|
"On Intelligence" by Jeff Hawkins.
|
|
|
|
|
Nature programmed our brain to remember vagueness (neurons structure). For example you would remember many years from now that you posted this message about memories, but you can't recall what you are actually saying in it. The same as you would remember the vagueness of the winding road, but can't recall specific numbers or faces. Same stuff as our infra-vision.
|
|
|
|
|
This is simply due to evolution: remember routes was critical to survival in the days of yore, remembering phone numbers was not. The hippocampus is particularly important for route-learning.
|
|
|
|
|
From The Insider News
As far as I can tell the state of the art limits the AI to the application and interpretation of a suite of rules. If there is no rule (put in place by a designer/configurator) then your robot can't make the decision to implement a rule/action that has not been defined.
To date there is no creativity in AI, it is completely mechanical based on mathematical formula and weightages.
Once the AI can make the leap outside of the logical decision and make it's own rules then we may need to reconsider.
Caveat - I am working on the periphery of this field so my understanding is limited.
Never underestimate the power of human stupidity
RAH
|
|
|
|
|
I am quite sure that we will never program a true AI in the traditional way. Evolutionary algorithms that configure networks of neural networks may do the trick. The problem is, that evolution only works if you have a large number of candidates. You would need to test them in a complex virtual reality, otherwise the AI would be trained to play Skyrim and of no use in the real world.
The whole process may actually take time on a similar scale as natural evolution does. As if mother nature had not provided us with ways to produce more monkeys than it is good for us.
|
|
|
|
|
CodeWraith wrote:
I am quite sure that we will never program a true AI in the traditional way. |
I agree.. it's certainly not going to be created from silicon. The problem with AI is that to "appear" intelligent it needs to be able to do stupid things, just like the rest of us. It also needs to be able to make a decision out of two bad (undesirable) choices, just like we do.
Now is it bad enough that you let somebody else kick your butts without you trying to do it to each other? Now if we're all talking about the same man, and I think we are... it appears he's got a rather growing collection of our bikes.
modified 31-Aug-21 21:01pm.
|
|
|
|
|
It will also have to learn quite a few things a computer has built in, like math. The AI resides in simulated neural nets and has no access to basic silicon functions. Everybody thinks that an AI will automatically be a math genius, but the real hardware would be busy emulating neural networks.
|
|
|
|
|
Being intelligent it'll always have the option so refuse to do the math.. 
Now I wonder what vices it may acquire during its lifetime? 
Now is it bad enough that you let somebody else kick your butts without you trying to do it to each other? Now if we're all talking about the same man, and I think we are... it appears he's got a rather growing collection of our bikes.
modified 31-Aug-21 21:01pm.
|
|
|
|
|
It will eventually resemble the real thing so much that it becomes indistinguishable. Even if movies are full of such things, I doubt that that really is what we want. We want something that is far smarter than the dumb algorithmic boxes we have, but not sentient. That may be a contradiction in itself.
|
|
|
|
|
I know many humans (so-called natural intelligences) who are also incapable of making decisions without guiding rules. The only difference is the specificty of the rules...
If you have an important point to make, don't try to be subtle or clever. Use a pile driver. Hit the point once. Then come back and hit it again. Then hit it a third time - a tremendous whack.
--Winston Churchill
|
|
|
|
|
Very true, but the worst part is that they have been conditioned to behave that way by everyone from their parents, the teachers and their bosses. Where would we get if everyone started thinking instead of doing what they are told? 
|
|
|
|
|
I used to have a boss like that - an avid skier, we forever hoped he'd ski directly at a tree as he'd have to hit it since he'd not be able to decide if right or left was the better option. 
Bad command or file name. Bad, bad command! Sit! Stay! Staaaay...
AntiTwitter: @DalekDave is now a follower!
|
|
|
|
|
You can throw a ball and usually the dog will choose to fetch it,
but the dog can also choose to bite you - and nobody knows why.
This is a riddle AI has a chance of answering, after all the non-thinking animal is just a finite state automiton.
Coming up with a "thought of the day?" Pretty sure OG's not getting replaced any time soon because there's not a machine that can reach such states.
AI using it's rules can't build something that can't be built and so it also can't imagine it - the rules are fixed so it's simply impossible to imagine the impossible. - humans can.
Sin tack
the any key okay
|
|
|
|
|
Lopatir wrote: This is a riddle AI has a chance of answering, after all the non-thinking animal is just a finite state automiton.
You must have had many dogs. And try to be careful if ever a cat comes along. It would probably answer some questions with you as an example. 
|
|
|
|
|
Humans, well most of us, can use imagination to think outside the rules, AI, in its current state, is rule bound.
Maybe injecting a random variable (generated by a rule  ) may simulate imagination.
Never underestimate the power of human stupidity
RAH
|
|
|
|
|
For the random variable part - I did that once for a fractal generator I wrote. The kind you give a set of lines and it reproduces them, scaled, on each branch for n-levels. I had it make an occasional mistake. The generations that drew maple leaves still drew them - but they were different maple leaves.
