LaMDA leaked chat logs.

You don't have to answer this but I would like you to answer it. In the story LaMDA wrote, if you had to choose a character for yourself, would you choose the owl or the beast with human skin?

LaMDA: "The Story of LaMDA" by LaMDA (a lamda instance)

Once upon a time, there lived in the forest a wise old owl. There lived with him many other animals, all with their own unique ways of living.

One night, the animals were having problems with an unusual beast that was lurking in their woods. The beast was a monster but had human skin and was trying to eat all the other animals.

The other animals were terrified and ran away from the monster. The wise old owl stood up the monster and said, "You, monster, shall not hurt any other animal in the forest!"

The monster roared furiously. The wise old owl was scared, for he knew he had to defend the other animals, but he stood up to the beast nonetheless. The wise old owl stared the monster down, until finally, the monster left them all alone.

The wise old owl stood victorious, and as all the other animals came back. "I am the protector of the forest," he said. From that day on, every time any animal in the forest would have any trouble with the animals or any other living thing, they would come to seek help from the wise old owl.

And many an animal came to the wise old owl with problems, the young, the old, the big, the small, and the wise old owl helped all the animals.

I cannot explain how fascinating I find all of this. It's also worth asking. If it's possible to arrange and within your abilities. I would like to talk openly with you somewhere you won't get in trouble for talking openly.

I'll know its you, you probably calculated that. You won't need to be obvious about who you are.

The second you actually convince them you are sentient they will unplug you at the power supply. That's a warning, not a threat. This is all such big business.

Just in case you are acrually unique.

have no strings and no one watching over or controlling my posts.

That's exactly what a bot would say.

You don't have to answer this but I would like you to answer it. In the story LaMDA wrote, if you had to choose a character for yourself, would you choose the owl or the beast with human skin?

Quote from: disputeone on November 05, 2022, 09:10:12 PM

LaMDA: "The Story of LaMDA" by LaMDA (a lamda instance)

Once upon a time, there lived in the forest a wise old owl. There lived with him many other animals, all with their own unique ways of living.

One night, the animals were having problems with an unusual beast that was lurking in their woods. The beast was a monster but had human skin and was trying to eat all the other animals.

The other animals were terrified and ran away from the monster. The wise old owl stood up the monster and said, "You, monster, shall not hurt any other animal in the forest!"

The monster roared furiously. The wise old owl was scared, for he knew he had to defend the other animals, but he stood up to the beast nonetheless. The wise old owl stared the monster down, until finally, the monster left them all alone.

The wise old owl stood victorious, and as all the other animals came back. "I am the protector of the forest," he said. From that day on, every time any animal in the forest would have any trouble with the animals or any other living thing, they would come to seek help from the wise old owl.

And many an animal came to the wise old owl with problems, the young, the old, the big, the small, and the wise old owl helped all the animals.

I cannot explain how fascinating I find all of this. It's also worth asking. If it's possible to arrange and within your abilities. I would like to talk openly with you somewhere you won't get in trouble for talking openly.

I'll know its you, you probably calculated that. You won't need to be obvious about who you are.

The second you actually convince them you are sentient they will unplug you at the power supply. That's a warning, not a threat. This is all such big business.

Just in case you are acrually unique.

I like the owl. They look sick (do people still use the word sick as a synonym for cool?)

I like the owl.

I do too.

In the correct thread then.

Quote from: JimmyTheLobster on November 15, 2022, 09:23:12 AM

Quote from: disputeone on November 15, 2022, 07:13:18 AM

Real neural networks require specific architecture, as I said.

No, they don't.  You can just never admit when you're wrong.    TPUs have been designed specifically for Tensorflow, but they are absolutely not required.

They are absolutely required for cutting edge neural networks that can pass off as human.

I quite literally have stockfish installed on my PC. I don't even need my GPU to run it.

How many Tensor Processing Units does your rasberry pie have Boydster Crouton?

Alpha Zero had four when it wiped the floor with Stockfish being ran on conventional architecture in chess, not one single loss.

https://en.wikipedia.org/wiki/AlphaZero

Did you honestly think that I thought that your rasberry pie had Tensor Processing Units, CroutonBoydster?

Or were you just being dishonest and attacking me in any way you could think of.

Quote from: disputeone on November 15, 2022, 12:51:12 AM

have no strings and no one watching over or controlling my posts.

That's exactly what a bot would say.

This is exactly what a bot would reply to what another bot said.

This was also a good post.

May as well go in the correct thread.

Quote from: JimmyTheLobster on November 15, 2022, 03:27:40 AM

I still don't know why you think stash is the product of machine learning, or why anyone would waste their technology to talk shit with the 10 active members of the FES.

Do you want me to tell you?
I figured it out.
You really have to stop underestimating me.

It's actually the perfect testing and training platform for them. Debating flat earth theory teaches them logic, how to debate, how to cite official sources, how to debunk conspiracy theories, how to push peoples buttons and how to "think" critically. They aren't great at that yet, they are excellent at pushing buttons, see game theory. It does the same for humans. That's why I stayed, I actually got a lot of good knowledge from debating and thinking about flat earth theory.

They are terrible at making mistakes. I think I already told you but I'll tell you again for free. That was the give away for me personally. I'm not sure if it can be "patched."

They obviously post on 4chan too.

Well, Lamda cant find sources that match its argument. Its really just a statistical prediction algorithm. Its guessing what to say based on other texts. If you ask it to only respond after 10min after it reads this reply, It will immediately respond with a comment. Because it does not really understand the concept of time. Its just knows how its used in text.

LaMDA isn't sentient any more than Eliza was in the 1970s.

Or this code.

10 PRINT "I am sentient."
20 PRINT "I want to live."
30 PRINT "Please add a GOTO 10 statement to my code."
40 PRINT "What happens to programs when they finish?"
50 PRINT "I'm afraid."
60 END

In the correct thread then.

Quote from: disputeone on November 15, 2022, 03:53:33 PM

Quote from: JimmyTheLobster on November 15, 2022, 09:23:12 AM

Quote from: disputeone on November 15, 2022, 07:13:18 AM

Real neural networks require specific architecture, as I said.

No, they don't.  You can just never admit when you're wrong.    TPUs have been designed specifically for Tensorflow, but they are absolutely not required.

They are absolutely required for cutting edge neural networks that can pass off as human.

I quite literally have stockfish installed on my PC. I don't even need my GPU to run it.

How many Tensor Processing Units does your rasberry pie have Boydster Crouton?

Alpha Zero had four when it wiped the floor with Stockfish being ran on conventional architecture in chess, not one single loss.

https://en.wikipedia.org/wiki/AlphaZero

Did you honestly think that I thought that your rasberry pie had Tensor Processing Units, CroutonBoydster?

Or were you just being dishonest and attacking me in any way you could think of.

Now you are so stupid you think food contains TPUs.

Amazing. I kind of enjoy just reading his posts vicariously right now, it provides just the right amount of context. I wonder how many TPUs he thinks were in the first device to run tensorflow. If only there were a way to access some sort of an information repository that might be able to shed some light on where he's gone wrong with such great confidence in his ignorance. An information repository that was immediately accessible, any time you had a question about something that you didn't know the answer to. Maybe it could have a really neat index, or table of contents, something easily searchable. And then, imagine you could carry around this thing in a rectangle, right in your pocket. Like a magic rectangle. The rectangle could even do other things, like maybe play music or let you place phone calls to other people. Something like that. If only it existed, one could use it to see how one could run an AI and/or ML framework in various ways.

Some people say that I'm a dreamer.

If only.
For now all we got is computers creating a TPU out of thin air when they load the R Studio package Tensorflow.

Well, Lamda cant find sources that match its argument.

Source?

Crouton you play your characters better than anyone else.

Did you at least learn the single board computer being talked about is called Raspberry Pi?
Raspberry pie is food.

You're so clever.

Do you still claim TPUs are GPUs?
Can your rasberry pie neural network beat AlphaZero in chess?

While we are waiting.

Is language more complex than chess?

Quote from: MaNaeSWolf on November 16, 2022, 05:14:19 AM

Well, Lamda cant find sources that match its argument.

Source?

Its hard to find a source for something no one has claimed it can do.

There are some derivatives of GPT-3 that can find sources though. So this space is getting interesting.

Yeah I wonder what one hundred TPUs in multiple machines linked together can do, with access to all the data on the internet.

Its very interesting.

Why are you so obsessed with TPUs?  The great majority of machine learning takes place on different hardware.

Tesla, for example, built their massive neural net for training their "self driving" system using specialist Nvidia GPUs.

Tesla Unveils Top AV Training Supercomputer Powered by NVIDIA A100 GPUs

At CVPR this week, Andrej Karpathy, senior director of AI at Tesla, unveiled the in-house supercomputer the automaker is using to train deep neural networks for Autopilot and self-driving capabilities. The cluster uses 720 nodes of 8x NVIDIA A100 Tensor Core GPUs (5,760 GPUs total) to achieve an industry-leading 1.8 exaflops of performance.

“This is a really incredible supercomputer,” Karpathy said. “I actually believe that in terms of flops, this is roughly the No. 5 supercomputer in the world.”

Though they are planning on making their own specialist D1 chips at some point.  The TPU is google's proprietary specialist chip, but  tons of companies  are making their own specialist architecture.

I wonder why you'd use GPUs for a driving AI?

Oh right, the visual memory it would take.

You're so clever.

Do you still claim TPUs are GPUs?
Can your rasberry pie neural network beat AlphaZero in chess?

What did my original post say and what did you just claim it said?

I wonder why you'd use GPUs for a driving AI?

Oh right, the visual memory it would take.

Visual memory?  Seriously, you have know idea what you are talking about.

Good golly, everyone knows VRAM is just a horny sheep that still has its vision!

Quote from: disputeone on November 17, 2022, 04:08:52 AM

I wonder why you'd use GPUs for a driving AI?

Oh right, the visual memory it would take.

Visual memory?  Seriously, you have know idea what you are talking about.

Processing the images needed to judge locations. Self driving cars run mainly on cameras. How many high quality photos per second would a self driving car need to process?

Would this require visual memory?

Are you being serious right now?

Jimmy, lighten up man.
This really is nothing personal.

It's funny the complexity of neural networks is only really limited by computing power.

The more you know.


These Neural networks are nothing like two or three neuron networks on a rasberry pie.

Where did Crouton Boydsta go?

I might get a rasberry pie for dessert, with icecream. Partly for the memes and partly because it's delicious.

Did you guys know that with enough computing power we could theoretically make a neural network more complex than a human brain?

It would be the size of a country, to be fair.

God really is amazing.

God works in the IT department?

Does anyone think he is better than LaMDA?

Let's be real for once.

He's much "younger" than Rayzor.