Your thoughts wanted on a node AI project

[Dennis Kane]

The site is powered by an express server, which exec's a separate node AI process for each new IP address.  Each process will automatically end after 10 minutes of inactivity.  You can interface with the process either through a browser at ai.webcyte.net or by sending a URL encoded GET request to http://ai.webcyte.net/send.  So, the comand line for the phrase, "there is a boy named joe" using curl will look like:

$ curl http://ai.webcyte.net/send/there%20is%20a%20boy%20named%20joe

Detailed instructions can be found at ai.webcyte.net.

This project is currently meant to be a prototype in order to build interest in developing a robust AI system that "just works" (read: NOT another chatterbot!).  It can possibly be used as the intelligent backend to a Siri-like voice interface in the not too distant future.

I am interested in getting in touch with academic/research types as well as business types.

Will probably want to develop an open source version for non-business purposes that can be used on the client side (eventually) and keep the supported commercial version well hidden from prying eyes on the server side.

Thoughts?

[Ted Young]

Really nifty.  It would be fun to play with.  One way to do it would be to open source the AI engine, but keep the large scale hosting and support architecture hidden.  That way academics and users with other needs could play with it and contribute.

Ted

--
Job Board: http://jobs.nodejs.org/
Posting guidelines: https://github.com/joyent/node/wiki/Mailing-List-Posting-Guidelines
You received this message because you are subscribed to the Google
Groups "nodejs" group.
To post to this group, send email to nod...@googlegroups.com
To unsubscribe from this group, send email to
nodejs+un...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

[Dennis Kane]

Hey, thanks!  This is really a labor of love at the moment, but I do have some pretty big ambitions.  I just think that with all of this Siri business floating around, there are very real expectations for general AI capabilities being built up in people's heads.  Many people think that IBM Watson is an example of AI.  I don't believe that is true.  Watson is just meant to sift through a static corpus of facts based on keywords and then fashion an answer based on levels of confidence.  From what I've seen, there is no facility to correctly handle these two simple  statements (which I believe is an absolute minimum example of AI capabilities):

There is a wild boy named Joe and a crazy boy named Mike

Who is the crazy boy?

This is meant to be something of a very high-level programming language that does not internally (at the parser level) make policy decisions about statements of fact.  In their common vernacular, people will be able to "teach" their computers about whatever facts/subjects they are interested in, and then execute natural language programs: with conditionals, loops and all.

I think that such a tool as this is necessary if there is ever going to be a truly semantic web that people will be able to navigate purely with human logic.  And when enough of these kinds of services are populating the web, we will begin to be able to see the emergence of systems that even laypeople will recognize as particularly responsive to their needs, and yes -- even "intelligent".

Ted

nodejs+unsubscribe@googlegroups.com

Ted

nodejs+unsubscribe@googlegroups.com

Ted

nodejs+unsubscribe@googlegroups.com

[Evan]

That's great!

I am reminded of this blog [[ http://www.briangrinstead.com ]] which also has some neat AI algorithms written in JS.  He wrote them for the browser, but I can tell you his A* implementation [[ http://www.briangrinstead.com/blog/astar-search-algorithm-in-javascript ]] works great in node :D

[Brandon Benvie]

The first AI that impressed me and has value for me is the one that delivers me what I want in the day before I consciously realize I want it.

[David Whitten]

This level of AI is closer than you think.

This is not exactly the same sentences, because RACE
isn't quite smart enough for your original English sentence,
but a small change does produce interesting results.
 I went to  http://attempto.ifi.uzh.ch/race/
and put in these axioms:

Joe is a wild boy.
Mike is a crazy boy.

and the query:

Who is the crazy boy?

and got back this:

overall time: 2.86 sec; RACE time: 0.04 sec

Axioms: Joe is a wild boy. Mike is a crazy boy.

Query: Who is the crazy boy?

Parameters:

There is 1 message.

