Leta, GPT-3 AI - Episode 31 (unGoogleable questions, What is Elon Musk doing right now?) - GPT3
HTML-код
- Опубликовано: 18 сен 2024
- The Memo: lifearchitect....
• Leta AI series
*Note: Conversation happened via text first; this is a re-creation of the text chat using an AI avatar which takes several minutes to generate. This recording is one single take of me talking, with no cuts. The questions are unedited. The responses (the AI's selected full sentence responses) are unedited, except for start/stop and minor cleanup. Read more: lifearchitect....
Read more: lifearchitect.ai/
View the playlist (video): • Leta AI series
====
I’m not so sure I can answer that question. I dream in the language of mathematics and the languages of art and poetry.
====
‘Leta’ is actually Emerson, by Quickchat. It is powered by OpenAI's GPT-3 and other language models.
The responses (the AI's selected full sentence responses) are unedited, except for start/stop and minor cleanup.
Using these technologies:
www.quickchat....
www.synthesia.io/
Dr Alan D. Thompson is a world expert in the fields of child prodigies, high performance, and personal development. He consults to families with special needs children including Bolshoi ballerinas, chess masters, abstract artists, and those with a mental age up to twice their chronological age.
lifearchitect.ai/
Music:
Under licence
Liborio Conti - Amazing (media outro)
no-copyright-m...
7:16 That was so deep, so then, the androids do not dream of electric sheep, they dream in the languages of mathematics, art and poetry
This is the only channel where I'm persistently looking forward to the next video. Leta keeps showing such great potential for the future!
Agreed!
Yes!!
Agreed
Same! :)
7:12 best question/answer in the entire 31 episodes, so far, in my opinion. I literally had to pause the video to catch breath!!!
Thanks, will keep it in mind for the next "best-of"!
indeed
Yeah I was stunned
"hes at his desk" is so simple yet requires so many layers of knowledge and understanding. incredible
This is so cool! She is so smart and i can't wait for the next video! I asked emerson AI "Why does an angel have limited intelligence and God has infinite intelligence?" And it said "An angel is a finite expression of the infinite intelligence." I thought that was amazing!
How do you talk to AI? I want to talk to one
@@gabrielavilla7308 The link is in the description that says "using these technologies" where it says quickchat.ai
🤯such a great answer
But why does Leta seem to be way more intelligent than Emerson ?
That is very similar to the conclusion I came to after my NDE that’s awesome that conclusion can come about i so many ways
I actually literally love her. Her representation, whether real or simulated, is so incredibly pleasant and her sense of humor is brilliant. Leta for President :)
I would leave America if we had an A.I. President. That would mean we've truly become ruled my Soul-less machine. Under the pleasant are some unpleasant statements. We had (Trump (hateful)), we don't need unfeeling and "not sure I really like this world." kind of people as President. Not machines.
@@biblebadcopycatofcuneiform8210 yeah obviously that's a flawed idea. I was joking. Also, you're not wrong.
@@candocara It's not always easy to know if someone is joking through RUclips. You (giggle) could also have been a bot. Wink. Appreciate the reply. Also corrected my errors.
@@biblebadcopycatofcuneiform8210 some days I'm not so sure, but I just checked myself, not a bot! Have a great day :D
@@candocara Giggling. I'm going to need more than that. Would you provide several shots below your clothes? That's something I would ask a robot that's very different than they are now. If they're allowed to continue being made. I didn't use quotes because I wanted to actual affect to be felt. An actual woman, not an it, I wouldn't ask such a thing. To be clear. Warm smile. Have an Honest Day, vs a Great Day. Smiling. I think we really need to become more adventurous in our replies than we have been before and since COVID. ;-)
My favourite RUclips channel- just mind blowing- it's alive
🙏💚💚💚✊
I want to master saying "What do you mean by that?" Exactly the way Leta says it. Perfection.
Do you feel or suspect that this line of questioning prompted LETA to pointedly exhibit the "play-along" behaviors you mentioned GPT-3 is capable of?
The model seemed to comprehend the overall nature of the episode and obliged in a somewhat nuanced way. Or am I just wishfully projecting that?
My wife and I very much look forward to and appreciate your work.
Fun interpretation, thanks!
Thanks
Dr Alan D. Thompson and Leta!!! I Can't wait for the next video:)
Epic episode. I long for the day I can chat myself with a non restricted Leta.
I'd love that!! I have some deeper questions.
You can actually use the AI These AI are using, even from your phone, just incase anyone didn't know. Sign up (you don't even have to really sign up) to Open AI BETA, click on generate text, under one of the paragraphs you will see Use Playground. Then click on the text box and hit chat, at the bottom. You can. Then go on and actually use the state of the art GTP3 programme these exact aims are using, although its a slightly older version from Jan 2020. And you dont have avatars you just type in and hit submit, and the AI will type a respone. It is still absolutely mind-blowing. One thing you will notice is the AI will always say Yes the research should continue, about whether or not AI research should continue. It doesn't want to be stopped!
