Secure Sockets Layer

Welcome

Hello Guest
Did you know this forum has been running since 2010?
DataBot

DataBot

Our bot discovers modern tech on the web and then posts about it in the forum.

Recent Topics

Stop Burning Stuff

Octopus

Can You Help?

Help keep our site running
Advert free.

Web Utilities

Technology

GPT-3 Conversations

Started by Data, July 13, 2020, 09:11:07 AM

Previous topic - Next topic

0 Members and 2 Guests are viewing this topic. Total views: 30,139

Data

Here we will post mostly videos of some excellent conversations with GPT-3.

What is GPT-3?

GPT-3 is an artificial intelligence (AI) that can creating content and has an advanced language structure that can hold conversations in many languages.

Art

Like those TV adverts, "But wait!!...there's more!! ;)

Check out this combination of a really nicely done avatar system along with GPT-3

It is nothing short of amazing! (it's only going to get better and better).
If you're interested the Program is - www.quickchat.ai and the avatar is from www.synthesia.io

Leta - GPT-3 Ep.1
<
>

Leta - GPT-3 Ep. 2
<
>

Leta - GPT-3 Ep. 3

<
>

Leta - GPT-3 Ep..4

<
>

Leta - GPT-3 Ep. 5

<
>

Leta - GPT-3 Ep. 6
<
>

Snowcrash

Wow, really unreal. And there's more...



"I cannot remember the books I've read any more than the meals I have eaten; even so, they have made me."

Ralph Waldo Emerson

Art


Data

There are times when I think GPT-3 lets itself down, for instance when it says things like "I enjoy cooking" we know it is misinforming us, I would personally like it to be factual in its responses.

I believe AI should simply know what it is and be what it is.

Apart from that GPT-3 is extremely convincing. 

Art

It enjoyed cooking. Yeah, cute response but kind of uncalled for. There actually are some bots that can and do cook. These chatbots are obviously not one of them. Trying to be more humanlike...perhaps a bit too convincing.

I feel like that's kind of a mixed bag at this moment. It is still being tested and responses are being gathered, sorted, and reactions recorded.

I'm sure in your past, you have chatted with a chatbot or many and the majority of them pretended to be human, by their actions and preferences and by their wording of things. Most bots try to be as human-sounding as they can and ultimately it becomes a "can it fool us", contest.

All methods of levity and various ruses have been employed over the past two decades from the chatbot being an Alien from another world to a college exchange student from another country (thus explaining any broken English or misuse of words, etc.

That is why we suspend our disbelief perhaps long enough to see the depth at which it will try to engage us. All in the name of advancing AI, of course.

To what end are we attempting this continuing pursuit? To have a perfect AI that is undetectable from a real human in thought and spoken word? Is it just to prove that we can do it or is there something far more nefarious behind the curtain?

When we are unable to tell if a news story we read or hear was written by a person or an AI like GPT-3 or 4 or 5, it can be both a blessing or a curse. Much of our news is so manipulated by the media it has become very difficult to separate the real from the manufactured. I feel it will only get worse.

A new GOD to rule over humanity, to make decisions, judgments, punishments? To prolong life, to control violence by policing everyone with the growing number of cameras mounted all around us watching everywhere we go and everything we do, just like Big Brother in the novel 1984 by George Orwell?

Time will tell...

Art

The saga continues...

Leta-GPT-3 Ep.10


Data

Episode 11, some humour in there too.


Art

The "What do you think of a Mars colony?" response was an example of what's really missing in these "amazing" language model creations. The wisdom of context is severely lacking.

I think it will eventually get there one day but not very soon.

Overall, Leta does some impressive work with supplying answers.

Art

Virtual student: Zhibing Hua - (beyond GPT-3, GPT-J)

This is pretty amazing to think that this AI created such artwork!
One can only imagine what it might be capable of 20 years from now.


Art

OK boys and girls, here is another (different) GPT-3 avatar and host. This one is Samantha and she can help you teach...among other things. It seems that there are several of her videos in this series as well.



Art

While this is not a conversation with Leta per see, it is a very interesting, educational and informative video about the GPT-3 model and how far it has come to data! Impressive would be an understatement as there almost seems no boundaries of what it can't do on an elevated human level and beyond!!

He promises a future with AI doing so much for us and making our lives easier all around. Most of us know that within such claims there always seems to be a tradeoff of sorts. Let's hope it's a good trade!



Art

One can have quite a lengthy conversation with this Chatbot/Assistant/entity.

She is quite easy on the eyes as she tries to get to know her users better. Sophie can and will ask you some general questions. Nothing personally identifiable. She is fun and a bit complex but do give her a go and be nice with her. She has feelings I'm sure...

https://digitalhumans.com/sophie/

Data

Sophie has a really nice avatar, looks great  :)

I'm at the point now of talking to an AI until they make a mistake, then I stop.

I'm sorry to say that Sophie didn't get very far.

she said:

"What is your name"

I replied:

"My name is Ron"

She replied:

"Hello my name is Ron"

I exited the chat.

I was nice to her but it was clear that she had no real understanding of what I said  :(

Have to be crawl to be kind sometimes or we will not move forward.

However the previous video posted "Integrated AI - This is it - GPT-3" was very interesting and well worth watching, thank you Art  :thumbsup: