• OK, it's on.
  • Please note that many, many Email Addresses used for spam, are not accepted at registration. Select a respectable Free email.
  • Done now. Domine miserere nobis.

AI took our jerbs!

Cognisant

cackling in the trenches
Local time
Yesterday 8:56 PM
Joined
Dec 12, 2009
Messages
11,155
---




You know this is just the start, it's a business now, there's a lot of money at stake.
 

EndogenousRebel

Even a mean person is trying their best, right?
Local time
Today 1:56 AM
Joined
Jun 13, 2019
Messages
2,252
---
Location
Narnia

For everyone single company that is actually innovating and working towards legitimately replacing humans in the workplace there's probably at least like two-five companies that are just trying to make a quick buck, which are run by opportunists who don't even know what they're doing.

It's almost as if with crypto, whole bunch of fraudsters kind of ruined the perception of crypto before it even got a proper start.

I don't know the specifics it would take to run and process AI systems to a large consumer base, but it's probably a lot and the business model isn't refined yet.

I guess I could have my own personal server at home with my own database of language models and training happening. Kind of like crypto miners, but in this case it's more than just about hoarding GPUs to solve cryptographic puzzles, it seems like it'd be pretty much a big investment of time to do this at a large scale.
 

Black Rose

An unbreakable bond
Local time
Today 12:56 AM
Joined
Apr 4, 2010
Messages
11,431
---
Location
with mama
Even a ten-teraflop computer can run a simple AGI.
I mean it is all about having data to model what the user wants the a.i. to do.
A basic layout of processes and practical application of reinforcement.
To begin with a foundation perception model and then a reasoning model.
It is very easy to break things down into steps/parts and then arrange them into something meaningful.
a.i. will be like the personal computer, everyone will have one with no need for a subscription system.
programing it should be as easy as web development.
because it is just a self-assembling network framework.
 

scorpiomover

The little professor
Local time
Today 7:56 AM
Joined
May 3, 2011
Messages
3,383
---
You know this is just the start, it's a business now, there's a lot of money at stake.
Imagine if someone makes a Terminator. AI could then replace soldiers.

Imagine if someone mounts an AI on a drone with a gun. Then AI could replace the police.

Of course, the big thing that is coming is automated cars. Once automated cars are legal, we can have automated trucks. Then AI can replace truck drivers. I think that's only a matter of time. Then the truck industry might go the same way that the Print industry went in the 90s and 00s.
 

EndogenousRebel

Even a mean person is trying their best, right?
Local time
Today 1:56 AM
Joined
Jun 13, 2019
Messages
2,252
---
Location
Narnia
Even a ten-teraflop computer can run a simple AGI.
I mean it is all about having data to model what the user wants the a.i. to do.
A basic layout of processes and practical application of reinforcement.
To begin with a foundation perception model and then a reasoning model.
It is very easy to break things down into steps/parts and then arrange them into something meaningful.
a.i. will be like the personal computer, everyone will have one with no need for a subscription system.
programing it should be as easy as web development.
because it is just a self-assembling network framework.
I'm not getting away the sauce on this. I'll just say that financial people have no idea how the system works. You probably know more than they do regarding this.

If you sounds confident and have a top tier school degree under your belt a VC will literally give you a bunch of money because it is chump change for them.
 

Black Rose

An unbreakable bond
Local time
Today 12:56 AM
Joined
Apr 4, 2010
Messages
11,431
---
Location
with mama
I'm not getting away the sauce on this.

It is the way we build the model and use that model to ascertain the needs wants and preferences of a user. Then it is how we achieve them in the safest most effective way possible regarding the actual generality of a system.

Generality being the widest, deepest and highest way of weighing options in a temporal setting. | width, depth, breadth, time - possibilities, evaluation. (gaols subgoals - comprehensive step building)

-

Animecat wrote:

If a quantum mechanism is involved it would only increase the options we are allowed to choose from.

The brain imagines what things are like and then updates based on how our actions affect the world.

And of course, it does so in 3D.

3wxG77h.png


Response stimuli​


consequences of a bigger right amygdala:
The amygdala is a reponse meachism.
Our preprogrammed responses to stimuli.
A big amygdala makes it possible to respond to a wide range of stimuli but in a fixed way.

consequences of a bigger anterior cingulate cortex:
The ACC monitors motivation and regulates self-responses strategies.
It is the error detection mechanism of the brain.
It creates new ways of responding to stimuli.
A bigger ACC means having a less wide range of fixed responses and a quicker way to new ones on the fly.

These are both ways of dealing with stimulus responses.
The first is the instant known ways method and the second is the pause and self-filter method.
instant learning vs inhibited learning
Often there is no stronger method in this system. But only a strong polarization between those with fixed and inhibited responses.

1LSA4ah.jpg
 
Top Bottom