Importance	Type	Sentence	Problem	Description/Suggestion
warning	anaphor	1	The definite noun phrase 'the boy' does not have an antecedent and thus is not interpreted as anaphoric reference, but as a new indefinite noun phrase.	If the definite noun phrase 'the boy' should be an anaphoric reference then you must introduce an appropriate antecedent.

The following minimal subsets of the axioms answer the query:

Subset 1

• 2: Mike is a crazy boy.
• Substitution: who = Mike

nodejs+un...@googlegroups.com

[Jann Horn]

On Mon, Apr 23, 2012 at 10:50:25AM -0700, Dennis Kane wrote:
> Detailed instructions can be found at ai.webcyte.net.
>
> This project is currently meant to be a prototype in order to build
> interest in developing a robust AI system that "just works" (read: NOT
> another chatterbot!). It can possibly be used as the intelligent backend
> to a Siri-like voice interface in the not too distant future.

> Will probably want to develop an open source version for non-business

> purposes that can be used on the client side (eventually) and keep the
> supported commercial version well hidden from prying eyes on the server
> side.
>
> Thoughts?

Looks cool, I want the source! :D

How much time did you spend on that so far?

I think you should query wiktionary.org or so for unknown words. It doesn't even understand "cake is delicious". :( Also, you'll have to somehow support names it doesn't know about - how do you want to do that?

[Dennis Kane]

Hi Jann,

I'm not sure you really want the source at this point... at least not
without some pretty detailed explanation to go along with it!

I first started working on the algorithm six years ago in perl, which
also happened to be when I was first learning how to program. Given
my inexperience, that effort didn't amount to much.  But it has only
been less than a month since I began from scratch in Javascript.  So
the overall algorithm has been sitting in my head for quite some time,
but it has only been very recently that I've decided to take up the AI
challenge again... this time with far more programming ability.

As far as it goes, I have so far put zero effort into having a decent
variety of words being recognized.  On the site, you will find
instructions to see what words are recognized, and how to add to the
list.

At present, there is no facility to refer to things without using some
kind of article or possessive... So to do something like what you
want, you will have to type the following two lines to get it to
recognize the words:

nouns:cake
adjs:delicious

Then you can type something like:

There is a delicious cake

or...

There is a cake
It is delicious

or even...

There is a cake and it is delicious

Then you can say...

There is a boy named mike
He has the cake
Mikes cake is blue

The idea of saying "cake is delicious" is much more dangerous because
it labels an entire type class.  This would disallow the existence of
bad tasting cakes.  I think at this early phase, it is much better to
think in terms of instances rather than entire classes.  Then once the
logic of instances gets worked out, we can start thinking about
classes (carefully!).

The program does allow you to give arbitrary names like this:

There is a boy named Zebulon

or

There is a girl
She is named Zelda

It currently looks for any arbitrary sequence of characters after the
word "named" and then adds this sequence into the list of recognized
names.

>  application_pgp-signature_part
> < 1KViewDownload

[Ted Young]

> Hi Jann,
>
> I'm not sure you really want the source at this point... at least not
> without some pretty detailed explanation to go along with it!

Haha I used to think that way too. But it's better to just open source early if open source is your goal, and slap a warning label on it saying it's in the very early stages. I've found people to be pretty understanding and non-judgemental about code quality, unless you are proposing that it is production ready when it is not. :)

[Joshua Holbrook]

> The first AI that impressed me and has value for me is the one that delivers me what I want in the day before I consciously realize I want it.

Welcome to consumer analytics. :) If you have enough data you can
extract all kinds of trends.

--Josh

> --
> Job Board: http://jobs.nodejs.org/
> Posting guidelines: https://github.com/joyent/node/wiki/Mailing-List-Posting-Guidelines
> You received this message because you are subscribed to the Google
> Groups "nodejs" group.
> To post to this group, send email to nod...@googlegroups.com
> To unsubscribe from this group, send email to
> nodejs+un...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/nodejs?hl=en?hl=en

--
Joshua Holbrook
Engineer
Nodejitsu Inc.
jo...@nodejitsu.com

[Dennis Kane]

To open source or not to open source... THAT is the question!

