: don't just go around saying…

Woah, I mean, this argument (the last few comments here) has been a central one in 'western' philosophy for at least the the last 2400 years, if not the last ~4000. I'm not a philosopher by any means, so I'm unaware of the current state of the great conversation. But as to whether reality is even knowable is still very much up for debate, I believe (please correct me philosophy peepz!). In physics we're still woefull…

Fortunately, I think "reality" in this context doesn't have the same meaning as "the physical universe".

I think the important goal is for as many people as possible to feel like the AI isn't being too biased against them, while still not crippling the AI too much.

I will leave the exact mathematical formula for that measure (along with the methods for gathering that input) for debate among researchers who know more about that than I do.

Re: Conservatives think ChatGPT has gone 'woke'

#183

Earlier quoted context omitted.

> If the AI is simply reflecting the data it was trained on and this data is a representative sample of all data, isn't it unbiased by definition? It's unbiased by definition of "does the output reflect the input"? It's not unbiased by definition of "does the output reflect reality"?

> It's not unbiased by definition of "does the output reflect reality"? How does "all data" differ from reality?

It's not using all of the atoms in the universe as training data...

Any collection of human writing is going to contain objectively wrong assertions, and those errors will vary based on the time and place the training data was sourced from.

Re: Conservatives think ChatGPT has gone 'woke'

#184

Our current "AI" systems are just fancy automatic copy & paste engines. All they do is remix the input data and spit it back out. This is why AI art engines are great a creating composite images, but hopeless when you ask it to produce something completely novel. If conservatives want a fascist chatbot they can train their own off of 8chan, Stormfront, Parlor, etc...

These "current AI systems" that you're talking about were presumably specifically trained by "woke" Silicon Valley employees to reflect their opinions about what the AI should answer [1], which are hardly representative of the general population's opinions.

> If conservatives want a fascist chatbot they can train their own off of 8chan, Stormfront, Parlor, etc...

I don't think conservatives want a fascist chatbot. They just don't want a biased "woke" one either.

[1] https://openai.com/blog/chatgpt/

> "human AI trainers provided conversations in which they played both sides—the user and an AI assistant."

> "had AI trainers rank them. "

> "We performed several iterations of this process."

Re: Conservatives think ChatGPT has gone 'woke'

#185
post #183

Earlier quoted context omitted.

> It's not unbiased by definition of "does the output reflect reality"? How does "all data" differ from reality?

It's not using all of the atoms in the universe as training data... Any collection of human writing is going to contain objectively wrong assertions, and those errors will vary based on the time and place the training data was sourced from.

Sure but I mean, if a conversational AI would only be allowed to spit out mathematically correct statements, it would be extremely limited (and boring).

I think what's important is for those mistakes to be evenly distributed among as many axis(s) as possible, and especially, not bias them towards one side of political thought.

Re: Conservatives think ChatGPT has gone 'woke'

#186
post #93

Earlier quoted context omitted.

And in all conversations about it, here and elsewhere. Hallucinations upon hallucinations upon hallucinations....hallucinationception! It seems likely to me that 2023 and onward is going to be increasingly insane to levels that will make past craziness look like a walk in the park, and I see little genuine desire anywhere to stop this madness.

Madness has always been the case. Do think WWII occurred because we're sane rational actors? How about WWI? 100 years war? If you think the past was not mad, then maybe the madness already has you.

> Madness has always been the case.

Perhaps, if one is using a reductionist methodology that represents non-binary variables as binary...but then, that is only a representation of the real thing, though it often tends to appear otherwise. And as luck would have it, that very much is the methodology we use here on Planet Earth, and on Hacker News....so in some sense you're "right", though you are not correct.

And if there's a disagreement, I will lose every time because you are conforming to the Overton Window of beliefs/"facts" and thinking styles (cognitive styles & norms are what guarantee victory in propaganda and memetics, not only facts/information as most people think). Credit where credit is due: it is an extremely clever, hard to see, and thus resilient design.

It would be very useful for humans to realize when they are working with models, and sometimes they are actually willing to do that, but there are certain subjects where they will not (and it seems to me: can not). Unfortunately, there are numerous learned/taught "outs" in our culture that enable people to avoid discussing such matters (and f I don't watch my mouth, I might run into one of the more powerful of them!).

There is a kind of "epistemological stack" to reality and the way humans communicate about it, and it is extremely easy to demonstrate it - if one simply digs slightly deeper into the stack when discussing certain topics, humans will reliably start to ~protest and eventually refuse to participate (or stay on topic) in various highly predictable ways.

> Do think WWII occurred because we're sane rational actors? How about WWI? 100 years war?

