Re: apple intelligence not so intelligent

Your Name <YourName@YourISP.com> wrote:

On 2025-01-16 22:18:10 +0000, Rick said:
 

On 1/16/2025 4:15 PM, Your Name wrote:

On 2025-01-16 16:02:46 +0000, badgolferman said:
 

Apple has come under intense scrutiny for rolling out an underbaked
AI-powered feature that summarizes breaking news — while often
butchering it beyond recognition.

<snip>
 
So, no different to any other idiotic AI nonsense. They should all be
avoided by any sane person because they're all just useless, over-hyped
crap that they are. Hopefully it will be just another quickly gone tech
fad.
 

 
AI is nothing more than software, which means that, like any other
software program, it is only as good as it the way it is programmed.

 
And hence why so-called "Artifical Intelligence" is not actually
intelligence of any kind. It's simpl;y a computer program doing what it
has been programmed to do.

I disagree. There are many features and outputs of AI that aren't or
weren't predictable based on the inputs you give them. Just look at some of
the errors that have happened with driverless cars.

It does NOT "learn"

It definitely does.

and it does NOT "think",
and it can never do either of those things.

I used to believe that too. Not anymore.

It'll be a long time before an AI will compete with humans, but elementary
thinking won't be that far away I suspect.

 
 

I agree that a lot of AI isn't very good (and Apple is clearly on that
list), but from the limited testing I have done, I think certain more
mature LLM products like ChatGPT and Copilot are actually quite useful
when used properly.

 
Not really. ChatGPT has numerous issues as well.
 
When it comes to the uselessness of "AI, you just have to look at the
absymal images that it creates - people with three hands, buildings
floating in mid-air, ...

That's old news. The cutting edge is almost indistinguishable. You need a
trained (human) eye to spot the very best.

The really scary part is that morons are trusting this garbage to do
important work like medical diagnosis and supposed self-driving cars!! 
:-(

Properly trained, and importantly properly used, ML/AI systems can be very
useful.

Researchers have to keep being reminded to not believe the hype, but focus
on the evidence.