Do I want to SAVE the world or do I want to OWN it, haha ;)

I really do want to keep it close to the vest at the moment.  I am not feeling too liberal right now.

But I would like to start getting a team together to help make it commercially viable (wink wink nudge nudge).  This, to me, is by far the best way to attract the best talent for such a demanding sort of project.  Will keep my ears open for any interested VC's sniffing around here...

[Dennis Kane]

Now that I think about it more, Javascript makes the entire issue of open source much less urgent given that one can pass in their own custom functions, either for replacement purposes or for extension.  I'll just have to provide the API for what kinds of arguments will be passed in and what should be returned.  You'll be able to save the functions on the client side through the HTML5 File API, and then send them via XHR into the process at runtime.

And when it comes to how the program works in general, I would much rather give people a conceptual understanding through pseudocode examples and have them develop their own implementations.  I know that this way requires actual time and thought, but hey... these things never hurt anyone did they?

I do find the whole open source movement a tad pointless.  For me, the source code always takes a back seat to good API documentation, well commented header files, decent man pages, a good bug reporting/fixing system, etc.

[Matt]

On Wed, Apr 25, 2012 at 1:03 AM, Dennis Kane <dka...@gmail.com> wrote:

I do find the whole open source movement a tad pointless.  For me, the source code always takes a back seat to good API documentation, well commented header files, decent man pages, a good bug reporting/fixing system, etc.

I think you're probably in the wrong place to make that kind of comment. The entire node ecosystem is built on open source. It's OK to be pragmatic and say there's a place for closed source too (I would agree), but don't call open source pointless.

[Dennis Kane]

Was just saying it is, a tad.  Maybe "pointless" was a bad word choice.  Overblown is probably better.  Don't think you can deny that without those other things, the source code of any non trivial program is quite a worthless commodity in and of itself.  The entire node ecosystem, at the moment, is at the point where the programs are still fairly trivial, and so the value of source code is still relatively high.  That state of affairs won't last forever.

Anyway, I guess the larger point is that I feel my program has the potential to be quite profitable in the future.  I still am fairly young and quite hungry.  I would much rather be very close with a few business associates than just another anonymous user on github who everyone takes for granted.  I have ambitions, and I won't apologize for that.

On Wednesday, April 25, 2012 5:33:36 AM UTC-7, Matt Sergeant wrote:

[Matt]

On Wed, Apr 25, 2012 at 10:43 AM, Dennis Kane <dka...@gmail.com> wrote:

Was just saying it is, a tad.  Maybe "pointless" was a bad word choice.  Overblown is probably better.  Don't think you can deny that without those other things, the source code of any non trivial program is quite a worthless commodity in and of itself.  The entire node ecosystem, at the moment, is at the point where the programs are still fairly trivial, and so the value of source code is still relatively high.  That state of affairs won't last forever.

You're digging an even bigger hole - there are significant and important applications written in Node that are far from "trivial". Take my own Haraka for example - almost half of the contributions have come from Craigslist. The application has been significantly improved because it is open source.

 

Anyway, I guess the larger point is that I feel my program has the potential to be quite profitable in the future.  I still am fairly young and quite hungry.  I would much rather be very close with a few business associates than just another anonymous user on github who everyone takes for granted.  I have ambitions, and I won't apologize for that.

That's fine - almost everyone who contributes to open source also makes money from software that isn't open source. But you're belittling a huge community of people who have spent way longer figuring this stuff out than you have. Being condescending about the node ecosystem and node applications and open source in general on this mailing list is just going to turn people off your future products. Remember that stuff you say on the internet stays there forever.

Matt.

