Thought on AI

[Spectric]

It just trots out modern sharpening dogma and is wrong on several points, if not all of them.

It can only be logical !

It would be interesting to hear how you define intelligence.

That is the part that can be open to interpretation, what is human intelligence and can human intelligence be replicated by logic when we know that many decisions we make are often not logical. So do we then look at human intelligence as being flawed by being more emotion driven than just logical and then look at machine logic as being potentially superior. Maybe a comparison of the two, a baby is born with nothing more than basic functionality that supports life and an empty brain which is both taught and self learns as it grows. A computer has to be built and then programed by a human, it is a digital system that attempts to mimic the analogue world around it and cannot think like a human, it follows the instructions programmed into it although this code could be modified along the way but again using an algorithm that has been programmed in by a human.

If you think about how we function then there is a lot that can be mimicked by computer, thinking of something like reading where we have learnt to reconise the language and patterns to interpret the meaning is already being done with ANPR cameras and more sophisticated facial recognition systems but we get more from reading something than AI which will just see data. The one thing AI does have a massive advantage with is speed, it could read vast amounts of data in the blink of an eye but humans interpret data in different ways so AI for logical outcomes and us for more thoughtful outcomes.

I think that to be classed as inteligent then any form of AI would start of as a system programmed by a human but at some point become self programmed with no original code and able to accept an instruction or task and decide how to complete the task without human intervention and engage in communication regarding the decision process.

 

[selectortone]

A computer has to be built and then programed by a human,

It won't be long before they are building and programming themselves. Then look out.

Edit: designing, building and programming themselves. That's the scary bit.

 

[Ed Weber]

It just trots out modern sharpening dogma and is wrong on several points, if not all of them.
Ditto with flattening stones answers.
Maybe you should also ask it for alternative views so you get a choice?
In the long term the benefits could be huge, but only if they are distributed. Going by past history of technological advance this may not happen without a struggle. https://www.theguardian.com/society...ersal-basic-income-about-to-transform-society

Saying it's "wrong on several points" is not an answer.
It may be a bland, generic answer but nothing is incorrect.

1. Consistency:
2. Efficiency:
3. Minimizing Errors:
4. Precision Work:
5. Physical Strain:

None of the reasons are invalid
It also said "Sharpening guides can still be beneficial" CAN BE, not are but can be beneficial. What's the problem.

 

[Jacob]

...

None of the reasons are invalid
.....

They are for me. And slower, especially in real terms if using the tool on an actual job. A little and often!

 

[Ed Weber]

They are for me. And slower, especially in real terms if using the tool on an actual job. A little and often!

So,;
1. Consistency:
2. Efficiency:
3. Minimizing Errors:
4. Precision Work:
5. Physical Strain:

These are all invalid and hove no place if using the tool on the actual job.
Do you ever give a detailed explanation for the things you post or all they all just broad, knee-jerk responses?

I suppose if you think of it, who would want to be contestant and efficient while reducing physical strain at the same time, how ludicrous does that sound.

 

[John Brown]

I've nothing against sharpening threads, but it would be a shame, in my opinion, if this should evolve into one, as I think it's interesting to read views on AI.

 

[SammyQ]

"AI"? Very poorly named. It is simply a step up from Google in that it (apparently) compares content across multiple sources and paraphrases what it finds. Considering the SISO mantra referred to above, the more S that is posted by wannabee narcissists, the more erroneous the compilation by AI will be. That's NOT intelligence. It's regurgitation without digestion; "a technicolor yawn"? well, may be a "technological" one.

 

[Jacob]

"AI"? Very poorly named. It is simply a step up from Google in that it (apparently) compares content across multiple sources and paraphrases what it finds. Considering the SISO mantra referred to above, the more S that is posted by wannabee narcissists, the more erroneous the compilation by AI will be. That's NOT intelligence. It's regurgitation without digestion; "a technicolor yawn"? well, may be a "technological" one.

I agree - it gives the illusion of ordered thought, mainly by being structured and grammatically correct. But then as it gets refined won't it get nearer to being intelligent, ie. in the sense of making as good a job as any of us, or better, in sorting through ideas and information? it already does this with maths - many thousands of times faster and more accurate than any of us could manage.

 

[TRITON]

I agree - it gives the illusion of ordered thought, mainly by being structured and grammatically correct. But then as it gets refined won't it get nearer to being intelligent, ie. in the sense of making as good a job as any of us, or better, in sorting through ideas and information? it already does this with maths - many thousands of times faster and more accurate than any of us could manage.

