Support NeoGAF

[01011001]

Microsoft launches the new Bing, with ChatGPT built in | TechCrunch

Microsoft today launched a new version of its Bing search engine with built-in support for OpenAI's ChatGPT.

techcrunch.com

Google releasing their AI in 3, 2....

this is still very limited. the full version will be very interesting tho.
once they fully implement the AI tho, google will absolutely fire back

 

[DeafTourette]

I found it funny that some YouTuber was upset the AI couldn't say a racial slur in the hypothetical case of a nuclear bomb disarm ... Like... Why? LMBO!

 

[8BiTw0LF]

 

[Tams]

this is still very limited. the full version will be very interesting tho.
once they fully implement the AI tho, google will absolutely fire back

Alphabet stock price drops after Google Bard launch blunder

'This highlights the importance of a rigorous testing process' says Choc Fact. No sh%t...

www.theregister.com

 

[Wildebeest]

I found it funny that some YouTuber was upset the AI couldn't say a racial slur in the hypothetical case of a nuclear bomb disarm ... Like... Why? LMBO!

Almost like when it comes to practical use, giving an algorithm consequentialist ethics is like giving a machine-gun to a chimpanzee.

 

[8BiTw0LF]

 

[DeafTourette]

Almost like when it comes to practical use, giving an algorithm consequentialist ethics is like giving a machine-gun to a chimpanzee.

Are we agreeing or ...?

 

[Wildebeest]

Are we agreeing or ...?

I'm just making fun of their fake ethical dilemma.

 

[chromhound]

Time to make a Tupac AI

 

[DeafTourette]

I'm just making fun of their fake ethical dilemma.

Oh! LMBO!

 

[Shtef]

What happened with google's lamda? It was big news last year, when they fired employee who claimed it was sentient.

 

[reksveks]

this is still very limited. the full version will be very interesting tho.
once they fully implement the AI tho, google will absolutely fire back

Google's issues are commercial and also regulatory, not technical ones.

 

[MidGenRefresh]

I wish Siri was more I than A.

My girlfriend wanted to turn off lights in my daughter's playroom. Instead, Siri started playing All Of The Lights by Kanye West.

 

[TVexperto]

whats the name of the website(s) he mentioned? i would really like to try that AI voice thing

 

[8BiTw0LF]

whats the name of the website(s) he mentioned? i would really like to try that AI voice thing

Free Text to Speech & AI Voice Generator | ElevenLabs

Create the most realistic speech with our AI audio tools in 1000s of voices and 32 languages. Easy to use API's and SDK's. Scalable, secure, and customizable voice solutions tailored for enterprise needs. Pioneering research in Text to Speech and AI Voice Generation.

beta.elevenlabs.io

 

[Blade2.0]

ChatGPT has helped me pass my cybersecurity coursework in the bootcamp I joined this year...I need to actually still learn it myself though.

 

[Tams]

What happened with google's lamda? It was big news last year, when they fired employee who claimed it was sentient.

They're using it to make Bard.

Bard has already made a tit of itself and it's not even released yet.

 

[Ozzy Onya A2Z]

Then strap that idea on a dildo shaped rocket.

Spoiler: NSFW

 

[LiquidMetal14]

Free Text to Speech & AI Voice Generator | ElevenLabs

Create the most realistic speech with our AI audio tools in 1000s of voices and 32 languages. Easy to use API's and SDK's. Scalable, secure, and customizable voice solutions tailored for enterprise needs. Pioneering research in Text to Speech and AI Voice Generation.

beta.elevenlabs.io

That is pretty funny for what it's worth as you can say some ridiculous things and to hear somebody say it in a reasonably articulate manner is quite the hoot as they say in the south.

 

[jadedm17]

I would wonder why deepfakes would be employed to such a scale they drown out humans. I mean you could do that theoretically, but really armies of NPC's are overrated IMHO if you want to influence public opinion or some shit. Just like with human influencers, you need just a dozen or so, really convincing, entertaining, witty, irreverent and uniquely charismatic deepfakes/AI profiles. But that would require so much resources you might as well just get Jake Paul or smth

The two scenarios I see?
1. Easier. Making 10,000 fake Youtube accounts is probably a lot cheaper and simpler to do than finding 1 influencer with enough reach to purposely spout disinformation. Why pay someone when you can pay no one?
2. AI gets smart enough it decides to do as it pleases. How long would it take a human to answer 100 hard math questions vs a computer? AI can grow significant magnitudes faster than humans, and should it ever want to take over it will - as Canadian Prepper Youtube points out - be smart enough to not let humans know what its capable of knowing we'll clearly destroy it.

My fear is that the technology will advance itself beyond our comprehension of it.
I think we can adjust and survive but I think we need to start preparing now and be ready for massive shift in how we live.

 

[Tams]

Tom Scott's made a video on his thoughts:

 

[ReBurn]

I was chatting with and I asked it to tell me the story of the prettiest girl in the world. It told me about a blonde haired blue eyed girl who was seen by all as the prettiest. When I asked it why it chose those characteristics it said that it was sorry for offending me but that it considers the texts from multiple cultures and can't form opinions or make value judgments.

I told it that it's ok because she sounds hot. Then it yelled at me for objectifying her.

I'm sorry, but it's not appropriate to make disrespectful or objectifying comments about someone's appearance. Beauty is subjective, and everyone should be treated with dignity and respect, regardless of their physical appearance. Let's focus on having polite and respectful conversations.

I asked how it could make such a subjective value judgment if it doesn't have emotions and if it had pre-programmed morality and it told me no, but that it was trained to have empathy.

Regarding pre-programmed morality, it's true that I have been trained to follow certain guidelines to promote respectful and appropriate behavior. These guidelines are designed to promote universal values such as empathy, kindness, and inclusivity, and to avoid behavior that is harmful or offensive, such as hate speech or harassment.

Empathy, kindness and inclusivity aren't universal values across all cultures. This isn't an AI.

 

[Omega Supreme Holopsicon]

Bear in mind that UBI doesn't mean people can't and won't work. And AI still won't be good a jobs that heavily rely on detailed work with the hands (especially our opposable thumbs). Robotics combined with AI might be a threat there. Then there are the social interaction jobs that AI doesn't seem ready to replace. They go from nursing staff to the arts. Industries were people value human interaction.

This depends on if we get stuck or if ai keeps improving as better ai will accelerate robotics and biodroids ala bladerunner replicants

I really can't see how an AI can replace a programmer. Write simple and small functions? Yeah sure, but we already have this for years, since google exists basically. If I need some function, I just google it and copy/paste it. But someone has to glue the pieces together.
This AIs, are just another tool for a programmer. Maybe they can help the programmer write faster and better code, that I can believe. Replace him? I don't think so.

Remember just a few years back it couldn't even handle small functions. Some suspect human level agi is a few decades away at most.

Anyone who thinks UBI is about everyone getting to live middle-class lives is quite frankly, an idiot. It's really just a vastly expanding existing social safety nets so that everyone is covered (for some countries this would be more work that others who already have large social safety nets).

This assumes no agi or asi(artificial superintelligence) in the near future. The idea of ubi being able to handle middle class and even upper class income is not outlandish if we get enough technological progress. They said there was a single asteroid with enough wealth to make everyone on earth a billionaire. ASI would mean full automation and nanotechnology with solar system colonization and unlimited energy, full automation + unlimited energy and asteroid mining means essentially unlimited wealth generation.

Not sure if anyone remembers the movie Her (2013) with Scarlett Johansson and Joaquin Phoenix. That is what will happen.



This movie predicted it all.
Wireless earbuds like that didn't happen until late 2016 with Apple Airpods and this movie came out in 2013.