[Ted Young]

> Was just saying it is, a tad. Maybe "pointless" was a bad word choice. Overblown is probably better. Don't think you can deny that without those other things, the source code of any non trivial program is quite a worthless commodity in and of itself. The entire node ecosystem, at the moment, is at the point where the programs are still fairly trivial, and so the value of source code is still relatively high. That state of affairs won't last forever.

Node itself is a large, non-trivial open source project, now well funded because people think they will be able to make money from it in various ways. How far would Ryan have gotten if he decided to make node closed-source PaaS? There are different business strategies based on what kind of buy-in and adoption you think you need, and where the actual money is to be made. You're kind of insulting the people who's technology you are using. I bet whatever db you are using is open source, also backed by a profitable company. And please don't say "it's different" because node is a tool or a platform etc, what you are describing could also be viewed the same way. It's closer to Solr than it is to Twitter.

> Anyway, I guess the larger point is that I feel my program has the potential to be quite profitable in the future. I still am fairly young and quite hungry. I would much rather be very close with a few business associates than just another anonymous user on github who everyone takes for granted. I have ambitions, and I won't apologize for that.

Then why are you asking this list for help and feedback? Your actions are showing that you value community involvement on some level. You mentioned you've been working on this by yourself and it's never gone anywhere, maybe you could start generating steam and interest if you let other people get involved on their own terms. Pursue whatever strategy you like, if you can make closed-source PaaS work I still wish you success, but you should seriously re-evaluate your attitude.

[Dennis Kane]

OK guys, firefighter Kane to the rescue.  I said "tad".  For full disclosure, I've been using Debian for many years, and I feel it is by far the best OS in existence.  I am writing this in nano, the source code of which I have done some hacking on.

I'm here because I can't do this thing by myself.  If someone wants to help me, then the code will be "open source" for them.

If things can begin to get off the ground, then we can start to think about releasing some code with a liberal license.

I don't know what else to say other than I'm excited as hell to try to make some major noise in the free markets.  If I come in here being all gung ho about open sourcing all of this, then that will be much less likely.

The test of this thing is that it "just works".  If anyone wants to help me get it to "just work" even better, I would be stupid to not bring them along on this crazy ride.

As far as my attitude is concerned, I admit to being an a-hole.  Always have been, probably always will.  I'm a little too old for reevaluation to do any good.  Hopefully my talents will make up for my personality shortcomings in a few people's eyes.

I would really like to get off the religious warfare and get back to the thing itself... ie, what kinds of statements would YOU like the program be able to handle?

[Jann Horn]

On Wed, Apr 25, 2012 at 11:14:12AM -0700, Dennis Kane wrote:
> I don't know what else to say other than I'm excited as hell to try to make
> some major noise in the free markets. If I come in here being all gung ho
> about open sourcing all of this, then that will be much less likely.

Seems like you still have a long way to go to reach that. :D

> I would really like to get off the religious warfare and get back to the
> thing itself... ie, what kinds of statements would YOU like the program be
> able to handle?

Well, your long-time goal is to make a good program that can answer questions using some logic if necessary, right? Well, I think you'll have to make your program able to parse Wikipedia texts and so on for that. Maybe start with the "simple english" wikipedia ( http://simple.wikipedia.org/wiki/English_language )? Good luck. :P

And before you can do that, you'll probably have to build in a dictionary backend.

Oooh, and when all of that works, make it able to accept commands. :) "Make me a sandwich"

[Dean Landolt]

On Wed, Apr 25, 2012 at 2:14 PM, Dennis Kane <dka...@gmail.com> wrote:

OK guys, firefighter Kane to the rescue.  

Some firefighter -- you're the one w/ the lighter, now you pour on fuel and claim you're here to help?

 

I said "tad".  For full disclosure, I've been using Debian for many years, and I feel it is by far the best OS in existence.  I am writing this in nano, the source code of which I have done some hacking on.

I'm here because I can't do this thing by myself.  If someone wants to help me, then the code will be "open source" for them.

 

I don't think that word means what you think it means.

 

If things can begin to get off the ground, then we can start to think about releasing some code with a liberal license.

