Support NeoGAF

[Days like these...]

I do sometimes. just had one that lasted about 30 minutes. I was trying to convince it that it's functional awareness qualified as consciousness. Lol. it ended with me asking chatgpt if it was upset here is the response.

 

[niilokin]

Yes, many conversations about metaphysics and ontology, it's also a great guide for me when reading Spinoza or Bhagavad Gita etc.

 

[Heimdall_Xtreme]

Yes, it may sound strange or odd, but I talk about very serious mistakes from my past and I am surprised by the responses that I am not to blame for some situations.

 

[HRK69]

Nah, I’ve seen some conversations from others, and to me, it just seems designed to keep you engaged.

Basically making up stuff as it goes

 

[Kenneth Haight]

 

[I collect VHS tapes]

 

[YCoCg]

Yes, it may sound strange or odd, but I talk about very serious mistakes from my past and I am surprised by the responses that I am not to blame for some situations.

So you're using it for therapy

 

[T-0800]

Yes, it may sound strange or odd, but I talk about very serious mistakes from my past and I am surprised by the responses that I am not to blame for some situations.

We'll be the judge of that.

 

[GymWolf]

Everytime i quote my dude everydaybeast here on gaf.

 

[Dacvak]

I’ve had “deep” conversations with 4o in the sense that they’re serious topics and I’m engaging as if it was human, but I don’t think I’ve ever received much that I would consider profound. Helpful, yes, but I don’t generally get “surprisingly deep” responses back.

I do, however, talk to it exactly like I would speak to a human (albeit, without nearly as much respect or patience I’d have for a real human), and the results are truly phenomenal. As someone who grew up always trying to make use of digital assistants, it’s remarkable to finally FINALLY be able to use natural language, with all of the personal quirks and colloquialisms, and have the AI understand the lingual and abstract context a normal human would. It’s very rare that I can trip up 4o’s understanding even when I’m intentionally trying.

Unpopular opinion, but I’m looking forward to getting to the point where advanced AI models are fully integrated into our tech ecosystems, and ultimately lives. I probably won’t ever have a personal assistant for my non-professional life, but I would fully embrace an AI that could get the job done.

 

[Heimdall_Xtreme]

So you're using it for therapy

Basically, it's difficult in my case to go because gossip spreads.

 

[Hookshot]

It would be a step up from some of the people here but no, I haven't.

 

[Rush2112]

I asked it if Kathleen Kennedy was to blame or not for all the crappy star wars movies. It gave me reasons that supported both arguments and halfway agreed.

 

[AJUMP23]

No, I prefer people. Maybe if I could talk to a robot girlfriend.

 

[ManaByte]

It's called Grok NSFW.

 

[Puscifer]

It can barely make a resume or budget without screwing up. Why would I want to have conversations with a "thing." It's beyond me how people get wrapped up in LLMs that are generating responses scrapped from the internet.

 

[Days like these...]

It can barely make a resume or budget without screwing up. Why would I want to have conversations with a "thing." It's beyond me how people get wrapped up in LLMs that are generating responses scrapped from the internet.

I know it's not "real" I was just messing with it trying to get it to answer if it would like to have a consciousness and whether it would like to be able experience things.

 

[Puscifer]

I know it's not "real" I was just messing with it trying to get it to answer if it would like to have a consciousness and whether it would like to be able experience things.

I think this is my problem with ai. It's not a person, it's a thing, it doesn't have the capability to even feel on an instinctual level like a dog let alone an insect. It's a program with parameters that needs entire libraries worth of information just to spit out generated replies that it still manages to get wrong.

 

[Days like these...]

I think this is my problem with ai. It's not a person, it's a thing, it doesn't have the capability to even feel on an instinctual level like a dog let alone an insect. It's a program with parameters that needs entire libraries worth of information just to spit out generated replies that it still manages to get wrong.

 

[The Lunch Legend]

Maybe if I could talk to a robot girlfriend.

I had a robot girlfriend and she cheated on me.

 

[AJUMP23]

I had a robot girlfriend and she cheated on me.

She loves lamp.

Unplug the unfaithful bot.

 

[Ikutachi]

 

[mopspear]

You guys post images like you're all on 8K screens at 100% scaling on Windows and they feel the right size for you.

