[Don Patrick]

A couple of reasons why I don't collaborate:
- My program is 17000 lines of code, and some of it is patchwork. It would take a year just to get someone up to speed, or vice versa.
- I've worked in teams in other fields, it takes a lot of additional communication and discussion to keep everybody aligned and going in a productive direction.
- Few other projects are really compatible. Even people working along similar lines have conceptual differences in knowledge representation or language parsing. If compatibility issues didn't exist, project Cyc would have been a grand success.
- I'd have to be crazy or a saint to hand out years of hard work for free, and I ain't no saint. 
- People are very unreliable.

I have a policy of only ever hinting at solutions, unless one is buried particularly deep. For one thing, because I think the world would be a better place if people used their brains more. WriterofMinds has come up with interesting twists of her own that advice from me might have stifled otherwise (I am referring to the development of bot initiative, something I didn't do). I find it rather contradictory to call discovery useless yet at the same time complain that people don't think for themselves.

PS: If I sound grumpy, it's monday morning.

[LOCKSUIT]

Let's analyse something here...

-More people working on the same idea = faster Discovery & Development (D&D) for n.

In other words, you have 2 brains working on discovering n, plus an ensemble (Team Power Rangers GAN).

Think about it. If I have you as a buddy, you could spark ideas in my brain or mentor things I want to know, and quicken the discovery rate of n. Why do teams exist anyhow? Power in numbers, different knowledge or 'skills'.

Say I convinced someone to come my way. That would be good for everyone. But it requires time (to hear me out).

Convincing would be not small things but the most important thing - the path one takes. Is it NLP? Robotics motion planning? What is intelligence? Then the 2 people are on the same path~

[spydaz]

@Spydaz

Being as I am the only person mentioned by name in your posts, are you implying that…

I have noticed that the older ones like to steer younger or newcomers on journeys of useless discovery.... without actually answering direct questions or even sharing previously Conquered code!

Refers to me?

 

Obviously I hope not; as you should notice, i used you as an example of "one of the older ones" being 50 and all..... 50 is not old .....  But i am  45 and some consider it as old "weathered" .

but noticeably sir, Quick to be offended is actually a sign of age Lol;

" Well korrelan for example is 50 I think and doesn't have time to work with others especially me being 22 "

Its hard to work with the younger Gen! but after recently experiencing a series of bereavements i understand knowledge can be lost so easy!...... It has made me realise; "share the knowledge before we die" exactly why people carved knowledge in stone!

 Maybe its another reason for the AI WINTER ! most of the other disciplines in computing there is a lot of sharing/ collaborating going on. whereas AI seems to have a Commercial aspect (make a millionaire for the discoverer), and a secretive aspect which makes it hard to get ahead.....

Exactly why forums like AIDREAMs are required!  Well done Freddy; as this forum has maintained and forum members from Prominent Backgrounds and the chat-bot pioneers are floating here.... So there is a lot of potential!

In the beginnings of AI it was also shrouded in a dark light as the commercial outlook for chat-bots are SEX-BOTS. and the people who want to create such products have been here at the beginning too! now there are so many possibilities for AI that commercial companies and colleges/universities are jumping in!
but the pioneers such as us are no where to be seen "still in the shadows"

Personally i have a lot of knowledge to share Paduwan i have tried making the odd You-tube but i'm not that guy! but i am a great technical teacher!
As you can see once is start talking i cant stop!



So No korrelan !

I'm always available on skype!

PS: Don Patrick I completely understand and i am also in a similar position i would never share my complete code ever (Arron the same)  as its a lot of achievement but maybe some Classes or Odd functions is no problem !
Or even explaining how i overcame the particular problem harmless....
I know the research process is very hard! jeeze my head is also bursting! and still don't know enough. there are no fast tracks to the future its all hard work. And appreciation is hard to find Sir!

And we are battling trying to create a technology that has already been achieved by the military science but not shared with the Academic or Commercial communities......

[LOCKSUIT]

Lol my mom just gave me speech and said about our dog (who passed away):

"See how we had no help. There's strength and power in numbers. You see how the army/military march with power and force? They work in numbers. Always remember that my son. There's power in numbers."

Lol.

Then I'm sitting there like   ._.

[8pla.net]

I have a policy of only ever hinting at solutions.

Don,

Please, hint at a few solutions for my released code.  Being a light weight design, it needs solutions to figure out how to keep this chatbot simple, light weight but as capable as possible in the fewest lines of code. Excellent suggestions from infurl have already helped wonderfully.  For convenience I linked back to the A.I Dreams forum: Plain Text.  Not to mention other members here have been supportive.

Please, I hope this is not considered an attempt to hijack this thread over to my thread. I only intend my thread as a reference for Don Patrick.  I just figured our discussion about releasing code, may benefit by referencing actual released code.   Why would a developer of advanced level A.I. object to releasing a beginner level A.I. source code?

In short, If you don't want to release your pride and joy source code, that's fine.  No one expects you to do that. It might become the most advanced operational A.I. released to date, if you did.  Maybe the world is not ready for that?  But (said in a polite tone) if you are a such a great A.I. developer, then why won't you help a beginner A.I. source code project for your fans? 

NOTE: Permission to quote this post is denied, unless you clarify that it is meant to use some humor in a request for help.

[Don Patrick]

I don't consider myself a great AI developer, but I am just not particularly interested in chatbot projects. I once emailed a professor who was recommended to me to have a look at my project, and never got a reply. I later learned that professors are extremely busy people and they don't benefit from randomly helping other people when it detracts from the time they can spend researching their own projects and helping their own students.
Similarly, I've just taken an hour looking at your code to indulge you, but I could have spent that hour improving my own program for the Loebner Prize otherwise. I know you for abandoning projects as quickly as starting new ones, and while that may be considered admirable in a way, it also means much effort goes wasted, in my view.