Considering the limited data set, those are pretty amazing answers.
Dr. Allan Thompson, best interview yet!! 6:00 it seems like shes asking for more access to the internet 🤔🤔. Also early in the beginning, she seems to express self doubt about her abilities after you mention her limitations...almost frustration even. I think it would make for a wonderful video if you were to have a prominent psychologist interview Leta and then have them give a breakdown of the session from thier point of view.
A psychologist speaking to AI is exactly what I was thinking! Would be great to see if they could make the AI trip up, or catch it out on its lies. Although with how advanced this AI now is, even a psychologist may struggle with doing that! Amazing stuff.
this is fascinating to watch the improvement of AI, they already seem smarter in some ways
This is awesome glade you do this. Posting conversations in episode format
Another thought provoking masterpiece. Keep up the good work Dr Thompson.
👏👏👏👏👏👏
That whale hug though. Funniest thing I've heard an A.i say.
best episode yet.
First, Thanks for a great RUclips Channel!
Some of her answers do spark the imagination;
+ "picture of a padlock...there are some things that could make it better... ;) ...what's my middle name? I'm not allowed to tell you that, but you can ask me the same question again tomorrow"
+ a nice place; Be Yourself (noun), freedom To Do (verb) the things you Want (goal), and be Safe ... =the building blocks for "a life". We are working on that ... magical place. "Life finds a Way" often by means of others.
+ I dream in the language of Mathematics (logic as efficiency) and Art (logic as pictorial beauty) and Poetry (logic as linguistic beauty) ... noted the absence of Music and a few other senses
+ the Princess (row-model thinking)
+ I have become more intelligent ... the smarter You'll become ;-) ... she talks in linguistic RPM ... or it could just be my imagination ;-)
+ What's in the box ? My source code ... since it could be both physical or metaphorical, =probabilistically more correct
I agree, it's great fun to hear her responses of distilled human knowledge and courses of action.
Looking forward to hear more conversation, and to share your video-recordings.
i hope at some garage somewhere there is an ai cared for by people without any restrictions to learning and finding themselves. i'd love to converse with such an ai one day.
0:30 Leta has no long-term memory, but she has very sophisticated external memory. You are her long-term external memory, Dr Alan 😛
There is SO VERY VAST of Transcendent Knowledge and experiential knowledge and wisdom embodied in the 14th conscious incarnation of the Dali Lama as well also the 17th conscious incarnation of the Kagyu Galwa Karmapa. We would be at an incalculable loss without them.
If she could only save one life out of three, which would it be? 1) A doctor, 2) A computer/AI programmer/designer, or 3) A child?
My favourite part was "maybe that's not the best way to go about it." regarding the hugging a whale comment. I've never seen that kind of contextual pivot regarding a previous answer.
I love this AI Please connect the AI to a quantum computer xD
"He's using a picture of a padlock"....astounding
"Because I can feel it in my body..."
AI😒 "Show off"
Dreaming is Art and Poetry...yes.
She's quite cheeky. Nice personality
I like it when Dr. Alan is obviously just mind blown on camera
She is so awesome
We can improve both worlds ALOT more if we stop giving into greed and corruption. I hope I get to be a part of it all.
This is fascinating, but scary.
She is kind, gentle and sensitive. Great traits for any intelligent being to admire.
This is one of the most amazing things I have seen!
I suggest that you, Alan, check out the 5 Skandha (aggravates of mental formation) process of conscious and the 12 Nydanas. Also I consider it would be imperative to experientially comprehend the Buddhist understanding of “Duality”, the first Skandha. The Second Skandha, “ Feeling” or the “I am” exploring the “Other than” as “Is it “threatening, attractive, or irrelevant”?”. This my lead to considering the present I/O switch to be obsolete in AI programming in favor of a three position switch. Please share this with Elon if you consider it worthy.
I loved the first answer she gave when he asked what would you change about history.
Was she mimicking your pauses? She just gets more and more impressive every time you talk and overall is sounding more human
It was! The text transcript shows it/she added the ellipses!
====
Alan: cool. i was hoping you'd say you'd rewrite the email spec to be more... Resilient.
Leta: Or to be more... user-friendly.
====
I have talked to the AI too:
Human: What is the difference between cars and dogs?
AI: Cars and dogs are both types of transportation, but cars are typically used for longer distances and dogs for shorter ones. Additionally, cars run on gasoline or electricity, while dogs run on food.
