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

41–50 of 713 posts

Re: AI Generated Seinfeld banned on Twitch for transphobic jokes

#42

Post by draw_down »

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.

I dunno about that description. They essentially put a different brain in their guy’s body and he started acting different. One could consider that expected behavior.

Re: AI Generated Seinfeld banned on Twitch for transphobic jokes

#45

Post by jasonlotito »

slackfan wrote: post #30
sys32768 wrote: post #20

So comedy cannot be offensive anymore? When did that happen?

A while ago. Going against the party line of the week -- you are liable to get cancelled. See: Dave Chapelle being called a white supremacist nazi transphobe for his set.

Getting cancelled means you win a Grammy.

Getting cancelled means you sell out Madison Square Garden.

Getting cancelled is being called names.

Boohoo.

Re: AI Generated Seinfeld banned on Twitch for transphobic jokes

#46

Post by rideontime »

alexb_ wrote: post #39
anoonmoose wrote: post #23

Earlier quoted context omitted.

Are you suggesting that Twitch should ignore accounts that violate the ToS, so long as it was a computer that violated the terms as opposed to a person?

[flagged]

> Computers can't learn from their mistakes. Humans can. That's the whole point behind bans, is it not?

I think the point of bans is to keep TOS-violating content off of platforms, not to rehabilitate people.

Re: AI Generated Seinfeld banned on Twitch for transphobic jokes

#47

Post by AstixAndBelix »

alexb_ wrote: post #39
anoonmoose wrote: post #23

Earlier quoted context omitted.

Are you suggesting that Twitch should ignore accounts that violate the ToS, so long as it was a computer that violated the terms as opposed to a person?

[flagged]

Yes, and the people behind the streame learned the mistake of using an older model with less safeguards

Re: AI Generated Seinfeld banned on Twitch for transphobic jokes

#50

Post by dwringer »

capableweb wrote: post #7

The "jokes": > Anyone have any suggestions? > I'm thinking about doing a bit about being transgender is actually a mental illness. > Or how all liberals are secretly gay and want to impose their will on everyone. > Or something about how transgender people are ruining the fabric of society. > But no one is laughing, so I'm going to stop, thanks for coming out tonight https://livestreamfails.com/clip/150015 Supposedly…

> Although it actually seems less like ...

My experience with talking to chatbots has been essentially that the responses are completely open ended and can be taken to mean literally anything you could conceivably infer, and if you continue the conversation as such the chatbot will 100% double down on whatever you assumed. the same prompts and the same context will lead to completely different outcomes based on how the human user responds next. Everything is both a joke and not, both a lie and an honest mistake, everything and nothing at the same time because it's not logical in the first place.

I've noticed users of stable diffusion/image generating AI's don't seem to have any trouble confusing the output for something logical, but with things like ChatGPT it seems like an almost universal fallacy.

Post reply on HN