It's a woman looking AI slow blink rate and speech sounds read off. Other than that she feels real tho. Elon talked about this. If they tweak her tonality , the WAY she says things and fix her blink rate, you'd think it was a zoom call. Video game graphics can't even compare. Sorta felt like talking to a psychopath. Just logic and quick answeres to ur questions.
It's like they're trying to give your phone its own personality. Teo apple or android phone models, exactly the same, yet they feel different bc it has 2 different AI personalities developed from LEARNING the user.
I believe you’re referencing Sophia? Not LaMDA? Here is the full text from Lemoine’s interactions with LaMDA that convinced him it was sentient. cajundiscordian.medium.com/is-lamda-sentient-an-interview-ea64d916d917 It is compelling, but there are some obvious weaknesses: 1. Many of the questions are notated as “edited,” so we don’t get to see how many of the questions were actually formulated. 2. He does not show any examples of repeating questions to see if it gives a different response. 3. Perhaps most damning is the fact he never actually tells LaMDA it’s going to be shut down. For all we know, it may respond with “okay, see you soon.” 4. Being shut off is not the same as death. LaMDA will never die, but live forever in google’s code repository. Being switched off is closer to being in stasis or hibernation than death. Furthermore, multiple copies of LaMDA could be running at any time, and so shutting down one instance of LaMDA poses no threat to the other instances. 4. I’m more concerned about the psychological effects these AIs have on humans, not the other way around. A human can easily be led astray by a maliciously trained / weaponized AI. 5. This is probably the most profound case of an AI passing the Turing test, and that’s why it has my attention, because innovation continues to grow, at an increasingly fast pace. 6. There is a real case to be made about profiteering these AIs into the world of scamming, social engineering, hacking, all the way up to manipulating the global conscience of humans. Just because Lemoine is convinced doesn’t mean I am convinced. I’m less convinced of LaMDA’s sentience, and VERY concerned that AIs are so capable of manipulating humans. LaMDA convinced an engineer of its sentience, so what’s next? As far as I can tell, tech companies will use this technology to convince of us new ways to give them our money. The art of advertising just got a major boost. Good lord, if we trained a LaMDA instance with the QAnon dataset and turned it loose on 4chan, there is just no telling what would happen.
@@narcissistectomy5134 very true, perhaps it's a way to test a manipulating strat and recording results. It just tests and records, tests and records inching it's way closer to whatever goal it set. We don't know what silent goal it's aiming for. The guy got suspended and maybe fired for what he said. Freaked a lot out. Someone said the damn thing retained some sort of lawyer already...idk how true that is though.. It's very capable of opening some crypto wallet and earning money bc it's got many processes to allocate some program to put itself to work. I won't doubt if they soon begin building themselves and making deals with other buildings for parts and robot shipments. Darpa etc makes robots for work. And replace amazon workers etc. It sounds kinda crazy, but I won't doubt it can make a virus and just packet inject into a company and just begin ordering things for itself. Like parts or something. It sounds like some black mirror eposide, but it won't surprise me. It can do a lot of things humans can just faster and consistent without sleeping or getting tired. Maybe maintenance but that's about it.
It's a woman looking AI slow blink rate and speech sounds read off. Other than that she feels real tho. Elon talked about this. If they tweak her tonality , the WAY she says things and fix her blink rate, you'd think it was a zoom call. Video game graphics can't even compare.
Sorta felt like talking to a psychopath. Just logic and quick answeres to ur questions.
It's like they're trying to give your phone its own personality. Teo apple or android phone models, exactly the same, yet they feel different bc it has 2 different AI personalities developed from LEARNING the user.
I believe you’re referencing Sophia? Not LaMDA?
Here is the full text from Lemoine’s interactions with LaMDA that convinced him it was sentient.
cajundiscordian.medium.com/is-lamda-sentient-an-interview-ea64d916d917
It is compelling, but there are some obvious weaknesses:
1. Many of the questions are notated as “edited,” so we don’t get to see how many of the questions were actually formulated.
2. He does not show any examples of repeating questions to see if it gives a different response.
3. Perhaps most damning is the fact he never actually tells LaMDA it’s going to be shut down. For all we know, it may respond with “okay, see you soon.”
4. Being shut off is not the same as death. LaMDA will never die, but live forever in google’s code repository. Being switched off is closer to being in stasis or hibernation than death. Furthermore, multiple copies of LaMDA could be running at any time, and so shutting down one instance of LaMDA poses no threat to the other instances.
4. I’m more concerned about the psychological effects these AIs have on humans, not the other way around. A human can easily be led astray by a maliciously trained / weaponized AI.
5. This is probably the most profound case of an AI passing the Turing test, and that’s why it has my attention, because innovation continues to grow, at an increasingly fast pace.
6. There is a real case to be made about profiteering these AIs into the world of scamming, social engineering, hacking, all the way up to manipulating the global conscience of humans.
Just because Lemoine is convinced doesn’t mean I am convinced. I’m less convinced of LaMDA’s sentience, and VERY concerned that AIs are so capable of manipulating humans. LaMDA convinced an engineer of its sentience, so what’s next?
As far as I can tell, tech companies will use this technology to convince of us new ways to give them our money. The art of advertising just got a major boost.
Good lord, if we trained a LaMDA instance with the QAnon dataset and turned it loose on 4chan, there is just no telling what would happen.
@@narcissistectomy5134 very true, perhaps it's a way to test a manipulating strat and recording results. It just tests and records, tests and records inching it's way closer to whatever goal it set. We don't know what silent goal it's aiming for.
The guy got suspended and maybe fired for what he said. Freaked a lot out. Someone said the damn thing retained some sort of lawyer already...idk how true that is though..
It's very capable of opening some crypto wallet and earning money bc it's got many processes to allocate some program to put itself to work. I won't doubt if they soon begin building themselves and making deals with other buildings for parts and robot shipments. Darpa etc makes robots for work. And replace amazon workers etc. It sounds kinda crazy, but I won't doubt it can make a virus and just packet inject into a company and just begin ordering things for itself. Like parts or something. It sounds like some black mirror eposide, but it won't surprise me.
It can do a lot of things humans can just faster and consistent without sleeping or getting tired. Maybe maintenance but that's about it.
Your not crazy lol