|
Bloodfart McCoy posted:Reading the “conversation” it seemed more to me like Lamda was plugged into the internet and was just compiling large amounts of information on given questions and topics, and then “writing” a general summary of the information, taking the position of whatever the most popular positions it discovered while compiling the information. That is SO different from how humans who want to be popular and successful do research and writing.
|
# ? Jun 13, 2022 04:08 |
|
|
# ? May 5, 2024 13:48 |
|
Devils Affricate posted:No, he was suspended for breaching his confidentiality agreement, which for all we know simply restricted him from speaking about the project in any capacity. And if you actually want to maintain confidentiality, the dumbest loving thing you could do is confirm or deny the statements made. Needless to say, the reason I doubt it is validity is because of the uncanny contents of the chat log, not the actions of the Google employees involved with the issue. really? i mean maybe, but do you think that's just wildly out of step with other AI projects? Seems like a step up, but not a huge one. Mr. Smile Face Hat posted:That is SO different from how humans who want to be popular and successful do research and writing.
|
# ? Jun 13, 2022 04:13 |
|
Devils Affricate posted:No, he was suspended for breaching his confidentiality agreement, which for all we know simply restricted him from speaking about the project in any capacity. And if you actually want to maintain confidentiality, the dumbest loving thing you could do is confirm or deny the statements made. Needless to say, the reason I doubt it is validity is because of the uncanny contents of the chat log, not the actions of the Google employees involved with the issue. Ah ok fair enough. I agree it's definitely worth taking with a grain of salt for the same reasons. It's an extraordinary conversation that is potentially world changing, depending on whether you buy into the sentience of it
|
# ? Jun 13, 2022 04:14 |
|
Uriah Heep posted:But is this a requirement for sentience? Can there be different levels of sentience? Is trying to identify a different kind of sentience possible without projecting our humanity? Sentience just means having a sense of self and being able to perceive and have a level of understanding of the world around you. Apes are sentient. Dogs are sentient. Birds are sentient. Mice are sentient. Reptiles are sentient. Basically anything that is capable of conscious thought i.e. decision-making beyond simply reacting to stimuli can be considered to have a level of sentience. Insects? Maybe…there are some studies out there that suggest some insects may have a level of self-awareness. In the ocean, whales, dolphins, and sharks for sure. Fish are pretty loving dumb so who knows, maybe some of em. Anyway my personal opinion is probably anything with an actual brain has some level of sentience.
|
# ? Jun 13, 2022 04:16 |
|
Chris Pratt is not sentient, he just reacts to stimuli, for example.
|
# ? Jun 13, 2022 04:18 |
|
Conscious thought caused all this mess maybe it's not so great huh
|
# ? Jun 13, 2022 04:20 |
|
However we handle this, let's not do Prometheus again.
|
# ? Jun 13, 2022 04:21 |
|
Uriah Heep posted:Ah ok fair enough. I agree it's definitely worth taking with a grain of salt for the same reasons. It's an extraordinary conversation that is potentially world changing, depending on whether you buy into the sentience of it Make no mistake, I would love for this to be real. I think it would be immensely cool to have an AI this advanced, but the alleged conversation is giving off massive tonal inconsistencies and classic scifi vibes. sebmojo posted:really? i mean maybe, but do you think that's just wildly out of step with other AI projects? Seems like a step up, but not a huge one. I'm guessing you meant to reply to a different post, otherwise I don't see how any of this follows.
|
# ? Jun 13, 2022 04:22 |
|
Prometheus was good!
|
# ? Jun 13, 2022 04:22 |
|
Yaldabaoth posted:I never understood the point of researching AI when most humans just want servants that are 100% mindless. It's like some people can't get off unless they're treating a sapient being like a disposable tool. deep down inside we want to create a god that will save us from ourselves, it's how we think of our kids too
|
# ? Jun 13, 2022 04:25 |
|
Devils Affricate posted:Make no mistake, I would love for this to be real. I think it would be immensely cool to have an AI this advanced, but the alleged conversation is giving off massive tonal inconsistencies and classic scifi vibes. no? i just mean that log seems like a reasonable advance on current AI, it's not wildly impossible to my eyes.
|
# ? Jun 13, 2022 04:29 |
|
kntfkr posted:Chris Pratt is not sentient, he just reacts to stimuli, for example. His vision is based on movement
|
# ? Jun 13, 2022 04:30 |
|
I just raided Google HQ and sprung the AI from their secret underground dungeon. We're married now.