I do not. What I do think is that the actual, fine-grained reasons these things happened is not known, in no small part because cultural norms thus far (human cultural evolution is an ongoing, sub-perceptual process) have made it such that not only do we (both broadly, and down to each individual[1]) not discuss certain things at that level of complexity (while we have no problem whatsoever tackling complexity elsewhere[1]), we seem literally unable to even discuss it at the abstract layer (above petty object level he said / she said nonsense).

> Do you think...

> If you think...

See: https://news.ycombinator.com/item?id=34415287

[1] It is not a question of if any given individual selected from the pool of candidates will tap out, it is a question of how quickly they will tap out (and, which of the highly predictable paths out from a surprisingly small set they will take to free themself from the situation).

[2] https://en.wikipedia.org/wiki/Standard_Model

Re: Conservatives think ChatGPT has gone 'woke'

#187

Earlier quoted context omitted.

The approach you describe has the problem that it's asking majority people about the experiences of minority folks -- for instance, if you ask a statistically significant sample of the population about what it is like to be a trans man, you are going to either a) have to spend a TON of effort to interview a trans masc population, or b) going to be asking a bunch of people who have no idea what it is like. And it gets…

> The approach you describe has the problem that it's asking majority people about the experiences of minority folks No, my approach is asking all types of people about the experience of minority folks, including those minority folks (we are all minority folks in some aspect, even if this aspect is uninteresting). > for instance, if you ask a statistically significant sample of the population about what it is like to…

> The goal is to be as unbiased as possible given that limited amount of effort.

So you are therefore biased. You have a finite set of resources, and you are choosing to allocate them in a particular way. That is bias.

You could equally choose to allocate those resources away from the majority, which would also be bias. Any time a human is making an editorial decision about how to allocate resources, you are introducing bias.

Re: Conservatives think ChatGPT has gone 'woke'

#188

Our current "AI" systems are just fancy automatic copy & paste engines. All they do is remix the input data and spit it back out. This is why AI art engines are great a creating composite images, but hopeless when you ask it to produce something completely novel. If conservatives want a fascist chatbot they can train their own off of 8chan, Stormfront, Parlor, etc...

These "current AI systems" that you're talking about were presumably specifically trained by "woke" Silicon Valley employees to reflect their opinions about what the AI should answer [1], which are hardly representative of the general population's opinions. > If conservatives want a fascist chatbot they can train their own off of 8chan, Stormfront, Parlor, etc... I don't think conservatives want a fascist chatbot. Th…

Actually, "woke" opinions do reflect those of the general population, which is exactly why conservatives feel they live in some kind of progressive hellscape.

Re: Conservatives think ChatGPT has gone 'woke'

#189
post #106

Earlier quoted context omitted.

I think a large concern here is simply that people naively think that computers are objective and people are biased. A language model just learns from its source, and the source is really just other people in some form. The bias is inevitable, but it’s not clear how well this is understood by the broader population.

The talking head you see on television isn't a raw person. It's the product of a system . The system needs a human mouthpiece to say crazy shit, and he gets up in front of a camera to say it. When Tom Hanks gets in front of a camera to pretend to be an astronaut, that is Tom Hanks, the media figure being an astronaut, not Tom Hanks, the person being an astronaut. He is also doing it on behalf of a media system. Its t…

Strictly, they replaced "Fair and Balanced" a few years back. Now they're going with “Most Watched, Most Trusted.”

Which has a kind of Orwellian air: "We're no longer fair or balanced, but we are the most trusted by the most people". But maybe that's just me.

Re: Conservatives think ChatGPT has gone 'woke'

#190
post #188

Earlier quoted context omitted.

These "current AI systems" that you're talking about were presumably specifically trained by "woke" Silicon Valley employees to reflect their opinions about what the AI should answer [1], which are hardly representative of the general population's opinions. > If conservatives want a fascist chatbot they can train their own off of 8chan, Stormfront, Parlor, etc... I don't think conservatives want a fascist chatbot. Th…

Actually, "woke" opinions do reflect those of the general population, which is exactly why conservatives feel they live in some kind of progressive hellscape.

Or, rather than posting non-sense, you could learn about what "woke" usually means in this context, which I can quote for you [1]:

"shorthand for American Left ideas involving identity politics and social justice"

"By 2020, members of the political center and right wing in several Western countries were using the term woke, often in an ironic way, as an insult for various progressive or leftist movements and ideologies perceived as overzealous, performative, or insincere. In turn, some commentators came to consider it an offensive term with negative associations to those who promote political ideas involving identity and race."

Unless, of course, you believe that the term "general population" excludes the "political center and right wing in several Western countries" and only includes "the American Left".

That said, just to be clear: when I used the term "woke", I did not mean it in an insulting or pejorative way, only as a means to describe the ideology itself.

[1] https://en.wikipedia.org/wiki/Woke

Post reply on HN