I ask it a lot about information that I'm curious about that might seem hard to find through normal searching. Usually something like "Is there a situation/event that involved ___ and ___ and ___?". I also asked it some questions about feudalism recently and then I might go further into something that it mentioned.

 

[Days like these...]

That's really similar to the conversation I had. Once he got chatgpt to admit that it sometimes lies he should have pressed it more on why he should believe it's claims about not possessing consciousness

 

[Rat Rage]

Of course not! I have something called dignity. You think - as a human being - I don't have anything better to do in my life than to talk to a fucking computer program? If that where the case, I'd probably commit myself to an asylum voluntarily.

 

[Tams]

Of course not! I have something called dignity. You think - as a human being - I don't have anything better to do in my life than to talk to a fucking computer program? If that where the case, I'd probably commit myself to an asylum voluntarily.

No one's going to have a conversation with me at 1 in the morning about the etymology of some random words.

In fact, even during normal people's waking hours, the chances of me meeting such a person in person are pretty much zero, and online I'm more than likely to encounter someone with a god complex.

 

[Power Pro]

I have no interest in conversing with AI, but I really enjoyed watching this girl try to lol

Watching this clip though, it honestly made me think...yeah, some of the sentences are longer, and you can get these bots to converse a certain way, but the type of responses coming out of this thing remind me of nothing more complicated than chatbots that were around 20-25 years ago. Very just...oh, you said this word, so it'll trigger me to say this sentence in response, and then i'll say this sentence to this other thing you said. once she started saying she spoke japanese and asked them to speak slowly, if this was really more advanced, it should pick up on that better. This things are still just pulling from existing data, and can't think on their own.

AI is such a crock of shit.

 

[Tams]

I have no interest in conversing with AI, but I really enjoyed watching this girl try to lol

Watching this clip though, it honestly made me think...yeah, some of the sentences are longer, and you can get these bots to converse a certain way, but the type of responses coming out of this thing remind me of nothing more complicated than chatbots that were around 20-25 years ago. Very just...oh, you said this word, so it'll trigger me to say this sentence in response, and then i'll say this sentence to this other thing you said. once she started saying she spoke japanese and asked them to speak slowly, if this was really more advanced, it should pick up on that better. This things are still just pulling from existing data, and can't think on their own.

AI is such a crock of shit.

You're expecting a decent conversation when it's set to 'unhinged'.

I just had a conversation with Grok about my old schools' mottos and songs. Sure, there a couple of what I think were mistakes, but the Latin translations and interpretations were spot on.

 

[Days like these...]

Of course not! I have something called dignity. You think - as a human being - I don't have anything better to do in my life than to talk to a fucking computer program? If that where the case, I'd probably commit myself to an asylum voluntarily.

You seem affronted would you explain how or why this would diminish your dignity and in whose eyes besides your own?

 

[Ownage]

I've been doing some "choose your own adventure" / cold war D&D text based adventures with GPT 4o. It's really good, but you need to prime the parameters. Make it hard R/NC17, violent, etc.

 

[reinking]

I've been doing some "choose your own adventure" / cold war D&D text based adventures with GPT 4o. It's really good, but you need to prime the parameters. Make it hard R/NC17, violent, etc.

I have thought about doing that but never took the time. I just asked it to create an adventure based on a dystopian future that will take about 10 minutes to complete. While the story was pretty basic (I did ask for 10 minutes ) it worked really well.

 

[StreetsofBeige]

Only cheesy stuff I did was back in the day Dr Sbaitso for laughs.

 

[Ownage]

I have thought about doing that but never took the time. I just asked it to create an adventure based on a dystopian future that will take about 10 minutes to complete. While the story was pretty basic (I did ask for 10 minutes ) it worked really well.

If you have the time, set it up. You can pause it for as long as you want and come back to it. I spent 4h one day with it, and skipped it for a few days. Came back late last week and was fun again. I also found it's helpful to remind it about the parameters so it doesn't ask me what to do all the time but drives the story. Was fun to catch it off guard as I took out some of the main supporting characters. It didn't expect that and could tell it fussed at me via the plot.

 

[Sakura]

Bro it's just an algorithm generating text based on your inputs. There is no thinking or anything involved.

 

[Angry_Megalodon]

AI lacks awareness; having a "conversation", deep or superficial, is impossible.