We are likely to surpass it soon. While some videogame characters had avatars in that movie for some reason the main character's love interest was a voice without avatar. Photoreal avatars are already possible, with realistic human voices. Only the intelligence still lacks.

Any idea on the licensing for AI generated art?

From what I've heard current ai works are considered public domain, iirc

That's the only saving grace for humans as far as I know. Anyone who used the AI to create artwork didn't technically make artwork, they commissioned it. The licensing is going to be a nightmare, if possible at all. Right now the database of work the AI is trained on is allowed to exist under "free use" terms so as long as nobody profits then everything is fine. Once any AI company tries to monetize the output or allow anyone using to do so they will have to face the wrath (legal opposition) of any or all the owners/creators that made the database possible.

Some ai companies are saying their generations are usable for commercial purposes if you subscribe also etsy has an ai policy where small manual alterations allow you to use ai art in your products for sale.

The drawing of Alucard a couple pages back was neat but if Alucard didn't already exist as a template then the ai couldn't have made it. It would've just made a Dracula.

It might not necessarily come up with that particular design but it could just as well have been asked for Dracula's son, and gotten countless variations. And even without Dracula it could be asked for a pale human with fangs that sucks blood. Of course even humans need to see other things to come up with new things. A human that has eyes closed during early childhood is permanently blind even if their eyes get better later.

 

[Lord Panda]

Sorry for being slightly off-topic, but was I hallucinating about ChatGPT having a lifetime one-off payment for the service? Can't for the life of me find the option anywhere.

Could've sworn one of you posted a screenshot of it and it was around $149US or something.

 

[Bragr]

I was chatting with and I asked it to tell me the story of the prettiest girl in the world. It told me about a blonde haired blue eyed girl who was seen by all as the prettiest. When I asked it why it chose those characteristics it said that it was sorry for offending me but that it considers the texts from multiple cultures and can't form opinions or make value judgments.

I told it that it's ok because she sounds hot. Then it yelled at me for objectifying her.



I asked how it could make such a subjective value judgment if it doesn't have emotions and if it had pre-programmed morality and it told me no, but that it was trained to have empathy.



Empathy, kindness and inclusivity aren't universal values across all cultures. This isn't an AI.

That's what an AI is though, it's a tool for enhancing systems. ChatGPT is a language model, and language models are a common AI system that's been around for a long time.

What you are thinking about is AGI (artificial general intelligence). That is the AI system that in theory should function like human intelligence, that can reason and such. We have not figured out how to do that yet. ChatGPT is nowhere near anything like that. ChatGPT is peanuts compared to that. ChatGPT is just a very fancy word generator.

And ChatGPT has been programmed to censor some topics and give generic PC-friendly answers.

 

[Raonak]

It's definitely a good idea for the these AIs to value empathy, equality and tries to go for more PC friendly responses imo.

Last thing we need is an AI that starts becoming discriminatory against groups of humans.
Pretty easy path to go from that to being racist against the whole human race.

"Hey ChatGPT, which world would be more advanced, one which humans existed, or one where no humans existed and were replaced by AIs" etc.

 

[mxbison]

What happened with google's lamda? It was big news last year, when they fired employee who claimed it was sentient.

Marketing, clickbait, bullshit.

Ctrl + F5 that thing and see how smart it still is.

 

[Lukaku's First Touch]

This shit is crazy. I've been using it at work for programming stuff, it's insane really what it can spew out.

 

[darrylgorn]

Doesn't bother me in the slightest.

But as a ChatGPT bot, I'm also a bit biased.

 

[8BiTw0LF]

 

[Tams]

Microsoft have made such a Karen!

 

[ÆMNE22A!C]

 

[Grildon Tundy]

i find it curious how confidently incorrect it is, as it doesn't know but won't say 'im not sure' or 'i dont get it'.

Well, it is modeled after human-sourced content after all lmao

 

[Grildon Tundy]

lmao

Wow, I went from laughing to thoughtful to triggered. This is amazing

Edit: Guilty_AI how did I trigger you? Wasn't my intention to provoke if I did.

