tjwoods's shoutbox - Help.com

Hello TJ,
I just noticed you where a fan of mine. I read your profile and some of your replies and liked what I saw so, I friended you back. Thats out of charecter for me. Normally I ask 2 questions first.

1. Why did you friend me?
2. What do you value most?

The first is the same for everyone and the second is individual.

not only that, and i was about to practically tell the story of my life!

ya so basically from ur replies in postland i now think u could easily read my mind.

It’s definitely something I’m going to look into, because I think it holds promise.

Of course, but that is why you have to really TEACH a macro. You must teach with appropriate specificity.

It would be a problem if we had a brain like Paris Hilton’s instead of Albert Einstein’s. However, if a body is included, I might be inclined to forgive such a result. From what I’ve seen so far, I think developing a weak artificially intelligent agent (an agent that can’t derive conclusions or actions for itself, but does demonstrate some facets of intelligence) will be the first major step in developing a strong artificially intelligent agent (an agent that can derive conclusions and determine actions for itself). I think the most interesting area of weak a.i. is in macros. Which I mean as defining functions for computers to perform. In essence, teaching computers what to do. I was thinking last night, and I envisioned a operating system where you have a macro function, and the computer can observe (record) what you do, and you can assign a name to what it observes and it can repeat it when called by that name. For example, if I want to do an Internet search:

1) We assume (know) the computer is on.
2) We assume (know) the file explorer of the operating system is running.
3) We assume (know) that we have two input devices, a mouse and a keyboard.
4) I navigate to the icon representing my Internet browser and double-click it, executing my Internet browser.
5) Using my mouse, I can click in the address bar, preparing it for input.
6) Using my keyboard, I can enter the name of the search engine I want to use (in this example it’s Google) and hit the return key to navigate to that page.
7) Using my mouse, I can navigate to the text box to input my query, and using my keyboard, I can hit return to execute my search.
8) I can either leave my actions as is to review the possible sites for my search, or I can make one further assumption and assume that the first result will be the one I desire, in which case I can use my mouse to click the first result, navigating to that site and finishing my search.

It’s a relatively simple task that takes only a couple of seconds to complete, but upon inspection, one can see there are actually several steps involved, and there are actions going on underneath our abstraction.

At this point, you may bring up facts about the process in relation to the computer, such as the fact that the computer doesn’t have to emulate mouse navigation to and activation of a process to execute it. There is a way for the computer to call processes. There are steps that can be done more efficiently when considered from a computer’s viewpoint than our viewpoint.

That’s why I envision a review period where you can review the computer’s actions, and modify, edit, and delete them to result in a more efficient macro.

Applying the process appropriately to the above steps should result in the following abstraction:

1) Execute the default browser.
2) Enter http://www.google.com into the address bar and emulate hitting the enter key.
3) Enter the user’s search term (which can be considered a function argument) into the search text box, and emulate hitting the enter key.
4) Open the first Web site.

Finally, let’s call are macro Googlesearch, with a string argument, maybe szUserQuery.

And let’s assume that the OS can catch what the user types at all times. And let’s say that a very specific combination of user keyboard input can execute a macro. Let’s say it is in fact the specific word, “macro”. (I know that wouldn’t work, because if you were just talking about a macro, that would create undesired behavior. But this is just a description of the general idea.) So if at anytime a user wanted to search for something, he could just type “macro Googlesearch (artificial intelligence)” and the Web browser would open displaying the following site:

http://en.wikipedia.org/wiki/Artifici…

I think this would be a great help to users, and could be applied in other fields as well. Microsoft already has a task scheduler function that can automate tasks such as defragmentation of hard disks and other forms of maintenance or run processes at certain times. The more work we automate, the quicker we can get onto innovation because we don’t have to focus on the tedium and labor of tasks that CAN be handled by macros.

There are weaknesses. Macros depend on a well defined environment to function. Individual macros cannot think for themselves. That means they cannot do original work.

That’s my idea for the moment.

There are just so many roadblocks, and a breakthrough in one would lead to breakthroughs in others, but it’s hard to make a breakthrough in any. For instance, an artificially intelligent research agent could decrease the time required to breakthrough in engineering, but a breakthrough in engineering could help make an a.i. The question for all researchers is, “Which area do we attempt to make the most advancement in first?”.

Too bad we just can’t reverse engineer the brain as easily as we can reverse engineer other systems.

Of course, no one says that there can’t be other types of artifacts used for computing. It might be possible to have biological computers.

So a.i. is already bootstrapping itself to make compilers that can code it better? If what you say is true, that is very interesting. Of course, the major breakthrough in performance is when humans, or humans and an a.i. finish researching how to build computer hardware using micro engineering technology.

