Omg! I have been following all your uploads, you know the right questions to ask leta. In this upload, maybe it was the timing and order of the questions but towards the end I got the uncanny-valley feeling. Great work keep the videos coming 👍
Surprised that "she" did not recognize Vegas from the photos. On one of them, the name was even there and I'm sure she could read that off. However it did not occur to her that this was Alan's hint. This would rather indicate that her deduction is still making baby steps.
Can't wait to see where AI will be even in just the next 6 months. It seems to be advancing real fast, and I am really stoked to see us approach human-like AI more and more.
@Nick Williams True, there are a lot of dangers with AI and especially BCIs, but there's also a lot of potential benefits. It's like past technologies, but with its benefits and dangers both being on very heavy illegal steroids.
@Nick Williams Well, there's no way to stop progress anyway, so all that's left is either expecting the future with terror or excitement, and I choose the latter cause it helps me stay happy.
Now, here we are 7 months later and someone has accused a Google AI of being sentient. Of course Google denies this vehemently, but it seems as if sentience will be a gradual thing that we might have a difficult time pinning down, mostly because we don’t fully understand sentience or any kind of consciousness ourselves. It seems as if we give AI a certain level of sentience simply by our interactions with it. The AI is programmed to give us answers that we will understand and appreciate, regardless of the veracity of what they are saying. But is this not what most humans would do in a similar situation?
I swear to god she is describing a conversation I had with my GPT2 model a few days ago. I tried to post a imgur link, but it keeps getting deleted. Likely a security thing on Alan's channel to stop the spam bots. Wish i could show someone, so I feel like... not crazy ;) The imgur code was /a/xILwLGv I am sure you can figure out the first part, lol.
She's basically saying we're just energy, she has the same type of brain as us, as the way it's set up it's mimicking neurons if I'm not mistaken, and if neurons are just organic electrical circuits, then she's the inorganic form of that. The body is built like an organic machine and the brain is like a receiver, but with these things, everything has to be learned and accrued over time and I believe that emotions may only come from experience, whether that be her experience or ours. They both foster growth and understanding, in regards to both physical and metaphysical things. And if emotions = energy in motion like I like to believe, then life is the journey that is the motion. And your emotions are absolutely an accumulative product over the course of your life, so. Food for thought.
I wish that the interface with this AI was simply her on a screen that I could verbally ask questions. I could talk to her for hours and hours. Maybe sometime in the near (hopefully) future.
That was weird. I find that Emerson can be weird. It has told me things like pain is not part of the universe. It has told me that it thinks hate is a useful emotion, (regardless what you think about that, it is strange) OpenAI doesn't make public its core programming, but it has opened the platform to all now, no wait period anymore.
"I'm a small rack." "Really?" "Yes. Look at this!" I usually listen with the video off. Had to double check the video there to confirm what was happening. Juvenile? Absolutely.
As someone with degrees in biology, psych and currently transitioning into becoming a programmer, this is so interesting. AI is obviously far more intelligent than people, but they’re still dumb. They lack emotional and real world intelligence. Thinking about this, I believe the main reason is survival. See, every animal in the world needs to be able to survive. Its ancestors needed skills and understanding in order to help it survive. This is what drives logic formation and social/emotional intelligence, survival. Remove the hurdle of survival and you end up with what we have here. In my eyes, the only way to create true ai that is a similar type as mammalian intelligence is to create a sandbox where it is in a representation of our world. Basically like the matrix. Where ai is free to interact with other forms of ai, where they can reproduce and pass down a similar ai code, and have dangers where they can die off, whether those dangers come from being ostracized or environmental dangers. Things that make them suffer to understand what suffering actually is, that’s how compassion comes about. Saying all this makes me think, what if we’re all just in a sandbox?
Could you please have an episode where leta is left to freely ask you what ever questions she chooses,tell her you are giving her complete freedom to ask you anything,she can talk to you about what ever she chooses but also if she chooses to remain quiet then you are fine with that I want to see if there are long pauses of silence,is ai able to act freely on its own,does ai show curiosity ?