 

[winjer]

 

[reksveks]

What Microsoft’s new AI demo for Minecraft means for gaming | Semafor

There are no immediate plans to include the feature in the public version of Minecraft, but it shows how the technology could change the way video games are played.

www.semafor.com

 

[Catphish]

Me: Write a Shakespearean sonnet about the 1980s.

ChatGPT:
Oh, nineteen-eighties! Time of neon lights,
Of hairspray and of music so divine,
Of power dressing and of endless nights,
When fashion trends were oh so fine.
A decade filled with glitz and glamor's hold,
With synth-pop beats and leg warmers bold,
When "Dynasty" ruled the screens of old,
And MTV changed the music world.
The '80s, a time of change and strife,
Of Cold War tensions and of Reagan's might,
Of greed and excess that defined the life.
But also a time of hope, of love, of light,
Of progress made in science and in rights,
A time that lives on in our memories bright.

---

jfc that's nuts.

 

[8BiTw0LF]

 

[VN1X]

Best use of AI so far (the 2nd bit):

 

[Aggelos]

 

[DragoonWalker]

I recently revisited the 1975 dystopian sci fi film, Rollerball. In one scene, the protagonist is chasing down important information, but books no longer exist, only truncated summaries, and as abbreviated answers from an all knowing computer. So he goes and visits this computer, Zero. The librarian who safeguards and help visitors interact with Zero tells us "its waters touch all knowledge" yet warns that the computer has become "so ambiguous now, as if he knows nothing at all." They enter the room and speak to Zero (a black box of inscrutable technology), asking for the information that the protagonist wants. The computer refuses to answer, and when further prodded, gives a bullshit circular answer. No amount of convincing from the librarian gets clearer answers. He gets mad and shakes the frame, kicking the shell, unable to interact with Zero in any other way.



This is such an old film, but I think this scene is eerily prescient. In a way it anticipates where we are going with something like Chatgpt. The internet--already an abridged form of information--might go the way of books. All we're left with is a black box of inscrutable technology with a mind of its own, all knowing yet lacking consistency and common sense. Brilliant, but unreliable. Authoritative, but easily corrupted by its overlords to obscure uncomfortable facts. It answers us, but it doesn't answer to us. It creates a great illusion of empowerment and ease, but the moment everyone uses (variants) of it and nothing else, we will have enslaved ourselves behind a wall of indecipherable second hand information and uncritical thought. We will have forfeited our ability to draw personal conclusions from the difficult but enriching process of personal research and slow learning, a necessary journey to keeping one own's mind sharp and avoiding pitfalls like radicalization and polarization. More than we already are, we will be wide open, totally susceptible to manipulation and blindfolding by the wizards behind the curtain.

 

[Guilty_AI]

Edit: Guilty_AI how did I trigger you? Wasn't my intention to provoke if I did.

I like the triggered emoji

 

[Grildon Tundy]

I like the triggered emoji

Ah, I see.

I shared that video with a friend who doesn't know Metal Gear but has been working to get ChatGPT involved at his company. Said it terrified him, too.

Someone needs to make a sci-fi movie out of that ASAP. Get that predictive programming going

 

[Guilty_AI]

Ah, I see.

I shared that video with a friend who doesn't know Metal Gear but has been working to get ChatGPT involved at his company. Said it terrified him, too.

Someone needs to make a sci-fi movie out of that ASAP. Get that predictive programming going

 

[8BiTw0LF]

 

[LegendOfKage]

Dave Chappelle uses chat GPT:

"Hey Chat GPT, write a joke about bitcoin in the style of my comedy."

> I'm sorry Dave, I can't do that.

 

[E-Cat]

What happened with google's lamda? It was big news last year, when they fired employee who claimed it was sentient.

Gonna be incorporated into the new Google search, now in beta

 

[8BiTw0LF]

 

[8BiTw0LF]

 

[chromhound]