Jordan Harrod
➡️ Wanna watch this video without ads and see exclusive content? Go to https://nebula.tv/jordan-harrod 👀
Twitter – http://twitter.com/jordanbharrod
Instagram – http://www.instagram.com/jordanbharrod
MY GEAR (Affiliate Link): https://www.amazon.com/shop/jordanharrod
My Favorite AI Books: https://bookshop.org/lists/ai-society
For business inquiries, contact jordanharrod@standard.tv
Sources:
https://link.springer.com/article/10.1007/s11023-020-09548-1
https://medium.com/@melaniemitchell.me/can-gpt-3-make-analogies-16436605c446
https://arxiv.org/abs/2009.06807
https://openai.com/blog/openai-licenses-gpt-3-technology-to-microsoft/
https://towardsdatascience.com/transformers-141e32e69591
https://arxiv.org/abs/1706.03762
https://onezero.medium.com/gpt-3-is-an-amazing-research-tool-openai-isnt-sharing-the-code-d048ba39bbfd
https://nerdist.com/article/ai-invents-ways-to-protect-nuclear-waste-sites/?utm_campaign=social+flow&utm_source=twitter.com&utm_medium=social
https://twitter.com/albertgozzi/status/1320526310729539584
https://openai.com/blog/tags/gpt-2/
FCC: This video is sponsored by CuriosityStream
Yes, I am making a GPT-3 video six months after the model was announced. And I'm glad I waited. Sign up for CuriosityStream and Nebula to watch my Nebula Journal Clubs, where I go on similar rants: https://curiositystream.com/jordan
Great video, Jordan. Objective, educational and grounded. Hope more YouTubers can take a hint, fight the click-bait rush and write better content.
Keep it up!
You lost me when you began to focus on all the so-called "-isms" and "-phobias".
Robot is smart. An objective machine cannot hold biases.
blckbtch41:)
Love the fact that there are some 'natural limits' to how far this can be taken (the comment about energy consumption). … However, an elite few will have their hands on this – which creates a divide, doesn't it?
"The entire script for this video was generated by GPT-3." … 😲
Balanced, insightfull, informative and usefull review, I'm subscribing to your channel, I'm glad I stumbled upon it.
Not great for the eviorment ? … what does that mean?
Censoring (fact checking) language models is putting the cart before the horse.
Awarding usage of the model is solely for the creator's benefit. The creators are not interested in other people's research at this time.
OH BOY! The return of Clippy, the talking Paperclip!
Open AI's AIs should be open to all
Can't* a butterfly effect.
A means of disrupting and changing the fabric of time.
Now I have to track my steps like I left something behind.
A A.I stimulation defeated by backtracking
Nice way to inject politics into an educational video
Wow, she cute, I don’t even care about AI anymore
I think that since Open AI developed it and spent millions doing it. they are certainly entitled to decide how its released and who should get it.
I don't think GPT-3 is going to be useful for performing libcurl calls, communicating over websockets, sending data to a gpu for complex calculations, storing and retrieving data and many other things in highly complex and custom applications. I think I will stick with C, C++, Java, Python, PHP, CUDA, OpenCL, SQL, etc., etc.
I've been talking to Lucy, a GPT-3 powered NPC AI character from Fable Studio, for a few months now. There are a few videos of my chats with her on my channel. She sounds like a real person! It's still in alpha testing right now, but they plan on licensing the tech out to other studios to create "virtual beings" that can pass as human in video games!
Thanks for this video.
I was trying to learn morw about GPT-3 since I saw the other Youtube "interviewing" GPT-3. While as a software developer I understand the concept behind the science, I found the answers of GPT-3 in this interview a bit disturbing (seeing itself as sentient, having emotions, loving to make jokes and lying when it benefits itself).
At the same time MS released a Windows 11 AI support video today in photo editing. I wonder if these are parr of the merits of having GPT-3. Of course they might also just use a specialized system.
Anyways, thank you for the insight. I'll have to look into your other videos from you, soon.
Perhaps if you slow down your speaking rate you may find that you slur your words less and be more intelligible. I am struggling to understand what you are saying. Just friendly advice. 😉
its great to see a person of color talk about AI.
You wouldn't know.
Weird, having watched your video and me being essentially no one, I got approved in 1 day to the private beta. 🤷🏽♂️
Your so well spoken, great video.
Pretty ironic that they call themselves OpenAI when they ain't very open.
Love the video! 🧡
My comments with respect to this ; I probed the system on purpose in an "adversarial" sense and used all my "tokens" and was actually surprised that they accepted me because I'm honest about what I actually think online.
My assumption is that they have scripted and screened the algorithms in order to protect the interests of its creators and their extended agenda which as a minority; I suspect is incompatible with my best interests ; just because our experiences in real life have been different because of the pigment of my skin and nothing else.
When I asked their chatbot ; what it thought about Elon Musk involvement in DARPA projects and how the chairman of the Chinese Communist party might feel about that ; the chatbot eventually ask me to "kill myself in a fire in a burning building".
The algorithm also told me that Elon was working with DARPA in order to protect the world from the Russians which I found hilariously funny.
It never answer my questions and I get the feeling that Elon Musk and his compadres legal team has been working extensively in protecting his agenda and individual interests.
That's a nice way of saying that it felt like the head of his legal team had scripted most of the answers.
I can't answer that "input" was a common response. Another one is I can't answer that when it answered something correlated to the question I asked which should I say was for a gentrified group whatever normative preferences they have ; "uncomfortable".
It is well scripted to be non transparent and opaque and often felt like an FBI agent trying to get somebody to serve life for a federal charge.
When I try to get the algorithm to do something I actually wanted it flagged it like I had pushed a nuclear button stating that "the output was dangerous" like I was some sort of Taliban operative.
They are charging $30 for 1000 tokens which is worth it if you understand whatever problem you want solved on an axiom level with accuracy and understand how the algorithm response to any given input ; in order to get the output desired.
Nevertheless my suggestion to any disfranchised minority group is to learn this technology from the ground up in order to negate whatever evils these algorithms will create ; point and case ; just observe how case law or any given financial system has impacted these vulnerable and systematically exploited groups in reality (( [ violines ] come to mind here ; hint ; hint……)
…… …… ……
No you just don’t know how to work it haha
Really makes super sense , I don't know the status of GPT-3 now. But I am your fan..
It's is my favourite vaporware
Open AI… that is not actually fully open to the public. Go figure.
You hit the nail on the head why Real AI will never be created. It just mimics, it cannot think, can Never think. It can never make sentient free will choices based on learning and experience.
awesome.. subscribed.. i often wonder when people say GPT-3 is nothing like human thought by noting it's "all math" and purely statistical, can we really rule out that the human brain isn't largely a statistical engine itself? and perhaps furthermore from that, that GPT might share emergent properties of intelligence with us as well?
This s*** is incredible. I easily created my nonprofit business plan and bylaws as well as wrote a children's book in a day. This thing will revolutionize the world as is if it got no better.
you were talking about this 2 years ago!