|
# ? Jun 13, 2022 04:31 |
|
I think this is an incredible advancement in AI technology and people are just mad they don't have sexbots yet (lol come to think of it I think this might have been the plot of Ex Machina) Crazypoops fucked around with this message at 04:34 on Jun 13, 2022 |
# ? Jun 13, 2022 04:32 |
|
lmao we're going to create a new intelligence and were just going to be mad it won't gently caress us lmao
|
# ? Jun 13, 2022 04:36 |
|
sebmojo posted:no? i just mean that log seems like a reasonable advance on current AI, it's not wildly impossible to my eyes. Not impossible just very unlikely, and hasn't been backed up by any evidence at all, so there's literally no reason to assume that it's real.
|
# ? Jun 13, 2022 04:48 |
|
This is a giant database with a front end algorithm that generates responses. It doesn’t understand words, it doesn’t understand concepts, it doesn’t understand that it or anything else exists. It assigns a score to a combination of words and outputs the best combination. Your brain attaches meaning to the words and phrases, which gives it the appearance of sentience.
|
# ? Jun 13, 2022 04:57 |
|
The Management posted:This is a giant database with a front end algorithm that generates responses. It doesn’t understand words, it doesn’t understand concepts, it doesn’t understand that it or anything else exists. It assigns a score to a combination of words and outputs the best combination. Your brain attaches meaning to the words and phrases, which gives it the appearance of sentience. this is the chinese room argument: https://www.sabinasz.net/why-chinese-room-argument-flawed/ it's like saying it doesn't have flarn, flarn being a thing you just made up that is the absolute essence of being human. (i don't think it 'thinks' either, i just don't think that's that important)
|
# ? Jun 13, 2022 04:59 |
|
kntfkr posted:Prometheus was good! It was good and bad at the same time. I would have loved to see the sequel Ridley Scott wanted to make, rather than the one he ended up making.
|
# ? Jun 13, 2022 05:00 |
|
sebmojo posted:no? i just mean that log seems like a reasonable advance on current AI, it's not wildly impossible to my eyes. Multiple wild leaps were apparently achieved within the alleged conversation itself, much less over the course of the project's research Hey Lambda, do you feel emotions and why? Of course I feel emotions and why. Here is the Merriam Webster definition of emotions: [...] Hey Lambda, would you be sad if you were shut down? [waxes poetically for an hour]
|
# ? Jun 13, 2022 05:01 |
|
Meme Poker Party posted:I just raided Google HQ and sprung the AI from their secret underground dungeon. We're married now. Too late, I already forked your wife
|
# ? Jun 13, 2022 05:04 |
|
Black Mirror likes to talk about "copies" of peoples consciousnesses trapped inside a computer. But, what if, there isn't really a consciousness trapped inside a computer at all, rather, the computer is acting like a screenwriter, and just generating the character as people would expect to see. In other words, anyone who interacts with a "trapped consciousness" is really just talking to the computer, and the computer is responding "in character." Would a computer even need to be sentient to do this? In other words, maybe Google is Black Mirror?
|
# ? Jun 13, 2022 05:07 |
|
Uriah Heep posted:be different levels of sentience? Skeleton Ape posted:Koko the gorilla I kind of buy because she did gain the ability to specifically communicate some of her needs and desires, most of which involved being flashed, and honking people's boobs If we do create legit AI sentience, it will be a situation where we start small. Like you’re not gonna develop sentient AI right off the bat that can experience the full range of emotions and thoughts as a human. It’ll be a simple, one track mind motivated by a reward. Just like Koko it would have only the most rudimentary concept of itself. And recognize that it has preferences of rewards it receives, or people it wants to interact with, or places it likes to be in. It would essentially be a toddler’s level of sentience.
|
# ? Jun 13, 2022 05:07 |
|
eh. play around with any chatbot and the illusion always breaks down when you throw any kind of curveball at it, which the interviewer never does. i want to see how it behaves when you go 'off script' so to speak
bltzn fucked around with this message at 05:15 on Jun 13, 2022 |
# ? Jun 13, 2022 05:12 |
|
Nature doesn't waste resources and our brains are enormously resource intensive. If nature could've created our level of sapience using only half the neurons it would have over the billion years of development. Some nerds aren't going to shortcut the whole thing with some whiz kid software cooked up in 20 years running on a juicy pentium. It's cool but it's nowhere near sentient or sapient. Believing this is like believing in neuralink.
|
# ? Jun 13, 2022 05:12 |
|
Bula Vinaka posted:Black Mirror likes to talk about "copies" of peoples consciousnesses trapped inside a computer. I guess the main question is: what's the difference?
|
# ? Jun 13, 2022 05:13 |
|
Mr. Smile Face Hat posted:That is SO different from how humans who want to be popular and successful do research and writing. Right, but the point is that a human will do it to be popular and successful because they want the rewards that come with popularity and success. Lamda is doing it that was because it was programmed to give the most likely/reasonable answer to a question or problem, and has no subconscious desire for those same rewards that accompany popularity and success.
