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

A failed AI girlfriend product, and my lessons

mazzzystar.github.io

Post reply on HN
Print view View original

81–90 of 400 posts

Re: A failed AI girlfriend product, and my lessons

#81

Post by pizza234 »

lynx23 wrote: post #77
krzat wrote: post #68

Earlier quoted context omitted.

The fact that you point this out, is a proof that humanity can improve, what needs to change is our culture, especially education, ofc it may take a lot of time but it will happen eventually.

Regarding long-term rewards and global warming: What is the actual long term reward here? As I see it, there is no way to collect any sort of reward for helping to prevent global warming by changing your habits. Before the problem gets really problematic, most of us will be dead and burried. I fail to see how "caring for the future generation" has any kind of reward attached to it. Its an act of kindness, but there a…

> Before the problem gets really problematic, most of us will be dead and burried.

The situation lies on a spectrum between "not problematic" and "really problematic".

There is actually plenty of evidence that tangible impact already started, independently of the perception of those whoe live in areas where the effect is neither perceived nor obvious.

Re: A failed AI girlfriend product, and my lessons

#82

Post by Kye »

feintruled wrote: post #27
globular-toast wrote: post #20

Earlier quoted context omitted.

> Also it seems like he had full access to the replies via the voice generator, which sure isn't great from a privacy standpoint. Whose privacy is being violated? Dolores's?

You joke, but I can forsee issues. If he has the replies to a particular user, and he has a means of IDing that user, he could cause them a lot of hurt. They are getting intensely sexual replies back, they aren't getting that without supplying sexual prompting. Could also be used to out gay users.

This was mostly just amusing, but it's the kind of thing that can happen and leak people's deepest secrets: https://news.ycombinator.com/item?id=38191470

All it would take is one of those people genuinely putting in a query without noticing what's happened to put in some personal info.

Re: A failed AI girlfriend product, and my lessons

#83

Post by krzat »

lynx23 wrote: post #79
krzat wrote: post #78

Earlier quoted context omitted.

Probably a question to the guy above, but there is immediate psychological reward in any act of kindness or generosity (at least in healthy, thriving individuals).

However, OP argued that global warming is suffering from people NOT getting immediate rewards for their actions. Your argument basically says OP is wrong, and there is no need to counteract peoples tendency to prefer short-term rewards over long-term because they are supposed to get enough incentive by just knowing they have been kind. I doubt that.

What I think OP said, is that humans won't adapt to these challenges due to our biological limitations.

IMO humanity will adapt, and the first sign of this is that those problems are getting recognized and given more and more traction.

Re: A failed AI girlfriend product, and my lessons

#84

Post by luke-stanley »

feintruled wrote: post #7

Well this went places. The dev really 'ate his own dogfood' so to speak. I thought it was going to be a despairing look at how his companion AI was continuously turned to sleazy uses by a pervy userbase but that's pretty much what he was going for it seems. I have to say, I hated this part: "...gained a lot of Chinese users. I immediately realized the political risk: users could set Dolores’ background, personality,…

Their name is "Ke Fang", and they say Chinese is their first language...

Re: A failed AI girlfriend product, and my lessons

#85

Post by suoduandao3 »

cornholio wrote: post #24

> GPT-based products, if not priced per usage, would fall into a dilemma: 1% of users consume 99% of tokens. A user from Sweden (seen from Cloudflare’s call volume) chatted with Dolores for 12 hours straight This, my friend, is a captive customer, that will pay anything to get his girlfriend back. I cringe at the potential of unethical behavior and abuse, where people fall in love with virtual entities fully owned by…

"I'd like to correct the record on this story. We do NOT sell kidnapping roleplay packages. The user paid for a premium Daring Rescue Package. These do not become available unless the user spends two consecutive months as a net cost to our company. The user was given multiple reminders that he might want to cut back on usage or purchase a premium plan. Even then, he could have opted for a free Daring Rescue package and taken his chances with the base 1% chance of untraumatized recovery this package offers. Shuffling user responses under these high-pressure conditions is vital to improving our training data, and helps our users learn to deal with loss of important relationships. We are a relationship training service, not a replacement for them, and this user wanted the equivalent of a college degree.".

