AI fails

Page may contain affiliate links. Please see terms for details.

C R

Legendary Member
The AI agents used by search engines are "stateless" in a sense. They only read from the model, they don't write to it. What you saw is the expected designed behaviour.
 
OP
OP
briantrumpet

briantrumpet

Shaman
The AI agents used by search engines are "stateless" in a sense. They only read from the model, they don't write to it. What you saw is the expected designed behaviour.

Rather undermines the 'intelligence' bit: a central part of human intelligence is the feedback loop that constantly refines the model of the world that our brains rely on to steer us through a chaotic world.
 
  • Like
Reactions: C R
OP
OP
briantrumpet

briantrumpet

Shaman
You don't need to convince me.

I can't remember which brain book it was that I read that is very persuasive about this central function of the human brain (not just in a generalised philosophical sense, but in light of latest neuroscience research) - but it mirrors the way that each of our brains builds a kind of 'predictive text' model of lanaguage that makes speaking & comprehending (aurally and in writing) practicable. It's constantly measuring what's expected and what actually happening, and where anomalies occur, trying to make sense of them, and either incorporating modifications to the model (or dismissing them as 'noise'). And it's all happening without conscious thought.
 
  • Like
Reactions: C R

C R

Legendary Member
I can't remember which brain book it was that I read that is very persuasive about this central function of the human brain (not just in a generalised philosophical sense, but in light of latest neuroscience research) - but it mirrors the way that each of our brains builds a kind of 'predictive text' model of lanaguage that makes speaking & comprehending (aurally and in writing) practicable. It's constantly measuring what's expected and what actually happening, and where anomalies occur, trying to make sense of them, and either incorporating modifications to the model (or dismissing them as 'noise'). And it's all happening without conscious thought.

It is even more surprising that it works so well for people like me who don't think in words.
 
OP
OP
briantrumpet

briantrumpet

Shaman
It is even more surprising that it works so well for people like me who don't think in words.

Yebbut words are the things that attach themselves to concepts which exist in parallel. Babies are amassing understanding of loads of things (people, food, gravity, movement, etc) well before they have language with which to express their grasp/understanding of concepts.
 
OP
OP
briantrumpet

briantrumpet

Shaman
Yebbut words are the things that attach themselves to concepts which exist in parallel. Babies are amassing understanding of loads of things (people, food, gravity, movement, etc) well before they have language with which to express their grasp/understanding of concepts.

And of course, humans existed before what we think of as 'human language' had developed.
 

Ian H

Shaman
It is even more surprising that it works so well for people like me who don't think in words.

Here you go: https://lingojam.com/HieroglyphicsTranslator
 
  • Laugh
Reactions: C R

PurplePenguin

Well-Known Member
k513o9ch2r8g1.jpg


Saw a Reddit discussion about this and what the names are. A few people got it right (I assume). Copilot nailed it in seconds and comprehensively justified its workings. It's definitely better at reading hand written scrawls than me.
 
OP
OP
briantrumpet

briantrumpet

Shaman
Yes, and despite Co-pilot's lengthy explanations, I still struggle to see a g and not a p.

Took me a moment or two, but it's not totally unlike my mum's g.

No AI would ever have read my gran's handwriting - it was a thing of strange beauty and total illegibility. My dad's mum used to say "I do like her handwriting, but I can't make out a word".

She should have been left-handed, but was beaten if she tried, and was forced to learn with her right hand.
 
Top Bottom