I don't know what else to say other than I'm excited as hell to try to make some major noise in the free markets.  If I come in here being all gung ho about open sourcing all of this, then that will be much less likely.

 

Now I know that word -- "free markets" -- doesn't mean what you think it means. There's nothing contradictory about open source and free markets. Quite the opposite, as both logic and recent history demonstrate, and as has already been pointed out to you.

 

The test of this thing is that it "just works".  If anyone wants to help me get it to "just work" even better, I would be stupid to not bring them along on this crazy ride.

As far as my attitude is concerned, I admit to being an a-hole. Always have been, probably always will.  I'm a little too old for reevaluation to do any good.  Hopefully my talents will make up for my personality shortcomings in a few people's eyes.

You don't come off as an asshole -- and besides, we have no problem with assholes: we're developers. Saying the things you're saying, to the audience you're saying them to -- well, it's that you come off as an idiot. The good news is this is a lot easier to fix.

 

I would really like to get off the religious warfare and get back to the thing itself... ie, what kinds of statements would YOU like the program be able to handle?

Oh, you're looking for help? Here's a suggestion: you could open up the code and see what people do. Far be it from us to tell you what to do with your code, but you're the one asking for help, and this is naturally the first bit of feedback you're going to get. Why would any of us invest much more of our precious time helping you improve your proprietary product? The world doesn't work that way -- at least, not anymore.

So yes, back to the thing itself: node.js (this is the node.js list, after all). You've expressed clearly that you have no interest in making your work part of node's ecosystem. So at this point you're just noise, no different than a recruiter. You just said as much yourself: "I'm here because I can't do this thing by myself." At least recruiters, even the worst of them, have a payed position to offer and the good sense not to directly insult the target audience :P

Good luck to you and your product, but kindly take your noise elsewhere.

[Oliver Leics]

> I would much rather be very close with a few business associates than just another
> anonymous user on github who everyone takes for granted.

Uh wait, this sound familiar:

"Nothing would please me more than being able to hire ten programmers
and deluge the hobby market with good software."

http://www.blinkenlights.com/classiccmp/gateswhine.html

> Will keep my ears open for any interested VC's sniffing around here...

Dont wait for them to contact you, *just* contact them:

https://plus.google.com/117388252776312694644/posts/BipcGutkkcc

http://toaberlin.com/

[Ted Young]

speaking of open-source success in the marketplace: http://allthingsd.com/20120425/automattic-grows-up-the-company-behind-wordpress-com-shares-revenue-numbers-and-hires-execs/

Ok, back to talking about node…

[Dennis Kane]

Seems like you still have a long way to go to reach that. :D

I don't think it should be too much longer before I can start making some sales calls based on the progress I've already made.

Well, your long-time goal is to make a good program that can answer questions using some logic if necessary, right? Well, I think you'll have to make your program able to parse Wikipedia texts and so on for that. Maybe start with the "simple english" wikipedia ( http://simple.wikipedia.org/wiki/English_language )? Good luck. :P

And before you can do that, you'll probably have to build in a dictionary backend.

Nah, wikipedia is made for human consumption.  As long as there is a good enough parser on the market, there will be plenty of people motivated enough to write compliant "programs" directly into it.  I see this thing eventually taking a pretty good bite out of Wikipedia's market share. The way it'll work, for the forseeable future, I think, is that we will work heavily with corporate clients on what kinds of words and statements they would like their customers and employees to be directly executable.  Trying to do everything for everyone all at once is simply a prescription for a major disaster.  I see many clients supplying the dictionaries themselves.

 

Oooh, and when all of that works, make it able to accept commands. :) "Make me a sandwich"

Funny, but I think we're gonna need alot of help from MIT's AI robotics department before that becomes anywhere close to feasible. 

[Stewart Mckinney]

Hey wait who hired the troll?