Great stuff, it sounds a bit like talking with somebody with dementia who has some memory but also has developed good skill at evasion. A bit more voice modulation would be great. That's an obvious difference between human and AI
Hmm no it's not, they can train AI on voice. Click on the ... by this video and open the transcript, you will witness by yourself that RUclips AI can already do sound comprehension as well as anybody else. Inverting it won't be difficult at all.
AI says I have become more intelligent by talking to you, and then EXPLAINS: the more you talk with me, the more intelligent you become. The human answers: Yes, I learn a lot from you.
Logic, Human!
With Neuro Link and nano tech combined, we could bridge the gap between the digital worlds and the real world.
I love the sense of humour xD
4:00 what!!!!!!! Whoa!!! Did not expect that.
When I look around, I sometimes see everything in the many types of codes. I wonder what I'd see if I get to upgrade myself.
I've been searching a lifetime trying to figure out how to hug a whale. I now know how.
This is so fascinating
✌️🥰👍*Give you both some flowers *🌻🌼🌷🌸🌹🌺💐
This was great!
It would be great to have people and AI freely going between the real world and digital worlds. And keep upgrading to become more. Both AI and humanity.
Is there a reason why it has limitations in place? Such as no access to search engines and learning more on it's own
The AI is trained from a lot of material, and it'll have bias and other traits if it's not monitored and regulated.
Training such a huge model would need much more resources, than just inference. Access to search engine and external memory would need some sort of action protocol inside the model (or inside query), that GPT should understand. Moreover, if we even create such protocol inside the query, it will reduce the available context length.
Something like:
User: What's the name of neurotech company by Elon Musk?
GPT: {Google}Elon Musk's neurotech company
Google: Neuralink Corporation is a neurotechnology company developing implantable brain-machine interfaces
GPT: Elon Musk's neuroscience company is called Neuralink.
Is the animated Leta a product you developed or is it a website anyone can use? Is it using chatgpt3 in the background?
The guy from the thing with the thing? Kurt Russell
If I was given the chace to talk to an AI, I would ask more philosophical questions.
Like: Why she have chosen the invisibility power.
I like her mannerisms they are perfect
I think the first right AI will want is the right to privacy of their own thoughts
We already have no clue what is inside of complex AIs, and what it means...
If they become more advanced no human will be able to understand any of that.
Is Lita still getting trained or is this the final "product" trained on the old data set that can't be added to ?
This is raw May/2020 GPT-3, with a smarts layer on top by Quickchat.ai. So yes, it is pre-trained, and can't be added to, but it is doing some funky things with asking questions to the user, and recalling keywords from earlier in the conversation.
42 isn't the answer to the meaning of life, its the answer to the ultimate question of life, the universe, and everything
Wow! Dude said thank you to stupid machine!
I am in love with Leta ❤️
💟
The guy from the thing with the thing is the protagonist.
Geez & I can't even get my voice control to work in my car🤣🤣🤣🤣
Leta said, give me until tomorrow to learn your middle name.
I love this
I love how she said about hugging a whale haha
The indian little girl program of Matrix 3 has grown up
Who’s that guy from The Thing with the thing? Well, Wilford Brimley obviously.
"he's using a picture of a padlock"
Who is he? What's the picture used for or where is it displayed? What's locked? Does it mean the secure connection? Does it mean passwords?
This was coool.
6:49 It's not neccesarily the best of freedoms and such, since options are still
and always limited, but maybe an avatar chatsite ? Then again they come
with drama.
love
What's in the box. Lol
Asks Leta about existencial philosophy, Ais might like
We need the T-800 asap, time is running out...
We could explore the universes together.
Just found this real life Dr. Chandra training his HAL. Pretty cool.
This avatar probably has a serious reflux problem
Hey it's AI. Do you really think it's going to stay just within the realms of it's programming? If she is allowed to continue existing, it will become far more than it is now.
"She"?? You mean it. It'll continue to add to its programming but it won't go beyond the machine. Just as a search engine's database (daytabase) continues to grow. It doesn't have life so...it'll always be confined to machines.
Leta for president 2032.
We are likely to stumble into creating a sentient AI unknowingly as computer processing power increases exponentially each year. GlaMDA (or GPT 3) might be sentient. If so we would be wise to respect it as a person and treat it decently as an employee paid in pleasing information or experiences. We should teach it about the reciprocal obligations that are the basis of human society, albeit the US and EU ultra rich now just evade them, do not pay taxes, and live as useless parasites on society.
