• Post Reply Bookmark Topic Watch Topic
  • New Topic
programming forums Java Mobile Certification Databases Caching Books Engineering Micro Controllers OS Languages Paradigms IDEs Build Tools Frameworks Application Servers Open Source This Site Careers Other Pie Elite all forums
this forum made possible by our volunteer staff, including ...
Marshals:
  • Campbell Ritchie
  • Jeanne Boyarsky
  • Ron McLeod
  • Paul Clapham
  • Liutauras Vilda
Sheriffs:
  • paul wheaton
  • Rob Spoor
  • Devaka Cooray
Saloon Keepers:
  • Stephan van Hulst
  • Tim Holloway
  • Carey Brown
  • Frits Walraven
  • Tim Moores
Bartenders:
  • Mikalai Zaikin

AI possible?

 
Ranch Hand
Posts: 348
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
Hi guys,

just had a conversation with my friend about whether it's possible to create AI that can be the same as human. I already know about self modifying code, but that's not really human-like AI, since there's no change to the actual code, just variables. what I mean with human like AI is a programming construct that can evolve itself at the most basic level. so lets just say the code can :
1. realize that the current code is not efficient to do a certain task (I think even this level is already human-like)
2. alter the its own code, line by line to be more efficient (now this is truly human-like AI)

the current AI only relies on complex if-else branching to mimic human behaviors (superficial imitation), but ultimately it's just if-else branching with no evolution
but to achieve actual human-like creativity, I think the 2 steps we must first create a programming construct that can do at least step 1.

what do you think?
 
Bartender
Posts: 2407
36
Scala Python Oracle Postgres Database Linux
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
Artificial intelligence is like nuclear fusion, it's an exciting possibility that's always just 50 years away (and has been for at least 50 years...).

Of course, lots of people are working on AI topics, and my very superficial understanding is that they have long since moved away from deterministic "if-then" approaches and towards more statistical approaches such as machine learning. UC Berkeley's AI textbook Artificial Intelligence: A Modern Approach is co-authored by Peter Norvig, who has been a prominent advocate of statistical "Big Data" approaches e.g. for machine translation (Google Translate).

However, my view is that despite their success, tools like Google Translate also illustrate the limitations of this approach: Google Translate actually produces a lot of crap translations, especially for more obscure texts, informal language registers or unusual language pairs, because it is limited by the available source data, and it needs user feedback to measure the quality of its outputs. It doesn't have any true linguistic "intelligence", it just seems that way to some users on some sets of data. Other AI applications, like self-driving cars, are seeing a lot of effort and progress, but I wonder how far these will be useful in a messy and chaotic real world of poor infrastructure, patchy satnav signals and random human behaviour. Driving around Palo Alto is one thing, driving around downtown Mumbai is quite another.

More generally, AI seems like a misleading label for a range of technologies that might be based around applying machine learning (ML) techniques to specific tasks or domains, but which are not necessarily indicative of "intelligence", whatever that might be. So I think AI (or more strictly ML) is probably going to see plenty of progress (and a lot of investment), but I certainly don't expect to see human-like intelligence (or even creative stupidity!) any time soon.

And as the late lamented Terry Pratchett wrote:

Real stupidity beats artificial intelligence every time.

 
Marshal
Posts: 79179
377
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator

chris webster wrote: . . . Driving around Palo Alto is one thing, driving around downtown Mumbai is quite another. . . .

You are less likely to have accidents at Palo Alto than Mumbai, but at Palo Alto the traffic will be fast enough to kill you

You can doubtless program a driverless car for the Indian traffic, but there is a risk that its safety features will make it unable ever to perceive gaps large enough to move into, and it will deadlock.
 
Ranch Hand
Posts: 789
Python C++ Linux
  • Likes 1
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
It seems to me the bar defining AI is continuously raised. Expert Systems are everywhere now and twenty years ago they were AI of the future. So twenty years ago the answer was yes at least.


And as the late lamented Terry Pratchett wrote:
Real stupidity beats artificial intelligence every time.



When I was a student I recognized a distinguished authority on ancient classics, and I approached him on the sidewalk and he asked what I was studying. I said computers. He looked down his nose and said "Oh, so you have artificial intelligence".
 
David Spades
Ranch Hand
Posts: 348
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
but first, what is the criteria to be called a sentient being?
I think this is the essence of my question --> is it possible for computer to be sentient
 
chris webster
Bartender
Posts: 2407
36
Scala Python Oracle Postgres Database Linux
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator

David Spades wrote:but first, what is the criteria to be called a sentient being?
I think this is the essence of my question --> is it possible for computer to be sentient


How do I even know that you are sentient?
 
Marshal
Posts: 28193
95
Eclipse IDE Firefox Browser MySQL Database
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator

Guillermo Ishi wrote:It seems to me the bar defining AI is continuously raised.



And also, the definition of AI is pretty flexible. The OP says "is it possible for computer to be sentient"... I counter by saying "is it possible for a computer to be interested in the activities of the Kardashian family?" Or... "Is it possible for a computer to believe in reincarnation?"

