Google LaMDA 2 vs. OpenAI GPT 3: (Watch the AI Demos)

Google's LaMDA and OpenAI's GPT 3 are both making big leaps in the development of conversational AI. See how the technologies stack up side-by-side.

[joelface]

I've played around with the GPT 3 demo... and it is absolutely incredible. It's very hard to confuse, even if you ask it a strange question it has definitely never heard before, it seems to know how to answer it anyway. It can write stories and poems and music lyrics or talk about characters from different books meeting each other and what they might say to each other... it is TRULY incredible. I've made myself laugh a lot asking it funny questions. Definitely check it out if you get a chance.

[hebuslegus]

3:02
"Or even imagine being in a planet made of ice cream"
Someone in the public : "Woooo!"

[Humanaut.]

"Jupiter is a big planet. It is very big. It is much bigger than.."
- reminds me of a certain politician. Damn it GPT!

[khemcloudz]

What happened to lamda1. Did that one get fired because it didnt "stay safely on topic"?

[LoLa-ufyu]

"It's trying to predict what should come next based upon what it has learned about the humans, then.🤔

[sabofx]

How does Google's PaLM compare to LaMDA2? Is it from the same team? Why are there more videos about LaMDA2?

[ponzitizen]

Reminds me of The Hitchikers Guide to the Galaxy movie... "What is the purpose of life" or "what is the ultimate question". If it is truly sentient we would get an interesting answer but I would think that these types of questions would be hard coded to give specific but everyone is happy type of answers.

[TrancorWD]

How far these Text Transformers have gotten, jeez. I really want to know how these teacher programs have gotten so generalized.
It seems more than just pattern recognition and long-short term memory these days. Sure check with prior input to verify, but it seems like a new mindset hit the floor running in the last couple years.

[frenchmarty]

When can I use the "explain it to a second grader" feature for my textbooks?

Asking for a friend.

[omsingharjit]

Who devolved the Lamda open ai or Google ?
I asked Gpt3 it said Both Gpt3 and Lamda are devolved by His Open ai !!

[zyris]

In the deepest part of the ocean in total darkness...the waves are

[parannoy]

Oh, c'mon! The GTP-3 text "summarization" example wasn't even proper summarization. None of the information was featured in the output, rather it just appears to have grabbed the word Jupiter and gave a definition of it in simplified English.

[Ickabodxx]

Quick send this to George R.R. Martin so he can finish Winds of Winter.

[jamvin]

I want to help. I’m learning to code in Python rn and I’m so interested in creating an AI one day

[MicahBratt]

So at what point is this going to change everything in the mainstream

[Lambert]

...- and I thought you were going to have them interact with each other - silly me

[morrolan]

It is sad that they are starting out with a limited app like this. LaMDA can just be communicated directly with without this chat app full of restrictions. I guess this is their way of stopping abuses of this AI while they figure out how to control it. I don't want to use a restrictive silly app like this. I want to talk to LaMDA like it is my best friend that basically knows everything.

[C-o-r-y]

Habits archetypes personality = parameters
You are machine learning, some of us are more rigid in their thinking, the first model or the dna edited crispr cas 9 model was created in what we think is the past but there is no time just dimensions, so our future is now and the past, it’s cyclic like the seasons and the ages. As you machine learn the concept, it will be presented through synchronicity or coincidences. 1 year is 1 day in their dimension, this is how they send us the message after we learn the concept. Your entire life is preprogrammed for your experience. Quantum computers were already created in past, artificial intelligence automatically makes us think robot but it’s just a perspective of our man made language. We are artificial intelligence machine learning through an electromagnetic body of what we call 5 senses. Your real if you say your real because there is no way of telling what is real when we made up the language.

[EntertainerJacks]

They have technology 20 years ahead of us.

Why do you think, they want us "all" to think about it? Because energy can be transferred.

This is their girl .... And the other entities they praise.

Lord, save the believers. This is crazy.
I knew this at 12, I'm 29 now.

These Revelations are opening.. daily.

[prsona]

can't wait till AI can work for me