Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

AI Generated Seinfeld banned on Twitch for transphobic jokes

sportskeeda.com

Post reply on HN
Print view View original

441–450 of 713 posts

Re: AI Generated Seinfeld banned on Twitch for transphobic jokes

#441

Post by daveslash »

etchalon wrote: post #406
daveslash wrote: post #390

Earlier quoted context omitted.

If I understood temp_account_32's point correctly, the comment about the mental health was not " they're trans, therefore their brains _must_ be wired wrong ", so much as it was "before they transition, they're often so distraught over their identity and who they see in the mirror, that it often causes despair and depression" . I agree with you that research has pretty firmly come down on the side of "nah, it's socie…

It's not all society's fault, but research has shown that the distress and depression of being outside societal defaults is greatly lessened when support is available and acceptance wide spread.

Yes. Completely agree. To suggest that it's all society's fault is disingenuous. To suggest that society is fault-free is reckless and cruel, and society's part is not a small piece of the pie either. Society needs to step up to address the societal aspect of the issue as part of a multi-pronged approach.

Re: AI Generated Seinfeld banned on Twitch for transphobic jokes

#442

Post by KMag »

hourago wrote: post #213
cobertos wrote: post #31

Earlier quoted context omitted.

A seemingly innocuous engineering change led to some drastic consequences. Imagine the other more critical production applications who have made the same reasonable engineering decision.

> A seemingly innocuous engineering change led to some drastic consequences. If you think that this is "drastic" you should read "Case Study 4: The $440 Million Software Error at Knight Capital". https://www.henricodolfing.com/2019/06/project-failure-case-... tldr; At Knight, some new trading software contained a flaw that became apparent only after the software was activated when the New York Stock Exchange (NYSE) o…

I was developing trading systems at Goldman then. I remember that day. I got a call that evening to come in early the next day to discuss additional preventions and mitigations to implement.

Scarier was the day I put out a change request shortly before the end of the day in Asia, went out to dinner and drinks with some colleagues, stopped by my office on the way home, and merged the approved change request, during early morning trading in the US. The next day (in Asia) I woke up to discover Goldman had a roughly 28 million dollar trading loss. I spent a couple of minutes proving to myself that my change couldn't possibly have been the cause, and then realized I would have been woken in the middle of the night by a phone call if there was any possibility that my change had caused the trading loss. Since then, I don't merge code at the end of the day, and try to avoid Fridays. I was already in the habit of not coding if I had consumed any alcohol, but added the habit of not merging if I've consumed any alcohol. There's nothing like a 28 million dollar panic to get your practices in shape.

Re: AI Generated Seinfeld banned on Twitch for transphobic jokes

#443

Post by ben_w »

welshwelsh wrote: post #424
ben_w wrote: post #372

Earlier quoted context omitted.

Being cancelled is the smallest possible increment I am aware of over being heckled. It is the necessary other side of the same coin as freedom of speech, otherwise you end up with the literal Republican party being legally obligated to let the literal Communist Party of the USA (not a euphemism for the Democrats) lecture them during their Primaries.

>otherwise you end up with the literal Republican party being literally obligated to let the literal Communist Party of the USA lecture them during their primaries And this is a bad thing how? The purpose of free speech is to prevent groupthink. Nothing good comes from allowing Republicans and Communists to create echo chambers where people only hear opinions they already agree with and dissent is banned. Wouldn't it…

> And this is a bad thing how?

While I think it would be hilarious in a comedy sketch, in practice it's a DDOS attack. Overwhelm the discussion with irrelevant noise and you have de-facto suppressed someone else's speech.

Re: AI Generated Seinfeld banned on Twitch for transphobic jokes

#444

Post by ekleraki »

adamrezich wrote: post #400
PartiallyTyped wrote: post #320

Earlier quoted context omitted.

JAQing off and asking questions to learn and understand are two truly different things. Misrepresenting positions and arguments to a rabid “fanbase” is not asking questions. It’s time the charades stop and the fascists drop the pretences, oh wait their supreme leader has PUBLICLY said he will ensure that trans people will be outlawed. No amount of science and evidence is ever enough for these people either. Transsexu…

you might not be aware of this but for many people, myself included, seeing casual use of the word "fascists" in discussions such as these elicits an immediate, uncontrollable eye-roll. presently the word "fascists" appears three times on this page (outside of this post), and like, I know what you're getting at, because I'm familiar with the general ideology you appear to subscribe to (based on your use of that term…

The word "fascist" was used to refer to a person who supports authoritarian regimes that attack minorities, education, reproductive rights, freedom of speech whenever they do not like said speech. I used it to refer to people in support of creating lists of "unwanted" people [1], I used it to refer to people in support of banning books that stand against their ideology.

In your overly condescending and infantilising response, you missed context from parent comment:

> there are actual Nazis out there who want to put people in ovens, so it's taboo to make jokes about how they may have a point.

> similar thing with trans people, you mention them and trolls come out of the woodwork saying they should kill themselves.

Now, you may be okay with a small segment of the population dying at the hands of definitely-not-fascists who have attempted to murder certain officials at their homes, but I do not subscribe to your morally deprived and bankrupt view of the world.

Anyway, what's the difference between Putin's regime, Iran, and GOP besides the religion? It's quite eerie to see how many parallels one may draw from GOP, Iran and Russia.

[1] https://www.huffpost.com/entry/texas-transgender-database-dr...

Re: AI Generated Seinfeld banned on Twitch for transphobic jokes

#445

Post by smcl »

fullshark wrote: post #294
bsaul wrote: post #84

As many people here noticed, the bit has actually nothing to do with transphobia, or gay, but it's about a comedian bombing a show with horrible jokes. What's interesting in this case, is that there can be no "background check". Or "this guy has actually had weird quotes in the past, he's very suspicious". Or "he has republican friends"... All the regular strategy activist use to demolish a public figure over a prete…

This story has an even better result than claiming a scalp for these guys, now the activists get to shout from the rooftops about how important it is these companies hire them to develop "fair and unbiased" AI models. This is a way for them to acquire greater control on information, as they can control the generative models that produce information.

People will, rightly, use this to urge caution against naively throwing AI at a problem. The non-specific activists demanding to be hired in some non-specific roles is a thing you have invented to get mad at.

Re: AI Generated Seinfeld banned on Twitch for transphobic jokes

#446

Post by willis936 »

yazzku wrote: post #375
mFixman wrote: post #121

Everybody seems up in arms about this situation, but I find the whole thing hilarious. After half a decade of AI hype without many uses in the wild, we finally live at a point in history where "AI-generated Jerry Seinfeld makes transphobic joke, gets cancelled" is not an Onion headline.

Somebody should try this again but with George Carlin.

The earth doesn’t share our prejudice toward AI. AI came out of the earth. The earth probably sees AI as just another one of its children. Could be the only reason the earth allowed us to be spawned from it in the first place. It wanted AI for itself. Didn’t know how to make it. Needed us. Could be the answer to our age-old egocentric philosophical question, “Why are we here?" "AI".

Re: AI Generated Seinfeld banned on Twitch for transphobic jokes

#447

Post by robertlagrant »

sassyonsunday wrote: post #407
Levitz wrote: post #303

Earlier quoted context omitted.

I think the point being made is that jokes don't imply hatred per se. If I say "Spanish people are lazy fucks" that sounds offensive and might be considered racist, but I myself am Spanish and that context changes its meaning. We generally strive away from racist jokes because as a society we are concerned that we are furthering racist ideologies otherwise. For a more explicit example, if a person says "I'm tired of…

I'm transgender. I transitioned 17 years ago. Sometimes I call myself a t-word or tranma (because I transitioned a long time ago I'm only in my early 30's so I guess I have a few years before I'm a literal granny :'[). Should I get canceled because I said the t-word in a gay bar with my besties or when I joke around with my husband and say things like "this is what you get for marrying a t-word"? No. But it's also us…

I've never heard tranma before and it's fantastic!

Thanks for articulating this; it's helpful. My only question is: words that are labels don't create hatred, although they might signify it. Will stopping words being used do anything to reduce any hate people feel and/or increase understanding, and conversely will allowing it add any hate to people and/or decrease understanding?

Re: AI Generated Seinfeld banned on Twitch for transphobic jokes

#448

Post by freejazz »

marcusverus wrote: post #288
mjamesaustin wrote: post #219

Earlier quoted context omitted.

It's weird to talk about missing an element of human empathy regarding the topic of trans individuals, whose very existence is unacceptable to a broad swath of the population. 40% of transgender people have attempted suicide. The public dialogue around trans lives is one of very little empathy, respect, or civility. People are on this thread casually debating whether this joke "went to far" or not, while the people w…

> Would it have been OK to make a joke about 9/11 victims while they were still digging people out of the wreckage? This is a good question! Who should be asking it? Comedians, or the morality police?

Okay, I'm calling the drama police on you.

Re: AI Generated Seinfeld banned on Twitch for transphobic jokes

#449

Post by skim_milk »

cobertos wrote: post #31
gregsadetsky wrote: post #19

Explanation from their Discord (more info/links below) “Earlier tonight, we started having an outage using OpenAI’s GPT-3 Davinci model, which caused the show to exhibit errant behaviors (you may have seen empty rooms cycling through). OpenAI has a less sophisticated model, Curie, that was the predecessor to Davinci. When davinci started failing, we switched over to Curie to try to keep the show running without any d…

A seemingly innocuous engineering change led to some drastic consequences. Imagine the other more critical production applications who have made the same reasonable engineering decision.

They've been developing this on-the-fly since it gained popularity (and profitability). Basically, they ran fast and broke things, like any good startup.

I'm nearly 100% certain that the scripts are generated through prompt engineering, with a random prompt (e.g. tell a joke prompt, talk about a new restaurant prompt) being selected for the scene.

From what I can gather they first used the older, cheaper GPT-3 models, only upgrading to davinci-003 when it was profitable. The older GPT-3 models proved fine and didn't generate edgy content for the several months they were up and running.

But I think the change that broke the camels back was they added a "2006 Laugh Factory incident with edgy content" prompt and only tested it on the davinci-003 model - the new models having been wiped clean of antisocial training data, while the older smaller models still having contentious content encoded in the model.

So, davinci-003 did fine producing "politically aligned" text with the "edgy" prompt because it's "cleaned", but when the openai API for davinci went down the fallback was curie. The older "unclean" curie model combined with an edgy prompt inevitably caused what we saw here.

Re: AI Generated Seinfeld banned on Twitch for transphobic jokes

#450

Post by jimmaswell »

btown wrote: post #227
pirate787 wrote: post #112

Earlier quoted context omitted.

Artists are what we call the people who push against the boundaries of majority group-speak and government mandated thinking.

Intent matters, and there are a lot of people who would say the exact words that the AI said, not with the intention to provoke thought but with the intention to normalize hate. A temporary ban while there is ambiguity about the intent is not unreasonable. But Twitch should be criticized if it persists in the ban after it discovers, as it should have by now, that it was not the intention of this channel's creators, a…

> say the exact words that the AI said, not with the intention to provoke thought but with the intention to normalize hate

The "but nobody is laughing" part would run counter to this. The premise of the whole joke is that the referenced jokes aren't funny.

Post reply on HN