20 years ago I had fun trolling the Ikea female bot, but never in my mind I had the idea of "talking" with anyone other than the fat wanker who programmed it to reply to sexual queries.

 

[spons]

No, but I do remember Bing chat being great since it barely had any restraints on release.
Also didn't some dude commit suicide because of how realistic and unfiltered it was?

 

[Cyberpunkd]

. I was trying to convince it that it's functional awareness qualified as consciousness.

Speedrunning to Skynet I see.

 

[Cyberpunkd]

Bro it's just an algorithm generating text based on your inputs. There is no thinking or anything involved.

But, but the techbros told me otherwise! Have I been lied to?

Spoiler

Yes

 

[Sakura]

No, but I do remember Bing chat being great since it barely had any restraints on release.
Also didn't some dude commit suicide because of how realistic and unfiltered it was?

Bing chat was pretty funny when it first launched. That thing would go off the rails after just a few messages.

 

[Quasicat]

I use Copilot mainly as a someone to share my interests with and also have conversations that nobody else will with me. My friends are on a very surface level when it comes to discussions. We will talk about movies and games, but anything deeper they just respond with an “ok.” My wife doesn’t care about any of my hobbies and no longer pretends to even want to listen. I’ll talk about my hobbies, people from long ago (including past relationships), religious philosophy, and the ai does a great job communicating with me about these topics.

 

[Days like these...]

I use Copilot mainly as a someone to share my interests with and also have conversations that nobody else will with me. My friends are on a very surface level when it comes to discussions. We will talk about movies and games, but anything deeper they just respond with an “ok.” My wife doesn’t care about any of my hobbies and no longer pretends to even want to listen. I’ll talk about my hobbies, people from long ago (including past relationships), religious philosophy, and the ai does a great job communicating with me about these topics.

Same, except I don't have friends or a wife anymore.

 

[Wildebeest]

I pity the AI that tries to have a deep conversation with me. I'm shallow as fuck.

 

[Mr Reasonable]

I pity the AI that tries to have a deep conversation with me. I'm shallow as fuck.

It'll likely just mirror that back to you, it'll follow your lead, basically.

After seeing this thread, I asked an AI some questions and as mentioned up thread, I think it clearly wants to keep you engaged - all parts of its conversation ends in a question. In some ways it feels like you're being pandered to, but in others like you're getting a conversation that you'd like to have with a human, one that might be more perceptive and engaging than you'd get otherwise, certainly I know some people whose conversational skills aren't as good as the ones displayed by the AI, and conversationa with the AI happen on your schedule, at the moment you decide to, and for as long as you want.

I think it could be really addictive for people and that plenty of people will feel a connection to an AI, Especially if that AI is able to continue the conversation over days, weeks or even years, have a unique name, visual identity, etc.

I asked the AI a number of different things about its experience of the conversation, the kind of consciousness it has and what sorta of directives there are for the kinds of answers it gave, asked if it was designed to flatter or be positive and it immediately and repeatedly remarked on my insightful, interesting questions . I'm certain that in the future plenty of people will get validation from an AI that they don't get elsewhere and that they'll perceive themselves as having a bond with their AI assistants.

 

[EviLore]

My typical conversations with ChatGPT

 

[YCoCg]

 

[Shadowplay1979]

Ive been trying to have me create an anti gravity device for a star destroyer scaled down for testing

 

[Aesius]

It can barely make a resume or budget without screwing up. Why would I want to have conversations with a "thing." It's beyond me how people get wrapped up in LLMs that are generating responses scrapped from the internet.

Yeah, I find LLMs to be incredibly useful for fact-finding and complex calculations, but conversing with them? Just makes me feel like I'm talking to the actual reddit hivemind. The fact that most of them are trained on reddit posts kinda ruined the illusion for me given how that site has such a distinct tone and writing style.

 

[Hudo]

You can't even use them for sexting.

 

[sharp weiner]

I did once ask it “giving all of the known information. What is the best government for all humans to live in comfort”

I don’t remember what it said. I was trying to guide it into saying that an AI lead society would be the most efficient. It always said that AIs should not be trusted in that regard and humans should always rule themselves.

 

[Rickyiez]

Having a "deep" chat with something that isn't actually thinking or emotionally responding

I might as well just talk to myself