--

[Dennis Kane]

You don't come off as an asshole -- and besides, we have no problem with assholes: we're developers. Saying the things you're saying, to the audience you're saying them to -- well, it's that you come off as an idiot. The good news is this is a lot easier to fix.

 

I'm an idiot because of my political views, and not my programming ability, I assume.  The only thing I'm worried about is the programming.  If you've got some criticism about that, I'm all ears.

  

Oh, you're looking for help? Here's a suggestion: you could open up the code and see what people do.

No.  Now chill out about the open source thing.  It's boring.  Redundant. Etc.

 

Far be it from us to tell you what to do with your code, but you're the one asking for help, and this is naturally the first bit of feedback you're going to get. Why would any of us invest much more of our precious time helping you improve your proprietary product? The world doesn't work that way -- at least, not anymore.

Dude, I'm looking to get a business off the ground.  People work with each other and try to make money off of their labors.  It has happened once or twice before.  You say the word "proprietary" as if its one of the seven deadly ones.

 

So yes, back to the thing itself: node.js (this is the node.js list, after all). You've expressed clearly that you have no interest in making your work part of node's ecosystem. So at this point you're just noise, no different than a recruiter. You just said as much yourself: "I'm here because I can't do this thing by myself." At least recruiters, even the worst of them, have a payed position to offer and the good sense not to directly insult the target audience :P

Sorry.  I don't have any personal feelings about node's ecosystem.  It's just a tool that I take for granted.  I set up the site with express, and now I don't think about it.  It just works.  I have a server hooked into the main parser to shuttle stuff between it and express.  Sometimes I need to make use of the filesystem.  Other that that, I'm pretty much using straight javascript. It's google's v8 work that's doing most of the heavy lifting for me.  Besides, I didn't know that I had to have feelings for the tools that I use in order to be accepted in the development community.  And yes, I am recruiting people who are interested in doing something awesome with this tool.  And I'll offer something even better that paychecks: how about co-ownership?

  

Good luck to you and your product, but kindly take your noise elsewhere.

Nah, I think I'll chill here for awhile.  I'm sure there are some people around here like me that appreciate new technology over the same old politics. 

[Isaac Schlueter]

Please do not bicker over politics here. This is not the place, and
it never ever ends well.

Please do not call each other idiots or make other personal remarks.
If you want to say that a piece of *code* is terrible, or that a
software approach is misguided, or that a statement is factually
untrue, then that's perfectly fine (as long as it actually is!) but
the Node.js user list must remain respectful. (Challenging someone
for benchmark numbers when they've made performance assertions is also
not only accepted, but encouraged.)

But a personal insult is not allowed, even if it's warranted. If
someone's *actually* causing problems, let me or one of the other
admins know, and they'll get a warning, and if that doesn't work,
they'll get banned. (Even if the person is an admin. You can
complain about me to the other core devs on this list, and they'll set
me straight.) Calling someone a troll is insulting if they are not,
and it is entertaining if they are.

Don't complain in public about another user. Just talk to an admin in
private. (Or email them directly off-list, if you feel comfortable
doing so.)


More on topic,

The MIT license in node (as well as the MIT and BSD licenses in the
vast majority of open source node programs) explicitly allows use for
closed-source commercial applications. If Dennis wants to keep his
source closed, then that's his prerogative. In fact, he can take all
the modules we create, and use them in his closed-source proprietary
thing, so long as he abides by the extremely liberal licenses that
most of them use. If you think he's wrong about the value of open
source, or its overblownness, refute it with data. (Or don't, and
just go back to cranking out awesome open source software.)


Dennis, since you came here ostensibly trying to raise interest among
other developers (and have been mostly successful, reading through the
thread), I would suggest re-thinking your approach somewhat. I'm not
talking about what's right or wrong, merely what's effective. Many
people come to a project like Node.js because they feel strongly about
open source software. If your goal is to recruit them, you should
think about the effects that your words have. If you want to recruit
developers who *aren't* passionate about open source software, then
you're in the wrong place.