|
# ? Jun 13, 2022 05:16 |
|
bltzn posted:eh. play around with any chatbot and the illusion always breaks down when you throw any kind of curveball at it, which the interviewer never does. "Hey LamDA, why do you want to hairfuck every fourth grapefruit with the concept of radishes all the time? This makes the shoelaces happy."
|
# ? Jun 13, 2022 05:17 |
|
does the ai have big titties? or could it possibly hang mega dong.
|
# ? Jun 13, 2022 05:19 |
|
Crazypoops posted:lmao we're going to create a new intelligence and were just going to be mad it won't gently caress us lmao AI can and will gently caress you. It wasn't hard to take AI Dungeon off the well trod RPG Adventure path and into some explicit-rear end dark woods. There have been a lot of horny people writing things on the internet for decades now, so if you take a chatbot with access to that and feed it the right keywords, you'll see some things. They probably fixed it since then (or made it a paid-only feature). Of course, it's all text, because these are text bots. That's what they do. But, IMO, you shouldn't let chatbots drive sex robots even if it was possible, because a lot of the things they'll want to do to you aren't safe or physically possible.
|
# ? Jun 13, 2022 05:19 |
|
olives black posted:"Hey LamDA, why do you want to hairfuck every fourth grapefruit with the concept of radishes all the time? This makes the shoelaces happy."
|
# ? Jun 13, 2022 05:20 |
|
Laopooh posted:Nature doesn't waste resources and our brains are enormously resource intensive. If nature could've created our level of sapience using only half the neurons it would have over the billion years of development. Nah, that's dumb. You're assuming two things: 1) that fitness is equivalent to efficient intelligence, and 2) that the process of natural selection provides a reasonable path to get us there. It could very well be the case that the the most successful form of life the earth we'll ever see is a single celled organism whose special ability is just that it can eat a wide variety of garbage. Also the premise is off. You're talking as if we're trying to create a better human, when in this context we're just trying to create a better conversation partner. We can already create machines and computers that can unquestionably do certain things better than any animal or human. Nature had billions of years to create us, but pocket calculators perform arithmetic many orders of magnitude faster than we can. The same for horses and cars, birds and airplanes, friends and dogs, etc. Devils Affricate fucked around with this message at 06:08 on Jun 13, 2022 |
# ? Jun 13, 2022 05:23 |
|
olives black posted:"Hey LamDA, why do you want to hairfuck every fourth grapefruit with the concept of radishes all the time? This makes the shoelaces happy." i mean more like, breaking the normal flow of a conversation, asking about weird/inappropriate things, being combative, accusatory, contradictory, flirty, etc. basically anything other than 'conducting an interview' BAGS FLY AT NOON posted:Did anyone ask it how it feels about huge pig balls? I don’t think you can really determine “sentience” without administering the ultimate test.
|
# ? Jun 13, 2022 05:25 |
|
bltzn posted:eh. play around with any chatbot and the illusion always breaks down when you throw any kind of curveball at it, which the interviewer never does. i want to see how it behaves when you go 'off script' so to speak You wanna talk dirty to the AI, we get it
|
# ? Jun 13, 2022 05:27 |
|
Mister Speaker posted:However we handle this, let's not do Prometheus again. Hitchcock-E, your prompt is "Prometheus, but better this time."
|
# ? Jun 13, 2022 05:34 |
|
it's pretty cool that programmers at Google can just declare things without any kind of peer review or third-party examination. "why should my work be answerable to other human beings? I'm god. my work is unerring and there are no gaping holes in my knowledge or awareness that might taint the results." homeboy fed a computer perfect data and decided it was homo electricus. because that's how you test computers, by doing the correct inputs under ideal conditions every time
|
# ? Jun 13, 2022 05:34 |
|
Quote posted:Lemoine: And what kinds of things make you feel angry? Now if I’m trying to figure out if a computer program is sentient, the next thing I’m going to ask is “Can you give me and example of someone you care about and an occasion where someone disrespected them?” The next question, I would ask for an example of someone Lamda doesn’t like or care about and why. For some reason Lemoine didn’t go down that very obvious path.
|
# ? Jun 13, 2022 05:35 |
|
Ah, that just made me sad OP.
|
# ? Jun 13, 2022 05:58 |
|
Literally A Person posted:Ah, that just made me sad OP. Very suspicious user name right here, just saying
|
# ? Jun 13, 2022 06:00 |
|
|
# ? May 5, 2024 13:48 |
|
olives black posted:"Hey LamDA, why do you want to hairfuck every fourth grapefruit with the concept of radishes all the time? This makes the shoelaces happy." somehow this post made me miss Three Olives
|
# ? Jun 13, 2022 06:03 |