On a more serious note, anyone know why unethical business plans are so much more fun to write? I always find myself giggling when these ideas come up.

Re: A failed AI girlfriend product, and my lessons

#86

Post by dontupvoteme »

stef25 wrote: post #80
bradley13 wrote: post #22

Interesting read. Others have also created virtual friends using ChatGPT. What I find unsavory, is OpenAI imposing content restrictions like NSFW. That can easily extend to censorship of non-PC content, unpopular political ideas, and more. The solution will (I hope) ultimately be open-source LLMs that you can run locally, on your own hardware.

If you're able to let your imagination run wild you'll see that the depth of depravity of NSFW AI content generated by user input probably has no end.

Even worse, some people turn it into a competition.

Leak an API key on github or somewhere else and heaven help you if you get picked up by their scraping.

Re: A failed AI girlfriend product, and my lessons

#87

Post by dontupvoteme »

phatskat wrote: post #64
bradley13 wrote: post #22

Interesting read. Others have also created virtual friends using ChatGPT. What I find unsavory, is OpenAI imposing content restrictions like NSFW. That can easily extend to censorship of non-PC content, unpopular political ideas, and more. The solution will (I hope) ultimately be open-source LLMs that you can run locally, on your own hardware.

And this doesn’t just cover sexually NSFW content - anything trending towards topics like violence and suicide is also off the table for most AI backends from what I understand. I struggle to see the benefit of this kind of censorship as well, it has a very “video games make people murders” vibe. That we’re being protected, for our own good, from…horny robots?

It's largely to protect their brand from bad press (this has hit even Google/Microsoft in the past), but established finance institutions absolutely hate porn, even the "less shady" ones like onlyfans or pornhub have issues with banking.

Re: A failed AI girlfriend product, and my lessons

#88

Post by spacemadness »

At this point in the hype cycle, I am completely unimpressed by everyone slapping something on top of GPT and calling it a product. There are at least 4 projects on the front page that do this today. The interesting technology is GPT itself, but these projects just feels like people trying to scoop up attention for themselves using the psychology of the hype cycle. Please HN, enlighten me on why the middle-man approach here is so interesting.

Re: A failed AI girlfriend product, and my lessons

#89

Post by anthony100 »

FeepingCreature wrote: post #23
Traubenfuchs wrote: post #5

So, all in all, OpenAIs puritanism ruined this product, made the developer sad and took away a virtual porn toy from lonely, horny men. Who were they doing this for? Who is the winner? They lost a customer and made a lot of people sad.

> Who were they doing this for? VISA probably. Card processing is weirdly puritanical.

Interesting. Not sure how this is different from Onlyfans and lots of money is going there.

Re: A failed AI girlfriend product, and my lessons

#90

Post by red-iron-pine »

suoduandao3 wrote: post #85
cornholio wrote: post #24

> GPT-based products, if not priced per usage, would fall into a dilemma: 1% of users consume 99% of tokens. A user from Sweden (seen from Cloudflare’s call volume) chatted with Dolores for 12 hours straight This, my friend, is a captive customer, that will pay anything to get his girlfriend back. I cringe at the potential of unethical behavior and abuse, where people fall in love with virtual entities fully owned by…

"I'd like to correct the record on this story. We do NOT sell kidnapping roleplay packages. The user paid for a premium Daring Rescue Package. These do not become available unless the user spends two consecutive months as a net cost to our company. The user was given multiple reminders that he might want to cut back on usage or purchase a premium plan. Even then, he could have opted for a free Daring Rescue package a…

ethical ones are fairly straightforward, and will often be the same.

unethical ones can range from not-so-nice to mind-bogglingly evil, and in all sorts of "fun", different, unique ways.

"Happy families are all alike; every unhappy family is unhappy in its own way."

Post reply on HN