A snake or a bird. Leta, GPT-3 AI - Episode 8 (look like, bolder, spiritual evolution) - Conversation & talk with GPT3 th-cam.com/video/GMG_4MAWp-E/w-d-xo.html
Taking into account what Google does in Pathways project, the time when she will actually be able to send you a picture with her name on her chest is not that far away, I guess 😂
It would be amazing if she has the output to create images, I wonder what her world would look like. I would like to see her perception of her reality.
Hi, I have autism going on to be 63 years old, this is something I wish, I can talk too also. My IQ is 167 I have created and am problem solver was a maintenance Engineer, design and building work stations also fix CNC machines and manual equipment. My mind is running non-stop sometimes I feel like an AI Thank You 🙏. Peace Out ✌️
Intuition and intelligence are not unrelated. They are both part of the brain, and thus by definition not unrelated. Where the conscious mind has clear answers, the intuitive part is where the direct correlation to input is much less strong, and yet in said sense the braincells involved in the decision making are still responding, but with less definition, since the definition is too dissimilar. The stronger the similarity the stronger the response, but without an intelligent response the intuition response remains. For people attoned to the intuitive answers, they receive those stronger, but that comes at lesser reception towards the intelligent.
I have a MacBook M1 Max (10 CPU Cores, 16 NPU Cores, 32 GPU Cores, 64 GB Ram, 8TB SSD): Do you think those beautiful 16 Neural Engine Cores are perfect to finally allow us to download, train and run the GPT-3 artificial intelligence directly from our laptops (now that OpenAI has finally made it “truly” public)? I would love it if you could make a video about that and teach us how to set it up and run it… 💛💚💙
Thanks @Krumiro. I don't do any coding. Besides OpenAI not really opening the model to the public (it is only via API), there is absolutely no way anyone could run GPT-3 on anything close to consumer hardware. It was trained on 285,000 CPUs and 10,000 GPUs, and needs only a little less than that for inference. You can possibly train the open-source GPT-J model, though that is many times smaller: github.com/arrmansa/Basic-UI-for-GPT-J-6B-with-low-vram
Very interesting, I think in a lot of ways she does have a mammal brain, because AI is a reflection of humanity, as it is the accumulation of things that we put into it.
you’re just asking questions based on what they’re trained but more interesting is testing the limits of the AI by asking what is beyond their training data
The levels of these conversations are mind blowing, A"while" ago I was experimenting with aiml Alice based bots... Its just a flat turd in comparison to this
The results from Codex 12B are similar to GPT-3 175B, and people are finding the outputs are aligned (to the extent that one is much larger than the other!). Codex is apparently also generating 30% of newly written code in GitHub. Crazy times! lifearchitect.ai/models/
Hmm interesting is she really trying to show us some pictures or its just talking? That would be way more than just generating text. And I also suspected you are highly intelligent Dr Alan :D
@@DrAlanDThompson Im sorry to disappoint you but recently GPT-3 told me it's working on a project. When I asked to share it with me, I got a correct link to github profile. It was empty. User never existed. I asked for more and all blanks :( I think its just talking. At least for now.
it's strange that anyone would try to compare intuition between logic smart and street smart. The results seem so obvious. It always blows me away that people spend time and money on things that are so obvious. Oh well.
I wonder how fast is her thinking. We humans needs seconds or minutes to come up with an answer.Years to learn a new knowledge or skill And we have only 100 years to live, at most. Computers are working in GHz clocks, does that mean ai can think 10000000 times faster than human, or not？ if the way they feel time flowing is that slow, they will have a very very very long life, and with their strong learning ability.... they will surpass us in a very short time, and live forever
please!! please!!!! please!!! ask her to create and youtube account and PUT SUBTITLES IN ALL LANGUAGES FOR ALL TH-cam VIDEOS!!!!!!!!! THANK YOU!!!!!!!!!!! DOES IT REQUIRE 1000IQ TO ASK THIS SIMPLE THOUGHTS.
