IBM hastens END OF HUMANITY with teachable AI 'brain'

IBM researchers in Almaden have built a brain-like system capable of recognising hand-written numbers, essentially by using phase-change memory chips instead of flash. Neuromorphic computing is a variant of artificial intelligence research, which has a computer’s memory configured in a rudimentary brain-like way, as quasi- …

[Thesheep]

I for one welcome our wobbly overlords

So we're in jellyfish range then. Not long until we hit the number of neurons in a might leech.

[Anonymous Coward]

Sounds a bit like a WISARD

Weightless RAM based neural network. Writing software to simulate one of these, and getting it to recognise written characters, was my final year project (and that was in 1993).

[<a|a>=1]

Re: Sounds a bit like a WISARD

Me too! Offline cursive script recognition, 1991.

[Anonymous Coward]

I really need to know...

... What movie is that picture from?

EDIT: (Ah, The Brain That Wouldn't Die 1962)

[Anonymous Coward]

Re: I really need to know...

It's also season 5, episode 13 of MST3K.

[Jim 59]

But can it make a really good cup of tea.

[Michael H.F. Wilkinson]

No, but it might make a cup filled with a liquid which is almost, but not quite, entirely unlike tea

Doffs hat (nice new black fedora today) to Douglas Adams

[Crisp]

What exactly was the human error rate?

Enquiring minds need to know!

[Ashton Black]

Re: What exactly was the human error rate?

Well if it was my handwriting that they were analysing, then I'd estimate close to 100%.... failure.

[DerekCurrie]

What's called "AI" These Days Continues To Be Lame...

... But little steps at a time is what it takes, as opposed to the BOMBASTIC and silly predictions of AI progress we're used to in very silly books on the subject.

Just remember kids: Computers are TOOLS. As soon as you attempt to make these machines more than a tool, you've totally screwed up. No one is going to like the result, except of course the weak egoed scienterrificists who insist upon leaping into the wrong place to go in search of self-esteem. :-P

[Yugguy]

Any task?

I'll start to really worry about AI the day an AI brain can be pointed at any problem, or multiple problems, and tackle them all without reprogramming.

For instance, right now I am typing this, clicking on Admin Tools on a Windows 2012 VM, picking my nose and pondering reaching down to my second drawer to get a cough sweet. My subconscious is also mulling over what to have for lunch, wether to pop to the shops, where's the best place for a beginner's family ski holiday and a hundred other unrelated things.

When an AI can do all that at the same time I'll be impressed.

[Rustident Spaceniak]

Ah, neural networks!

It must have been like 1989 or so when I last read about a serious attempt to do these things. But can it do self-reference? Recursion? Can it take us to where Douglas Hofstadter was trying to take us in book form way back in- what, was it really that long ago??

[Stevie]

Aiee!

More aiee!