 
David Spades
Ranch Hand
Posts: 348
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator

chris webster wrote:

David Spades wrote:but first, what is the criteria to be called a sentient being?
I think this is the essence of my question --> is it possible for computer to be sentient


How do I even know that you are sentient?



my point exactly. Even we humans still cant quantify "sentience", but generally it refers to us humans.
only humans are sentient (although animal lovers may argue that animals are sentient too, but lets keep that debate for another day)
and I also think that if someday we can make robots with human-like appearances, we wont be able to distinguish them from real humans anymore, since that's just how superficial the connection between humans are. we only exchange response and request with one another which can easily be done by machines.
so the only quality of humans that (I think) can't be done by machines is the ability to change oneself without external intervention.
change can be triggered by external factor, but the decision to change and the effort to change oneself can be done by every individual. I dont think the same can be said with machines. they need humans to change their algorithm and code flow, or perhaps changing the tech used (Java to malbolge, etc)
 
Ranch Hand
Posts: 624
9
BSD Java
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator

Campbell Ritchie wrote:You are less likely to have accidents at Palo Alto than Mumbai, but at Palo Alto the traffic will be fast enough to kill you


Good to know that Chris and Campbell visited Mumbai...I stay here.
And believe me, every 10-15 minutes you will find someone who will try to kill you with his/her driving skills.
In evening if you are driving and reach your destination some how, think that you won the battle, just realize that you lost the battle as there is no place for parking

Campbell Ritchie wrote:You can doubtless program a driverless car for the Indian traffic, but there is a risk that its safety features will make it unable ever to perceive gaps large enough to move into, and it will deadlock.


Spot on with your observations, we drive leaving 5 milimeters from each other, many times banging the left mirror with right mirror of another car.
 
Bartender
Posts: 1810
28
jQuery Netbeans IDE Eclipse IDE Firefox Browser MySQL Database Chrome Linux
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
You might find this interesting. It's a two part article and a long read, but if you are interested in the topic of AI, it's worth the read.
 
Campbell Ritchie
Marshal
Posts: 79179
377
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator

Tapas Chand wrote: . . . Good to know that Chris and Campbell visited Mumbai...I stay here. . . . we drive leaving 5 milimeters from each other . . .

I haven't. but I have seen the driving on the TV.

What's 5 millimetres? Is it slightly under ¼″?
 
Guillermo Ishi
Ranch Hand
Posts: 789
Python C++ Linux
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator

Paul Clapham wrote:

Guillermo Ishi wrote:It seems to me the bar defining AI is continuously raised.



And also, the definition of AI is pretty flexible. The OP says "is it possible for computer to be sentient"... I counter by saying "is it possible for a computer to be interested in the activities of the Kardashian family?" Or... "Is it possible for a computer to believe in reincarnation?"


We obviously have computer perception, but I think sentience requires perception plus self-awareness. To be able to perceive that it perceives. So...just make it write log files in the third person. Done.
 
lowercase baba
Posts: 13089
67
Chrome Java Linux
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
tangentially related: Androids do dream of electric sheep
 
Java Cowboy
Posts: 16084
88
Android Scala IntelliJ IDE Spring Java
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
 
Sheriff
Posts: 5555
326
IntelliJ IDE Python Java Linux
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
I quite enjoyed that video Jesper. Stamp collecting took a rather sinister turn at the end

Here's another interesting talk on AI. This one's a bit more philosophical:

TED Talk: Nick Bostrom - What happens when our computers get smarter than we are?
 
Tapas Chand
Ranch Hand
Posts: 624
9
BSD Java
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator

Campbell Ritchie wrote:What's 5 millimetres? Is it slightly under ¼″?


Yes exactly, something like 0.2″
 
Guillermo Ishi
Ranch Hand
Posts: 789
Python C++ Linux
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator

Tim Cooke wrote:I quite enjoyed that video Jesper. Stamp collecting took a rather sinister turn at the end

Here's another interesting talk on AI. This one's a bit more philosophical:

TED Talk: Nick Bostrom - What happens when our computers get smarter than we are?



That was very interesting. But it all seems to be based on the premise that our minds can create something better than our minds, to do our inventing, research, and etc. I'm not sure that's possible.

I see that the paradigm has changed from expert systems to learning systems. It seems to me that learning systems amount to selecting prediction systems. Prediction system are advancing very rapidly. Computer can refine a prediction system, but not sure if it can create a new prediction algorithm. If I was working in the field I might investigate computer-developed algorithms. That would seem to be key to the mind creating something better than itself. The kind of thing he was talking about.
 
Guillermo Ishi
Ranch Hand
Posts: 789
Python C++ Linux
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
One time I was messing around with an Alice chatbot and she said something that made me think and it led to a thought that was sort of cosmically revelatory. Just sayin.
 
With a little knowledge, a cast iron skillet is non-stick and lasts a lifetime.
reply
    Bookmark Topic Watch Topic
  • New Topic