AI and creativity.

[cookie2]

Whatever it takes?

[LP]

The fears over AI are greatly exaggerated, it's basically software, and currently it requires a high bandwidth 24x7 power intensive connection to the web to work, and because the web is full of shizen AI gets about 40% of it's answers wrong.

The AI we are exposed to like ChatGPT or OpenAI pretty much requires a server farm the size of the MCG to deliver 40% wrong answers in the timeframe people expect from a Google search.

The real problem is dumb-ar5e people believing the AI is always right without question, or nefarious types find ways to us AI to manipulate or con others. Either way the problem is people, not a hunk of silicon or a few lines of code.

Finally, the best of the best technology fails with even rudimentary human intervention. For example AI Drones can fall out of the sky or run into buildings just from someone turning the lights on or off, because machine vision isn't all it's cracked up to be, and it's not really intelligent in the way a person knows all trees are trees just from seeing just one up close.

[Thryleon]

Im skeptical about AI, because ultimatey, machines don't learn without good data being fed to it.

Anyone who has ever had to look at a web application firewall can see that they are bloody hopeless and start blocking legitimate traffic for stupid reasons. 

[kruddler]

You guys are talking about AI as it currently exists.

I'm talking about AI in the future, how far into the future is anybodys guess and its almost impossible to fathom.
Just like 100 years ago, it was impossible to fathom a contraption that could be used as a calculator doing basic mathematical operation any pre-schooler can do. When it eventually came to be, it was as big as a room. Now everyone carries that, and 100 other inventions capable of so much more, in their pocket.

What does 100 years from now look like??

30 years ago the thought of self driving cars was nothing more than sci fi nonsense. Now it exists.

Chat GPT is like the first fish that every crawled out of the water and onto land.
AI i'm talking about is humans taking over the planet.
There are a lot of steps from then to now, but it will happen. Time being the only unknown.

[LP]

I think we must terminate this debate, where is [member=26]madbluboy[/member]‍ when you need him?

I think the idea of an all seeing all conquering Cognisant AI is greatly overstated, and if we create something with such a high level of intelligence then it's unlikely to unilaterally turn on it's creator. Not all AI if it ever exists is going to become Norman Bates!

We can imagine a lot of stuff, most of it will never be real!

A bigger issue is poor quality robotic automation, and a lot of what the media refers to as AI is mostly robotic automation, and it's not intelligent in any way! Intelligent in it's design but not encompassing intelligence.

Pretty soon Windows 11 users will be fully exposed to MS CoPilot and other similar "apparently smart" technologies, but they are just forms of robotic automation based on LLM and not AI at all! LLM is basically the data on the internet reduced to a library of relationships that can be searched for an answer, it's not that smart and most of the answers it will send right or wrong are parroted replies originally authored by humans before being repeated or rehashed by the LLM robotics. None of it can pass a Turing Test. It's rote learning on a massive scale, not understanding.

[kruddler]

I think we must terminate this debate, where is [member=26]madbluboy[/member]‍ when you need him?

I think the idea of an all seeing all conquering Cognisant AI is greatly overstated, and if we create something with such a high level of intelligence then it's unlikely to unilaterally turn on it's creator. Not all AI if it ever exists is going to become Norman Bates!

We can imagine a lot of stuff, most of it will never be real!

A bigger issue is poor quality robotic automation, and a lot of what the media refers to as AI is mostly robotic automation, and it's not intelligent in any way! Intelligent in it's design but not encompassing intelligence.

Pretty soon Windows 11 users will be fully exposed to MS CoPilot and other similar "apparently smart" technologies, but they are just forms of robotic automation based on LLM and not AI at all! LLM is basically the data on the internet reduced to a library of relationships that can be searched for an answer, it's not that smart and most of the answers it will send right or wrong are parroted replies originally authored by humans before being repeated or rehashed by the LLM robotics. None of it can pass a Turing Test. It's rote learning on a massive scale, not understanding.

Mate, you are still thinking 5 minutes down the road. You need to think decades into the future.

We are a lot closer than you think.
All it will take is some crazy Elon Musk style billionaire to go rogue and things get pretty tricky pretty quickly.

Nano-bots can do many weird and wonderful things and are expected to be a big player in our lifetime.
AI is already streets ahead of us in terms of some kinds of diagnoses.
Put the 2 together and we have AI finding issues and fixing us (for a small fee of course) in the not too distant future.

"The first ultraintelligent machine, is the last thing man will ever need to make."

Once this is achieved, it can design and improve upon itself quicker than we can keep up and its about then that we are just along for the ride.

We can already see some of this stuff occuring in our lifetimes.
The movie hidden figures about the african american women who were doing the calculations for NASA back in the day....us as humans, couldn't do the same thing now because we use computers to do it for us.
Logarithm slide rules etc, that had all come and gone before my time. Thats a computer thing now.
The better computers get at this stuff, the less we need to know it. Humanity, is largely getting dumber thanks to our current technologies.
We already have social media sites reading our minds and predicting our wants and needs as well as feeding us (mis)information changing how we vote without us even realising it.
As i mentioned earlier, our cars are starting to drive themselves, and it won't be long before we don't know how to drive ourselves.

The more technology does for us, the less we can do it for ourselves.
Soon that same technology will be better at everything we can do for ourselves.

Some of the smartest people in the world, Hawking,  Musk etc can see this as a potentially catastrophic threat, and i certainly don't see any reason to not agree with them.

Again, its just a matter of when....not if.

Individually, we may be long gone before we even get close.

[Lods]

If I'm not...I'll  be on their side.  

[kruddler]

If I'm not...I'll  be on their side.  

A nice gesture, but they wouldn't need you.

[tonyo]

Here is an AI version of Johnny Cash singing 'Barbie Girl'


https://www.youtube.com/watch?v=HW1qjsqS1zM

[PaulP]

Here is an AI version of Johnny Cash singing 'Barbie Girl'


https://www.youtube.com/watch?v=HW1qjsqS1zM

I don't know whether to laugh or cry. The irony is good for a giggle (although I suspect the joke will wear thin pretty quickly), but come on ? Johnny Cash ? There really is nothing sacred left in this world.