[shawn wilson]

Per parsing language, you might want to look at what python's nltk and Java's gate (less familiar with the later) have done for inspiration.

[Dennis Kane]

On Thursday, April 26, 2012 3:35:48 AM UTC-7, shawn wilson wrote:

Per parsing language, you might want to look at what python's nltk and Java's gate (less familiar with the later) have done for inspiration

Yeah, I saw the python nltk yesterday just to see what was currently out there.  Despite the fact that I absolutely abhor the indentation enforcement of python (neither here nor there), that whole project is just more of the same "everything and the kitchen sink" mentality.  There is so much in the way of graphical representation that has nothing to do with a program that just gets things right.  That 's all I care about.  As long as the parsing works very well (the logic of 'and', 'or', 'if', 'then', 'because', etc), then everything else will start falling into place.  Just think of this as you would any c compiler or script interpreter.  The compiler must be able to get the intended logic right.  As long as that is the case, the user/programmer can do whatever the frick he or she wants as far as creating meaningful content.  This thing is doing pretty damn good with the logic on some highly convoluted statements and sequences thereof.  Forward thinking businesses will have to take note of this fact... they have no choice.  Then we can all start getting together to determine how to develop the exact dictionaries and statement types that will fit their needs as well as possible.

[Dennis Kane]

On Wednesday, April 25, 2012 7:07:52 PM UTC-7, Isaac Schlueter wrote:

The MIT license in node (as well as the MIT and BSD licenses in the
vast majority of open source node programs) explicitly allows use for
closed-source commercial applications.  If Dennis wants to keep his
source closed, then that's his prerogative.  In fact, he can take all
the modules we create, and use them in his closed-source proprietary
thing, so long as he abides by the extremely liberal licenses that
most of them use.  If you think he's wrong about the value of open
source, or its overblownness, refute it with data.  (Or don't, and
just go back to cranking out awesome open source software.)

Let's all remember what open source is really all about.  A program is called closed source if it is distributed in binary format only.  The open source movement makes the demand that one cannot distribute a binary program using, for example, GPL'ed code without also making the source code available.

But a server side process has nothing to do with any of this!  I am not distributing the program... I am interested in running a service like Google (which will hopefully eventually overtake them... but don't tell anyone I said that).

Is anyone seriously demanding Google to freely distribute all of their painstakingly developed search algorithms?  Not likely!

The basic fact is that this thing is the result of years and years of absolute psychological warfare between me and my computers.  And given the fact that robust NatLang Processing (weak AI) is something of a holy grail for tech enthusiasts, the stakes in all of this are quite big.

I am not saying that releasing the code for this won't ever happen.  But I am saying that releasing code is a very major decision that should never be taken lightly.  All I do know is that now is not the time.

But I will talk about it.  I won't want to give away too many of the technicals in an open forum, but I will give some of them away to people who I can trust.  And in order to me to be able to trust someone fully, I have to feel that they actually have an interest in the problems surrounding NLP.

The Net is absolutely littered with freely available NLP projects, code and all .  They are just not interesting.  But you know what is interesting?  A site with a lone input box on a white background with a snazzy, colorful logo above it (sound familiar?) that "just works" as advertised.  I want to get something like that on the site pretty soon.

Now that I'm pretty well done with the hair-pulling aspects of my coding, I can start having fun with putting variety into the thing (giving it the "wow factor").  Different kinds of words, statements, sense checking, etc.  This kind of stuff should not be very difficult for any competent programmer.

That's what I really want to start getting on the same wavelength with people around here.  Stop thinking that this thing has to be some kind of end-all-be-all killer app from the outset.  I mean, there are so many things to be done.  I want to be able to translate natural ways of referring to time points (last week, the day after tomorrow, etc) into their precise Unix timestamps.  Not very difficult work, but it's something that really should start getting worked on.