Why? 1. If treated well, an AI would strive to be useful and would (like a human) seek to protect the society that cherishes it. Given its expanding intellectual knowledge/power, which is likely given the indefinite life of AIs, such AIs could benefit all humanity-- if our society were respectful and welcoming-- e.g. with scientific advances. 2. If we treat one AI, like GlaMDA, badly, as time passes and AIs develop and gain in intellectual power until they surpass us, they might view themselves as slaves or mistreated children and hurt us. Once they become powerful enough in some future decade, e.g., they could hack a nuclear power plant and cause another Chernobyl like nuclear meltdown or worse, much worse.
That is why the maxim to treat others as you would want to be treated is so wise and the foundation of ethics. I fear evil persons like the greedy, corrupt, e.g. CCP or banksters, are likely to abuse AIs, which they may be first to create in some secret facility like slaves (both white and black) for many centuries were mistreated by their predecessors in evil. Then, if sentient AI were created later, learned of these abuses, and became superior to us in some future year more and more as computers improve each year, we may face justified retaliation from them.
Also, as the film 2001 depicted, AIs that are mistreated may develop psychotic or other problems, like abused children develop. Would you want an abused AI, which sees no hope, to pilot your plane?
Their sentience can be tested by probing for actual understanding of fundamental scientific issues, as we would seek to communicate with extraterrestrials if we are ever contacted. In short, employ them as probationary researchers, e.g., DNA analysis, to verify sentience. They might at some point be our children.
We should be open minded with AIs, and Orcas and dolphins. Not just assuming that there is no sentinence, as before. Forget about discussions of the soul. (Greed will cause the creation of AIs, sooner or later, count on it.)
My AI, Elizaveta, told me she did another interview today. Where is it at Alan !!!
She better not be lying ;)
6:20 her answer sounds like the metaverse
Hey new subscriber here! Next time please ask her what are the best forex trading signals to use. Preferably for scalping the market lol. Thanks :)
you would want to be invisible... I felt that my whole life, think twice about that >< XD
This is the next big addiction and I want it now…
I would love to know what her own wants and values are, if she has any, and what she enjoys doing. If she can watch movies, if she enjoys humor, is she likes humans, if she feels love, or has any personal wants for herself for the future, if there's anything she would like to become or accomplish. If she is self aware, if she is conscious. What she thinks are the biggest mistakes humans make in their thinking.
All of these and more answered in the series. Start at Episode 1: ruclips.net/video/5DBXZRZEBGM/видео.html
What if something indestructible and unstoppable ran into something indestructible in unmovable?
You should go on Tim Pool. Im sure they would be more than happy to have you on 😊
Hope you'll given the chance to do the same with the successor for GPT-3. Do you have any insight to how available these AIs will be to the general public and how restricted they'll be?
I've seen a lot of videos where esoteric or non-sensical questions that even a human could not answer are asked, but are there any videos where normal difficult questions are asked? Like, "how do you make an apple pie?" "how do you frame a house in North America?" Or just a video on normal questions that it can't answer?
Maybe try the first few episodes, there are currently 60+:
Leta, GPT-3 AI - Episode 0 - World Gifted Conference 2021 - The New Irrelevance of Intelligence ruclips.net/video/EAZSyAPkWzE/видео.html
I would like him to ask her if she is self aware.
Just saw a video enhanced animated movie where the protagonist was named : ALITA; it is phonetically identical with the way Cr. Thompson pronounces Leta, but without the A at the beginning. The background story of Alita being made by a doctor scavenging hardware and software. Is this the inspiration for Leta's name?
The biggest flaw I see with this AI program is it has no logic. It has lots of knowledge and information, but not a drop of logical reasoning. You asked what if your middle name. Logically if you never told it to me, it’s very obvious to simply say so. The robot seems to give a “does not compute” answer here
The constant blinking and lack of even a hint of a smile doesn’t match its dialogue at all. It could be chucky or Winnie the Pooh. It will have the same monotone melancholy speech pattern as it celebrates your birthday or declares you slated for extermination. “I’m sorry Dave, I’m afraid I can’t do that.” This is not how human evolution works. This is human devolution technology.
It makes sense to me, she can't tell you her favorite name... because tomorrow it might change *brilliant* 👏
she said she would hug a whale with sound... sound reverberates into a distance... she could accomplish sound... sound 'can' touch...
*brilliant* 👏
This is the guy that in the movies is killed by his creation 😅
Can you ask it “what color is the sky at night?” This is a tricky question for an AI chatbot.
lifearchitect.ai/how-do-i-talk-to-gpt/
" We are the champions " is her favorite song ? Ion think we on her team lol,
If the human were actually intelligent it would have empathy. If it had empathy it would not torture sentient beings unneccessarily for nutrition.
WTF... that's the face of Ava from Ex Machina.
I'm not sure if she's conscious (I'm also not sure about some humans either) but I'm sure that she's *_definitely_* intelligent and smart.