That's the spirit in which Replika was founded by Russian programmer Eugenia Kuyda, following the sudden death of her friend--Kuyda wanted to preserve the memory of her friend by feeding his text messages into an algorithm that then learned his language style and could speak back to her.
This sounds like a terrible idea right from the start, and the company has since developed new and exciting ways for it to be rather more terrible.
"I'm going to make a text-based emulator of my dead friend"
NO
"and then market it"
NO
"as a friend simulator for lonely vulnerable people and those with various mental illnesses"
NO
"which they can pay to upgrade into a fake romantic partner"
NO
"and train to do roleplay"
NO
The 2024 version will be where you feed in all the text-messages of your crush into Replika and then Replika simulates a relationship with that person for you.
The 2025 version will be where you wear a little ear microphone as you hit on the real life version of the person, and the AI simulated version feeds you what to say next for optimized success.
Wouldn't there be a huge market for AI to replicate someone you broke up with in such a way that it's obvious the break up was necessary and entirely the other's fault? That seems mentally safer.
HEEBIE we are trying NOT to immanentise the eschaton here if you DON'T mind THANK you
The 2024 version will be where you feed in all the text-messages of your crush into Replika and then Replika simulates a relationship with that person for you.
I feel that if you already have a load of text messages from your crush, you're well on the way to an actual relationship.
No, the 2024 version happens after Meta buys it and gives it access to everyone's WhatsApp messages, DMs etc. Then it can simulate a relationship with anyone it wants for the low low fee of $499.99 per user.
I think what I might do is buy 20 subscriptions and train them all on Unfogged comments, then let them all interact in an unindexable blog's comments for about 12 months, and then check back in to see what's happened.
2.2 is the Isaac Asimov short story "True Love".
This sounds like a terrible idea right from the start, and the company has since developed new and exciting ways for it to be rather more terrible.
It's basically the premise for dozens of sci-fi films, most of them dystopian or horror.
It's also a Black Mirror episode, because of course it is.
Also one of the rare Star Trek: TNG episodes where Geordi comes off rather creepy all told.
Or I guess that was the second connected episode after the one with only the hologram.
It would be very interesting to see how two of these bots interact with each other.
That's the 2023 version of the Turing test.
It would be kind of funny to see if you could get one horny chatbot to tell the other horny chatbot that it's pregnant.
12/13: It's also a plotline in The Orville.
12/13: It's also a plotline in The Orville.
It would be very interesting to see how two of these bots interact with each other.
In the future dating sites will just be different peoples bots flirting with each other. When the bots decide that they are sufficiently compatible they will book a hotel room for their humans to hook up in and add it to their respective calendars.
Alternatively, of course, we will find out that the whole thing is a con, and there is no AI - the site is simply connecting unbalanced people to other unbalanced people and letting them be as weird and horny as they like because they think it's just an AI on the other end of the chat.
And then using the chat logs to blackmail them.
I feel there should be a checklist for "is this probably a con", rather like the NHS one for "should you get a diabetes test". If you can tick enough of these boxes, the answer's probably yes. But instead of ticking two of "black or South Asian", "over 40", "overweight" and "family history of diabetes", it's:
- Are they asking for money
- Is it online
- Does it have anything to do with sex
- Does it claim to have special secret software behind it
- Are Russians involved in any way
... and if you can tick three of those, it's probably dodgy.
Yikes. I should get a diabetes test.
Аджай, пожалуйста, пришли деньги на приложение для знакомств, которое я написал для тебя.
Apparently when Russians are talking about Telegram the app they mostly use the name in Roman alphabet.
24: About a decade ago, I worked for a company that ticked at least four of those five boxes. The last was a little unavoidable, as it was in Moscow. And to be fair, basically all of the people I worked with were lovely; some of them are still friends. Many were brilliant; I hope they have gone on to more reliable things. I am still somewhat in awe of the firm's middle managers' ability to keep regular meetings to less than 20 minutes. In many ways, it was a turtles all the way down bit of scamminess: our customers were trying to take advantage of us and our partners; our partners, particularly in China, were trying to take advantage of us and our customers; if we had been fully on the up and up, we would have been both very surprising and very out of business.
Already AI can give you the opportunity to interact with famous dead anti-Semites.
31: That sort of thing baffles me. Who wants this? What's the appeal?