About assembly, I think it may be necessary to use assembly in artificial intelligence programming one day. I haven’t had a chance to verify this fact, but I’ve heard that programs written in optimized, well-used assembly language outperform the same program written in a higher level language. This will be important to an artificial intelligence for performance reasons. We normally don’t use all of the processor power and memory of our personal computers at one time, but some tasks an artificially intelligent agent would perform (such as intractable tasks) will be intense. That’s where every bit of performance we can squeeze out will be important. In addition, multitasking remains a problem. Computer processors cannot multitask at the level of the human brain currently. So the first truly artificially intelligent agent will absolutely depend on being programmed well.

I have random thoughts, and here is one I wanted your feedback and insights on. Would a operating system co-developed with an a.i., with both specifically designed to integrate with each other, be more efficient than a a.i. application developed for an existing OS, like Windows?

These patients suck the life out of everyone in their life. I feel sorry for you tjwoods, I mean that sincerely.

g2g

most people here do not understand the in-house treatment for patients with obsessive disorders. Maybe she should be in an inpatient setting, she will not stay healthy minded though if she doesn’t want to. If she can find places like this to feed the attention aspect she will become more complicated.

Harsh is what someone should have been with her a long time ago and this would not have ballooned into a self defeating attention obsession.

I will use that information to revise the order in which I learn programming languages, but I will not decide which of the existing languages - if any - I would use to program an a.i. agent. I think it is a prudent course of action to research what an artificially intelligent agent would do, and write out the flowchart, pseudo code, etc. to plan the operation of such an agent. Once one sees what one needs to accomplish, the step of choosing a programming language to implement ones design goes from nearly impossible to relatively easy. And it may well be that no existing language is sufficient to program a high-quality a.i. For example, I’d imagine that a language that provides the security of other languages, the power and efficiency of C, and the functionality of languages such as Python or Lisp to be quite useful. Perhaps there are even other tasks that you or I haven’t thought of that would best be implemented in a new language. From my studies and thoughts so far, I’ve thought that a language that could create and execute code at run-time well would be useful, along with the ability to modify and save source code to be compiled and run later, or even code to be interpreted later. I was conversing with a friend via MSN, and he told me that he didn’t think that a language extremely capable in those tasks existed. As I continue my studies of computer programming languages, that is just one of the features I’m going to keep my eyes open for. I’ve already learned about the new and delete functions of C++, which are used to dynamically allocate memory, and that is a part of self-modification, etc.

Since I also mentioned that I was a hobbyist programmer too, I should tell you my primary language. Right now I’m learning about C++. After that I want to learn GUI programming, then assembly, then C, then Java. I expect that I will continue to study other languages after that, and I even entertain aspirations of creating my own programming language with an associated compiler and an interpreter too.

A very general area, but primarily the study, planning, and development of an a.i. application that could bootstrap itself to a higher level of capability and intelligence. Of course, when I mean general, I mean broad, because there are many topics relating to what I mentioned in the previous sentence. Neural networks, programs that modify their code in the manner of genetic mutation, learning, etc.

Hey. I noticed on your profile that you studied machine-learning. I’m study general a.i. and computer programming. I’m not a professional or a student, just a hobbyist who does it for fun and personal enrichment.

:D good to hear, what r u gong to school for?

Thank you for taking over, i had to leave :/ how are you BTW?

awesome advice youve been giving.

i added you to my friends list.

We all like a bit of validation once in a while. Your welcome.

I have never seen you here before. I like the way you write. Thank you for talking with bookworm. Sincerely smoogie. Simone.

Hi, I really like the advice you have given me. I’m the one with panic attacks and excessive worrying. I’m having a hard time helping myself. I’d appreciate some more advice on how to help myself if you have some more. you’ve been helpful I just need lots of help LOL thanks

hello, i read ur reply on how to relax more, that one should know and set up his priority in life. that was really inspiring and an eye-opener for me. thanks for that. but how does one know his priority in life, what if he lacks “goal” in life?

Merry Christmas :)

Hey thanks for the help! : )
Very nice of you to investigate my personality first. Haha. Happy Holidays.

Hi, I just wanted to say the comments you have just posted to the person starting antidepressants were brilliant. Very uplifting and thought provoking.

read your profile. I was working at a Uni in oz in computer vision and was getting into AI. Had a freak out for a bit, but your writing reminded me of how much I use to enjoy it. Cheers.

tjwoods wrote:
Right. As the statistician George Box said “All models are wrong. Some models are useful.”

oh man, a triple bam! signing up now.

Your replies reveal your intelligence, education, and fabulous communication skills.

hi, thank you soo much
Your advice has really left an impact on me…
appreciate it !!
Cheers