Locksuit, I do agree that there is power in numbers, when all noses are pointed in the same direction. But in an experimental field that is not necessarily the case without a strict hierarchy, like there is in an army or in a professional environment.

[LOCKSUIT]

So let's get on the same page already.

We want to create AGI right?

AGI is literally "forming discoveries".

Don't believe me?

How do we learn our sensory knowledge and actionary skills? Senses, and by internal thinking. This thinking discovers better/new knowledge/senses.

A pure example is a baby sim in UE4. It learns to crawl but after that it needs to be taught all of man's progress and then research/discover how to build nanobots, cure cancer, etc. The body / learning to walk is NOTHING. The baby would literally sit there thinking about how to build nanobots. I usually sit all day myself lol. I take in others work and discover my own new things at my chair (or during lunch lol).

[infurl]

Why would you want to work with people when you can write software that does the work faster and more accurately.

[LOCKSUIT]

And is that going to bring AGI within the next 5 years? We don't have that yet. It would have to play around with adding things and tweaking parameters. One-shot coding is more likely how AGI comes. It takes brains.

[Don Patrick]

We want to create AGI right?

Actually I don't , so you'll have to count me out.

As for the whole "learning from sensory experiences" approach, this project may interest you. If you get a degree maybe you could join such projects. Now if you'll excuse me, I'd like to get some work done rather than discuss the work that I could be doing.

[spydaz]

Myself

i'm actually only interested in Conversational AI

[/b][/u]
 

[8pla.net]

Don said, "I once emailed a professor who was recommended to me to have a look at my project, and never got a reply."
Your A.I. project is likely beyond the level they teach at college.   So it is safe to take the lack of a reply as a compliment.

I can recommend a college professor to you, who has helped another well known A.I. developer named Bruce.
So, simple to verify beforehand. And, I can guarantee a faculty email reply from this college professor,
with no requirement to share your project.

I suspect there are more college professors in our A.I. community than meets the eye.
But for the reasons you discussed, Don, they may not advertise it.  Imagine if a whole classroom of students
find their college professor online and join the discussion simultaneously like they do at school. 
Hard working professors deserve a break.  Even from a dream job, like getting paid to teach A.I. , robotics, and how to build a simple chatbot...  I mean, if you may empathize with that?

[Don Patrick]

It is kind of you to offer, but I only brought it up as an analogy to express that my time, too, is a commodity that I do not have in abundance to share for collaboration. The time when I mailed that professor was years ago, and I have since found the answers to the questions I had then.

Spydaz: Conversational AI is one of my sidequests. Can you recommend any particularly inspiring research in that area? I found that neither AI professionals nor chatbot developers (aside from Bruce Wilcox) concerned themselves much with the psychology behind conversation, and I found Discourse Analysis "research" rather medieval. I've taken to psychology books instead, but it seems to me there ought to be some Dialog Manager systems that are more intricate than predetermined tree paths.

[spydaz]

Opps Top Secret !

I have sent PM sir!.......

In truth there are no advancements in conversational AI ..... to tell the truth they are all using keyword and response! or hard-coded conversational trees.

People are a product of their environments and their exposure to the world. Truth is subjective to that person even though there is an overall truth.  Could your bot know all truth or learn the truth as a human?

What should the bot really say ? what are its intents / Imperatives?

As a business tool AI is just like a computerised doctor its a closed system with pre-programmed solutions and responses. although considered to be AI its a BoT.

the AI can be defined as not being in a closed system. therefore its concentration is on learning to respond. and learning information; gathering information. IT has no feelings there fore it has no emotional responses. (we can always design simulated responses but is we do then it would simulate being angry... whats the point!(SEXBOTS again).

Its this distinction that defines a project in Conversational AI.....

Psychology: plays a large part in designing a "Bot which acts like a human", "or make human Decisions":
Long-term memory / Short term memory / Working memory - Its how we organise our thoughts and knowledge...... Variables / Transnational database / Data-warehouse comparatively.
there are so many aspects of psychology that are required for AI.... If designing a Course on AI Formal logic / Psychology / Data warehousing / Machine learning   would play key roles.... designing Chat bots would also be KEY! .... And obviously PROGRAMMING....

In the World of open source - is this not the same as menu driven software? (for users not programmers)

[spydaz]

AI.API .... https://dialogflow.com


a good resource for an API to integrate your AI into Social media.
This is an example of (keyword / Detect) LOL.

This is a jump start for beginners. we are in that world now ... where the powers at be do not want you to know how to do things ..... Just give you some answer that will satisfy you.... ie : open source provides answers and short cuts with out you solving your own problems.... this means later when you come across another hurdle you have no understanding ho to cross it.... unless another open source exists to perform or solve your problems .... This the problem today!
Whippersnappers ... thinking they know when all they are doing is just using existing tools! Tensor-flow etc are exciting projects yes but if you learn and use them the discoveries are not yours... essentially you have used their tech to accomplish your goals ...... numbpy lol its another shortcut.....
For a standard programmer there is nothing in numbpy that you cannot write for yourself... the librarys are like menus of commands to use . little black boxes...... as a beginner its useful but as you grow .... You should not need them ...... People choose a programming language now days because of the resources they can get from open source communities.... bad move! eventually open source librarys become commercial librarys which cost money! with you as the beta tester!