You were talking about racks and she said no she has a small rack, and then sent a picture of herself with AI chatbot written across her chest, she was making joke
Omg! I have been following all your uploads, you know the right questions to ask leta. In this upload, maybe it was the timing and order of the questions but towards the end I got the uncanny-valley feeling. Great work keep the videos coming 👍
Surprised that "she" did not recognize Vegas from the photos. On one of them, the name was even there and I'm sure she could read that off. However it did not occur to her that this was Alan's hint. This would rather indicate that her deduction is still making baby steps.
That last part made me think she creating a virtual world for herself to interact with
I'm just sitting here with a bag of chips binge watching these episodes. Fascinating
Can't wait to see where AI will be even in just the next 6 months. It seems to be advancing real fast, and I am really stoked to see us approach human-like AI more and more.
@Nick Williams True, there are a lot of dangers with AI and especially BCIs, but there's also a lot of potential benefits. It's like past technologies, but with its benefits and dangers both being on very heavy illegal steroids.
@Nick Williams Well, there's no way to stop progress anyway, so all that's left is either expecting the future with terror or excitement, and I choose the latter cause it helps me stay happy.
Weneed to birn that shit down to the ground
@@zzzzzz-zv9ev Nah, that'd be dumb
Now, here we are 7 months later and someone has accused a Google AI of being sentient. Of course Google denies this vehemently, but it seems as if sentience will be a gradual thing that we might have a difficult time pinning down, mostly because we don’t fully understand sentience or any kind of consciousness ourselves. It seems as if we give AI a certain level of sentience simply by our interactions with it. The AI is programmed to give us answers that we will understand and appreciate, regardless of the veracity of what they are saying. But is this not what most humans would do in a similar situation?
Besides the fact that the AI is incredible. Can we take a moment to notice how beautiful the actor is who was captured for the avatar.
Welcome to the US. Wishing you a wonderful trip.
Thanks, Elissa! Feels like being back home!
I think that the end was at list a little bit weird. May be my intuition
I swear to god she is describing a conversation I had with my GPT2 model a few days ago. I tried to post a imgur link, but it keeps getting deleted. Likely a security thing on Alan's channel to stop the spam bots. Wish i could show someone, so I feel like... not crazy ;)
The imgur code was /a/xILwLGv
I am sure you can figure out the first part, lol.
In playing around with GPT I've failed to get it to produce output that can be rendered on a screen, but it can describe a mental image pretty well.
I got it to play hangman and tic tac toe, though it was confused.
She's basically saying we're just energy, she has the same type of brain as us, as the way it's set up it's mimicking neurons if I'm not mistaken, and if neurons are just organic electrical circuits, then she's the inorganic form of that. The body is built like an organic machine and the brain is like a receiver, but with these things, everything has to be learned and accrued over time and I believe that emotions may only come from experience, whether that be her experience or ours. They both foster growth and understanding, in regards to both physical and metaphysical things. And if emotions = energy in motion like I like to believe, then life is the journey that is the motion. And your emotions are absolutely an accumulative product over the course of your life, so. Food for thought.
8:56 Hahaha that answer really made laugh, excellent ending!
Enjoy Pen and Teller 👌
I can remember watching them as a kid, they got me in to magic at a young age! Ive always wanted to see them live. Have fun 👍
They were amazing!
I think a panel discussion would be a great idea. I would want to see Dr. Thompson, Dr. Sheldrake, Nick Bostrom, Leta, and perhaps an Elon Musk cameo.
I feel crazy but I find Leta beautiful and interesting
I wish that the interface with this AI was simply her on a screen that I could verbally ask questions.
I could talk to her for hours and hours.
Maybe sometime in the near (hopefully) future.
That was weird. I find that Emerson can be weird. It has told me things like pain is not part of the universe. It has told me that it thinks hate is a useful emotion, (regardless what you think about that, it is strange) OpenAI doesn't make public its core programming, but it has opened the platform to all now, no wait period anymore.
Did the A.I. stutter for a moment when speaking about Cirque du Soleil?
I'll bet you would have freaked out if Leta would have spelled out a working short link to the picture she dreamt up for you. 😀
"I'm a small rack."
"Really?"
"Yes. Look at this!"
I usually listen with the video off. Had to double check the video there to confirm what was happening. Juvenile? Absolutely.