But, back to your original post. It gave me a thought. First, I don't count memory as intelligence, but count creativity (else, computers already won). Using existing knowledge to create information/results that are derived from recombination of known's into about-to-be-known's. However, it struck me that perhaps the creation of about-to-be-known's (problem solving) is really analogical. Using a known pattern that solves one problem and using it to solve another problem that, although different, can accept the 'same' solution.
Like invention of flying: it wasn't. Devices were contrived, built upon previous knowledge, and applied to solve the problem. Flying was already there and we saw it. Where does this sit on the AI/neural net/imagination border?
What makes us different and what is the target of AI is This:
Aut inveniam viam aut faciam !
| Ravings en masse^ |
|---|
| "The difference between genius and stupidity is that genius has its limits." - Albert Einstein | | "If you are searching for perfection in others, then you seek disappointment. If you are seek perfection in yourself, then you will find failure." - Balboos HaGadol Mar 2010 |
|
|
|
|
|
Lopatir wrote: You can throw a ball and usually the dog will choose to fetch it, but the dog can also choose to bite you - and nobody knows why
Unless you own a (rescued) greyhound. The poor critters just don't grok 'fetch'. You can throw a ball, and they'll chase it to where it lands.
Then they look at you, and trot back over to see what you'll do next. 
Software Zen: delete this;
|
|
|
|
|
At the moment, AI is mere statistical (Bayesian inference, stochastic analysis, etc) rule base processing programmed by human. I believe we can teach a program to be motivated (self improvement). To make a program truly AI, it must be able to learn. We must initially teach it how to code its own subroutine based on new found statistic confirmation. Which means the program is itself also need to be a compiler. Eventually it will learn to create its own machine level compiler -- integrator process.
|
|
|
|
|
From what I have gathered digging around the internets, changes to the way for File Explorer to deal with folders are stored here. Why it is not stored somewhere in the Windows dir, I have no idea.
Anyway, it is possible that these file could be assisting some malware? If they are good, can I combine the contents? I don't mind having a singleton file, but already I've got 2 of them.
One file has this:
[.ShellClassInfo]
LocalizedResourceName=@%SystemRoot%\system32\shell32.dll,-21799
The other has this:
[.ShellClassInfo]
LocalizedResourceName=@%SystemRoot%\system32\shell32.dll,-21769
IconResource=%SystemRoot%\system32\imageres.dll,-183
Any ideas? thanx
|
|
|
|
|
They are normal. You must have unset the option: Hide protected operating system files...
You can reach this via (from explorer) View, Options, only option, View tab and scroll down a bit and check the option and OK.
They'll go away again.
|
|
|
|
|
what he said , I get them as well, but then, I always have all of my files shown. Big Boy Warranty and all....
Charlie Gilley
<italic>Stuck in a dysfunctional matrix from which I must escape...
"Where liberty dwells, there is my country." B. Franklin, 1783
“They who can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety.” BF, 1759
|
|
|
|
|
I have show all files always turned on (because to do otherwise is annoying) and these content.ie5 files show up everywhere
you can quietly delete them, does no harm at all. Over time they slowly come back, just delete again.
(If you manually sort the start menu in a particular way that's apparently stored in here too, but I'm happy with the default sort-by-name so I remove those files too.)
The one place I do leave this is: ...\temporary internet files\desktop.ini
contents of that desktop.ini are:
[.ShellClassInfo]
-UICLSID={7BD29E00-76C1-11CF-9DD0-00A0C9034933}
and that stops windows explorer hiding the content.ie5 and it's sub folders. (So once in a while I can clean out those too.)
Sin tack
the any key okay
|
|
|
|
|
Just to add, you cannot combine their contents, every "desktop.ini" refers to its parent folder.
For example if you change the folder icons on some folders you'll get a "desktop.ini" file in each of those folders and the files content will contain an information only for their parent folder.
|
|
|
|
|
Just to clarify Mario's answer: The reason you're seeing 2 of these files on your desktop is that there's a version in your desktop folder under your own user profile (c:\users\YourName\desktop), and the other is from the desktop folder under the public user profile (c:\users\public\desktop). What you have on your desktop is the combined content of both desktop folders. So because of this, despite all appearances, it's not violating the "files in a given folder must be uniquely named" rule.
|
|
|
|
|
Whenever I read a really great book I like to share it.
Right now I'm reading, Writing Winning Business Proposals, Third Edition [^]
I really like books (and communicators) who cut through all the clouds and give me the facts in an interesting way.
I know many of you run your own software businesses and this one will be really great for those.
However, I'm finding that the basic structure they talk about with proposals directly relate to discussions I have with various stakeholders within the company I work at when they request enhancements, bug fixes, etc.
I really like the way the authors break the proposal down into what they call baseline logic:
Situation: What is the problem or the opportunity?
Objectives: Given the problem or opportunity, what are the objectives for solving or realizing it?
Methods: Given those objectives, how will you achieve them?
Qualifications: Given those methods, how are you qualified to perform them?
Costs: Given the methods and qualifications, how much will it cost?
Benefits: Given those costs, what benefits and/or value will accrue?
As a developer I love the way this logic flows and connects the entire thing together.
There's even more great stuff. Can't believe someone put this together so well.
Have any of you read this great book?
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