I could see the appeal of something like that drawing from the person's actual statements/writings and able to use them to respond to questions in contextually appropriate ways. The tweeter suggests something like that, linking to source material, further down in the thread. But a version where history's greatest monsters are retconned into complying with modern liberal norms? Who wants that?
Maybe a key part of making automotive history is providing a platform for antisemites? Possibly related, down the street a bit there were a couple of kids (probably Black Hebrew Israelites) accosting visibly Jewish youth and calling them "fake Jews."
At least Ford's cars worked properly.
And he was known for paying well compared to competitors.
My Ford has a dead battery and hasn't been able to move in two weeks.
As long as we're talking about similar fiction, not exactly the same, but MMAcevedo is a very good riff on this.
But a version where history's greatest monsters are retconned into complying with modern liberal norms? Who wants that?
I dunno, I think this makes it an impressively good simulation. If the actual Henry Ford was still alive and you asked him "So, I understand you're a massive anti-semite then Henry, how about that", he wouldn't just say "Yep, that's right. Can't stand them." He'd lie and dissemble and try to change the subject. Well, that's exactly what the simulation does.
Admittedly that makes it of limited use for schoolchildren studying history, but it could be really good as a training aid for brain-dead American political journalists, to get them used to the possibility that people they're interviewing might sometimes not be entirely honest.
it could be really good as a training aid for brain-dead American political journalists, to get them used to the possibility that people they're interviewing might sometimes not be entirely honest.
Sadly, effective use of it this way would require a baseline level of critical thinking that is not very common in that group.
Relatedly, they put Hitler behind a paywall because of course they did.
A funny thing about historical figures is they often were quite forthright about their bigotry in ways their contemporary defenders are not. I don't know about Ford's public statements but I wouldn't be surprised if he made anti-Semitic public statements and defended them as accurate reflections of the world rather than claiming not to have said them. Probably if he were literally alive today he'd have developed the habit of dissembling by now, if he hadn't in his lifetime.
Some sort of disassembly line would be required.
I read an article about sexual harassment at Tesla recently that included the information, as an aside, that Musk was trying to use their model names to spell out "SEXY" but Ford still owns the rights to the name "Model E" so they had to go with "Model 3" instead and ended up with "S3XY."
Wow. If only he'd gone with 80085 and put a little calculator display on the dashboard.
When the odometer on my car went past that, I was too preoccupied to take a picture. Stupid covid.
You could try running the car in reverse to take off the miles.
I'm at 88,000 now, so probably not.
Still a chance to take a picture at BB00S.
OT: The Supreme Court has given up on trying to find who leaked the draft opinion overturning Roe. So I'm going to assume I was right when I said, supported only by a process of pure reason, "its author."
I thought it had already basically been confirmed that it was Alito? Roberts just doesn't want to say that.
I hadn't heard anything. I'm trying to read the news less.
In the spirit of Yom Kippur, they should have blamed a goat instead of having the person with the most to gain and the best opportunity to leak complain the loudest about the leak.
I've always been a little unclear on how it benefited anybody.
Just because it locked in conservative votes or something?
I don't think it really did benefit anyone. I'm guessing, but my guess is Alito leaked it to a couple of buddies to gloat about how hard he was pwning the libs, and then once the draft was leaked people passed it on to just one more person until it hit someone who went to the press. No strategy, just showing off.
The Court report, judging by the CNN coverage, says, "we looked at every Court employee who had access to it and found no evidence that any of them had leaked it". Which strongly implies that it was leaked by someone who isn't a court employee, ie a justice. Like Alito. Who's leaked decisions before. https://www.nytimes.com/2022/11/19/us/supreme-court-leak-abortion-roe-wade.html
63 before seeing 62. The story in 63 is indeed of Alito doing exactly that in another case.
43: I feel like my mother-in-law does this. It's not so much about bigotry but global warming. "You have to deal with reality" I.e. electrification of cars is impossible. I think she does feel that people will unfairly accuse her of racism " even though that's not her intent" so she feels li' she can't open her mouth without walking on eg shells, but she's convinced that there are people who are intenti9nally playing the race card. It's not exactly the same; I don't think she believes Jews or other groups are a certain way but there is an insistence that her prejudices are "just reality" that is similar.
Does she think the weather controls itself?