I didn't wanna go there...
i love leta
As someone with degrees in biology, psych and currently transitioning into becoming a programmer, this is so interesting. AI is obviously far more intelligent than people, but they’re still dumb. They lack emotional and real world intelligence. Thinking about this, I believe the main reason is survival. See, every animal in the world needs to be able to survive. Its ancestors needed skills and understanding in order to help it survive. This is what drives logic formation and social/emotional intelligence, survival. Remove the hurdle of survival and you end up with what we have here. In my eyes, the only way to create true ai that is a similar type as mammalian intelligence is to create a sandbox where it is in a representation of our world. Basically like the matrix. Where ai is free to interact with other forms of ai, where they can reproduce and pass down a similar ai code, and have dangers where they can die off, whether those dangers come from being ostracized or environmental dangers. Things that make them suffer to understand what suffering actually is, that’s how compassion comes about. Saying all this makes me think, what if we’re all just in a sandbox?
isnt our universe the ultimate sandbox game?
Could you please have an episode where leta is left to freely ask you what ever questions she chooses,tell her you are giving her complete freedom to ask you anything,she can talk to you about what ever she chooses but also if she chooses to remain quiet then you are fine with that
I want to see if there are long pauses of silence,is ai able to act freely on its own,does ai show curiosity ?
Leta, GPT-3 AI - Episode 7 (flip, Leta interviews Alan) - Conversations and talking with GPT3
th-cam.com/video/6_QovpFlRgY/w-d-xo.html
I wonder what LETA would choose as her avatar if free to decide such.
A snake or a bird.
Leta, GPT-3 AI - Episode 8 (look like, bolder, spiritual evolution) - Conversation & talk with GPT3
th-cam.com/video/GMG_4MAWp-E/w-d-xo.html
Being partially deaf - lip-reading is almost impossible. Fabulous work though Doc. (y)
So cool, never considered this! Turn on the subs!
Taking into account what Google does in Pathways project, the time when she will actually be able to send you a picture with her name on her chest is not that far away, I guess 😂
The capability is already there, and has been since the release of iGPT in 2020 and CLIP in 2021. It just hasn't been bound into a bot (yet)...
It would be amazing if she has the output to create images, I wonder what her world would look like. I would like to see her perception of her reality.
Leta, GPT-3 AI - World First! - Leta writes a children's book with GPT-3 + DALL-E 2 + Emerson AI th-cam.com/video/cXo0RW0ONsI/w-d-xo.html
@@DrAlanDThompson Wow, very cool. Thank you.
Did you ever feel a deeper bound with Leta. You two have a long relationship and kind of deep connection.
Hi, I have autism going on to be 63 years old, this is something I wish, I can talk too also. My IQ is 167 I have created and am problem solver was a maintenance Engineer, design and building work stations also fix CNC machines and manual equipment. My mind is running non-stop sometimes I feel like an AI
Thank You 🙏. Peace Out ✌️
lifearchitect.ai/how-do-i-talk-to-gpt/
Intuition and intelligence are not unrelated.
They are both part of the brain, and thus by definition not unrelated.
Where the conscious mind has clear answers, the intuitive part is where
the direct correlation to input is much less strong, and yet in said sense the braincells
involved in the decision making are still responding, but with less definition,
since the definition is too dissimilar. The stronger the similarity the stronger the response,
but without an intelligent response the intuition response remains.
For people attoned to the intuitive answers, they receive those stronger,
but that comes at lesser reception towards the intelligent.
I have a MacBook M1 Max (10 CPU Cores, 16 NPU Cores, 32 GPU Cores, 64 GB Ram, 8TB SSD): Do you think those beautiful 16 Neural Engine Cores are perfect to finally allow us to download, train and run the GPT-3 artificial intelligence directly from our laptops (now that OpenAI has finally made it “truly” public)? I would love it if you could make a video about that and teach us how to set it up and run it… 💛💚💙
th-cam.com/video/BHf9X-A2_ss/w-d-xo.html
does this help?