Still learning. but more importantly, we've found another semi-intelligent entity, so lets see how we treat it this time around.

 

[John Brown]

While I agree that AI has some way to go, nobody has yet pinned down what special characteristics of "human intelligence" would not, ultimately, be achievable. When I say nobody, I'm not talking about those on this forum. Nobody really knows what consciousness is, or why it evolved. I've yet to read a convincing argument for the existence of free will. Roger Penrose proposed something vague about quantum effects, but I found it dubious.
On the other hand, maybe I'm out of date, maybe things have moved on?

 

[John Brown]

"AI"? Very poorly named. It is simply a step up from Google in that it (apparently) compares content across multiple sources and paraphrases what it finds. Considering the SISO mantra referred to above, the more S that is posted by wannabee narcissists, the more erroneous the compilation by AI will be. That's NOT intelligence. It's regurgitation without digestion; "a technicolor yawn"? well, may be a "technological" one.

Isn't that what humans do? We consume a big bunch of information from somewhere, used to be storytelling,/written matter, now mainly online, filter it and combine it in some way, and pass it on.

 

[SammyQ]

John, you missed out: "compare a new statement to existing information and make a judgement on its veracity".

That's called discerning thought.

 

[paulrbarnard]

Isn't that what humans do? We consume a big bunch of information from somewhere, used to be storytelling,/written matter, now mainly online, filter it and combine it in some way, and pass it on.

I agree it is a similar process. Even leaps of intuition are simply building on sparse bits of information.

 

[Ed Weber]

I've nothing against sharpening threads, but it would be a shame, in my opinion, if this should evolve into one, as I think it's interesting to read views on AI.

Point taken
As far as AI stands today, the consumer stuff we're exposed is fancy autocomplete.
You'll get no more intelligent answers than the material supplied, garbage in, garbage out.
The datasets typically aren't known, so how could I be confident that the generated answer is correct?
If I agree with woodworker A on most all topics and disagree with woodworker B, I have a bias. If they use woodworker B's material for the dataset, the answers will reflect it. Meaning the AI would have a bias toward B, if it's the only information it can draw from.
You really need to know what information is being fed into the searchable database if you're going to be able to rely on the answers it gives.
Simply drawing information from the internet is dangerous, especially these days where everyone thinks they're an expert on everything.
I'm not scared of AI, I'm scared of humans using AI.

 

[John Brown]

John, you missed out: "compare a new statement to existing information and make a judgement on its veracity".

That's called discerning thought.

That sort of comes under the umbrella of filtering, in my opinion.

 

[Jacob]

...... Nobody really knows what consciousness is, or why it evolved.....

Galen Strawson is your man on consciousness, viz "panpsychism" which isn't as odd as it sounds.
Consciousness is easy to understand if you compare it with "unconsciousness". Consciousness is just the brain machine being switched on i.e. "awake".
It evolved to deal with those issues where you need to be conscious and aware of external events. Most of what goes on in the body is totally unconscious - millions of little internal processes and interactions.

 

[Jacob]

Still learning. but more importantly, we've found another semi-intelligent entity, so lets see how we treat it this time around.

What was the semi-intelligent entity first time around? I must have missed it, unless you are thinking of Liz Truss/Boris Johnson & clones.

 

[Terry - Somerset]

Assume both AI and humans have access to the same sets of data - digital searches and (perhaps) film and other recordings of various experts and non- experts providing inputs to a question.

Asked a common question - let's stick with sharpening for the moment - humans may come up with a range of answers based upon not just the data but personal relevant experience.

AI can access and analyse more data more quickly. It can link sharpening data to other peripheral information (metallurgy, chemistry, etc) in a way which most people cannot. But I assume the same AI system using the same data will always give the same answer.

Perhaps what we perceive as intelligence in people is bias and preconceptions at work. Where personal knowledge is extensive (sharpening again) individual bias may be understandable, but it is a personal opinion not a universal truth.

Where the subject matter is far removed from personal experience, differing opinions seem likely to be the product of cultural, social and environmental factors. This is less likely to evidence superior intelligence - more an inability to assimilate large amounts of data and form rational conclusions.

 

[Jacob]

... let's stick with sharpening for the moment - humans may come up with a range of answers based upon not just the data but personal relevant experience.
,,,

Exactly .
The big difference between AI and us on sharpening is that we have to bloody do it!
AI doesn't know what it's talking about, quite literally.

 

[Ed Weber]

If AI sharpened, it would be consistent, repeatable and precise, not ot mention faster. All the things you seem to think are antithetical to freehand sharpening.

Is there anything you do like?