Hold your horses!
This Alan guy happens to be Dr Alan Thompson, Who are you to disparage his character so easily?
I don't believe in argument from authority and he doesn't challenge the AI in any way while presenting it as if it was just another person, indulging in fantasy and doesn't do enough to inform the viewers that it is just a presentation.
You display ignorancet on the subject of GPT series AI.
What is your basis of that? I have done my research, the one that keeps saying things about it that isn't true is you. You display unnecessary admiration of the AI that hasn't been earned. I think it is that that reminds me of a trumpist. Those may be harsh words, but I think you need to hear it, even intelligent people can be infected by the times of today. We share the same love for AI, but it is a tough love for me, I won't characterise it for something that it hasn't shown.
At the beginning of the test series (62 videos so far) he explained how he present the discussion with Leta as a female "person" to show the remarkable similarity to a real person speaking. Any lack of expression is not the AI 's fault but the limitation of the Synthesia corp. that is used to give Leta its human features.
The GPT-3 Leta video series
more.......
https://lifearchitect.ai/leta/
It does say it in the description so maybe he isn't as deceptive as I first thought, I still think it is a bit deceptive though, cause the fluent conversation he is having is an illusion and he plays along with it unlike what would be the actual experience. I can see how that can fool someone to think that it is more intelligent than what it is, cause it truly
is only an illusion to the viewers, if you understand how it works you would see that.
Other than the presentation Leta's texted responses are not in any way edited. The text is translated to voice by a text-to-voice translator ad is an exact copy of Leta's responses. There is no cheating, no subterfuge, no duplicity.
You are prejudicial in your criticisms. That is not critiquing, that is prejudice from ignorance.
It is deceptive to the casual viewer, it would take effort to break the illusion, I don't think it is made perfectly clear in the description that it is a presentation, a casual viewer would think that it was a realtime conversation, not put together.
I am going to be nice to you and ignore that ad hominem
I didn't use it as an argument, that's just the perception I get.
, but it is you who is the trumpist claiming the AI is a fraud from ignorance.
I didn't claim that it was a fraud, that is a strawman argument.
Accusing Dr Thompson of fraudulent practices is not civilized discourse.
One should always question authority.
I have seen all the videos and there are signs of emergent intelligent thought patterns in some of the AI's answers to questions and unsolicited voluntary tangential observations.
Great! I haven't seen any, can you show even one?
How many of the videos have you watched?
Probably not all of them, so if you have any video that clearly shows that it is conscious then be my guest (and the presentation won't fool me, so don't bother with good presentations).
The lack of flaws suggests an emergent artificial intelligence with signs of sentient awareness.
Human intelligence has flaws and we are still sentient and conscious. Lack of flaws aren't a suggestion of consciousness (though the flaws it does show are the flaws you would make if you're just blindingly given certain directions pointing at words at a dictionary without thinking about which words you choose.
Do you believe that you could flawlessly pass say a Mensa test? Would that prove you are not human?
No, that is what you are saying, that flawlessness suggests consciousness, that is not what I am saying, I've been pretty clear that it is the kind of flaws that would suggest one way or another.
I bet the AI could pass a Mensa test. Now that would be a legitimate test of IQ, no?
I don't think it could pass it, maybe it would be a test of IQ, but do you really think that low-IQ people are less conscious than high-IQ people? What does IQ have to do with consciousness? I believe even dogs and cats have consciousness, would they be able to pass any IQ test?
You should heed your own cautionary tale and inform yourself of the facts without the juvenile knee-jerk responses. No one claims Leta is human, but GPT3 and its bigger sibling GPT4 are beginning to show the evolutionary emergence of sentience from their extraordinary complex patterns of data processing.
In what way is it beginning to show emergence of sentience?
Max Tegmark explains how that may be the case with AI. See post #20.
Maybe the next generation, but I think we have to have a general AI for that to happen, and also uninterrupted realtime processing (which would enable a train of thought) and long term memory. To me that would be the minimum requirement for consciousness, sentience and self-awareness for the AI as a persona or "individual". Perhaps there are seeds of consciousness starting to emerge, there's not enough evidence to make any conclusions and without that evidence I won't make any claim that it is, and consciousness, self-awareness and sentience are three different properties and I believe that there could be a general consciousness of the system but not of the AI specifically as separate from the system, and I don't think it is self-aware as the AI persona or sentient as the AI.