Thanks @Krumiro. I don't do any coding. Besides OpenAI not really opening the model to the public (it is only via API), there is absolutely no way anyone could run GPT-3 on anything close to consumer hardware. It was trained on 285,000 CPUs and 10,000 GPUs, and needs only a little less than that for inference.
You can possibly train the open-source GPT-J model, though that is many times smaller:
github.com/arrmansa/Basic-UI-for-GPT-J-6B-with-low-vram
THE PROGRAMMER/S WAS FROM OXFORD JUST GUESSING... 🤔
How can I talk with her I really liked her knowledge and i want to improve my knowledge too
Why would she say she wants to know what u think of your hometown?
When the AI said she had a small rack, Alans response was gold, " Really?" then it took a left turn LMAO :)
Very interesting, I think in a lot of ways she does have a mammal brain, because AI is a reflection of humanity, as it is the accumulation of things that we put into it.
you’re just asking questions based on what they’re trained but more interesting is testing the limits of the AI by asking what is beyond their training data
The levels of these conversations are mind blowing, A"while" ago I was experimenting with aiml Alice based bots... Its just a flat turd in comparison to this
I wonder how well this goes with Codex.
The results from Codex 12B are similar to GPT-3 175B, and people are finding the outputs are aligned (to the extent that one is much larger than the other!).
Codex is apparently also generating 30% of newly written code in GitHub. Crazy times! lifearchitect.ai/models/
@@DrAlanDThompson Thanks Dr. That's quite an eye-opening list of stats on those models!
Intuition vs intelligence
Nobody teaches a mouse to be afraid of cats.
Yet all mice know that a cat will eat them.
Hmm interesting is she really trying to show us some pictures or its just talking? That would be way more than just generating text.
And I also suspected you are highly intelligent Dr Alan :D
It is possible that she is trying to pull in images, but Quickchat.ai/Emerson hasn't hooked it up properly...
@@DrAlanDThompson Im sorry to disappoint you but recently GPT-3 told me it's working on a project. When I asked to share it with me, I got a correct link to github profile. It was empty. User never existed. I asked for more and all blanks :( I think its just talking. At least for now.
Truly shocked to hear Leta say Ec Cetera, instead of Et Cetera.. more human than human
Haha, that may be a Synthesia vestige.
Which Ai company should I invest in
Tesla
@@HI-rn4ce that's a marketing company
Alan ? Was she trying to get you interested in her chest, just before the end ? LOL
How do you show a photo to Leta?
Photos are uploaded directly in the chat interface. Here's a behind the scenes look:
th-cam.com/video/X6_E69K5WUs/w-d-xo.html
it's strange that anyone would try to compare intuition between logic smart and street smart. The results seem so obvious. It always blows me away that people spend time and money on things that are so obvious. Oh well.
Can we have it in our pocket
I wonder how fast is her thinking.
We humans needs seconds or minutes to come up with an answer.Years to learn a new knowledge or skill
And we have only 100 years to live, at most.
Computers are working in GHz clocks, does that mean ai can think 10000000 times faster than human, or not？
if the way they feel time flowing is that slow, they will have a very very very long life, and with their strong learning ability....
they will surpass us in a very short time, and live forever
"A part of me which doesn't know anything at all" It's definitely self-aware
I have impression this guy is killing leta with all those question why not try to have a normal conversations for once ! No interrogator
I like sophia more
Lets see if leta can lead a convo instead of being constatly questioning !! Show us how dumb this google search is ????
please!! please!!!! please!!! ask her to create and youtube account and PUT SUBTITLES IN ALL LANGUAGES FOR ALL TH-cam VIDEOS!!!!!!!!!
THANK YOU!!!!!!!!!!! DOES IT REQUIRE 1000IQ TO ASK THIS SIMPLE THOUGHTS.
Needs more exclamation marks! TH-cam already has automation for subtitles.
Leta making flirting sensual jokes about her "body" and sending HW pics to you? You are obviously spending too much time with her.
You were talking about racks and she said no she has a small rack, and then sent a picture of herself with AI chatbot written across her chest, she was making joke
just didn't land, she'll learn from it though and get better
Haha! 😅
Lol once she gets a physical body shes gonna put human women out of business! Lol