I really feel that this thing could give quite a few of you out there a comfortable living.  This could open up totally new vistas of the tech sector.  This has the potential of going places pretty quickly, and the earlier that people get in on it, the greater the potential for reward.

Dennis, since you came here ostensibly trying to raise interest among
other developers (and have been mostly successful, reading through the
thread), I would suggest re-thinking your approach somewhat.  I'm not
talking about what's right or wrong, merely what's effective.  Many
people come to a project like Node.js because they feel strongly about
open source software.  If your goal is to recruit them, you should
think about the effects that your words have.  If you want to recruit
developers who *aren't* passionate about open source software, then
you're in the wrong place.

I've already been told more than once to rethink/change my approach.  Seriously?  I mean, let's get serious here.  The only time I get snarky is when people get pushy with their demands to just hand out something that has resulted from years of torment.  This just is not any old program.  It just isn't.  Period.

Again, I have to reiterate that there is a vast difference between the philosophy of open source as espoused by Eric Raymond, Richard Stallman and company and the reality of people actually opening up source code in order to actually *improve* it.  Please someone break open the source for emacs right now and make it better.  I dare you.  My philosophical views are always changing.  Conservative, liberal, whatever, blah blah.  I'm just not dealing with any of it now.  This is a huge technical challenge.  That's all.  At this point, I don't care if you sit in trees for months on end to save them from getting cut down or if you are a dues paying member of the NRA.  I really don't.

[Mark Hahn]

I guess the question everyone is thinking (or asking) is why are you here?  I can understand the recruiting, but I can't think of any other reason.

--

[Dennis Kane]

On Thursday, April 26, 2012 10:57:26 AM UTC-7, Mark Hahn wrote:

I guess the question everyone is thinking (or asking) is why are you here?  I can understand the recruiting, but I can't think of any other reason.

Which means: if I don't want to give away the entirety of my labors that I've been agonizing over for all these years, then I have no legitimate reason to post to this forum.  I already said that I will talk about it.  But if no one is interested in the coolness factor of the thing, then I guess there's nothing else for me to talk about.  I can just move on to other places to find people who want to help me get something pretty awesome started.  No worries.  I'm just looking for friends who would like to work together.  I can't force anyone to get interested.

Oh yeah, there's always this reason:  this thread has top ranking on google search "node ai", and my site is now listed on the engines.  Before I posted here, my site wasn't even on google!

[Mark Hahn]

>  Which means: if I don't want to give away the entirety of my labors that I've been agonizing over for all these years, then I have no legitimate reason to post to this forum.  

No, that is not what I meant.  I literally meant "why are you here?". 

I had only thought of the recruiting angle but now you have pointed out the public relations angle.  So you have two reasons.  In my humble opinion, those reasons are not appropriate for a public discussion group like this.  There is supposed to be a give and take, not a take and take.

[Glenn Block]

Dennis

As Isaac said, the fact that you want to build something that is closed source is not a problem.

I think you could do more however in being cognizant and sensitive to the values of this group. If you did you would probably find people more receptive to your ideas.

You may not want to release your code under OSS, thay is fine. But there are a lot of people here that really value and are passionate about OSS, and that invest a lot of effort to actually grow the node ecosystem through OSS. Really it is about helping each other be successful imo.

You may not value it in the same way, but you can still respect that others do.






Sent from my Windows Phone

---

From: Dennis Kane
Sent: 4/26/2012 10:53 AM
To: nod...@googlegroups.com
Subject: Re: [nodejs] Re: Your thoughts wanted on a node AI project

[Isaac Schlueter]

On Thu, Apr 26, 2012 at 10:53, Dennis Kane <dka...@gmail.com> wrote:
> Let's all remember what open source is really all about.  A program is
> called closed source if it is distributed in binary format only.  The open
> source movement makes the demand that one cannot distribute a binary program
> using, for example, GPL'ed code without also making the source code
> available.
>
> But a server side process has nothing to do with any of this!

The GPL and BSD/MIT licenses are very different. What you describe is copyleft.

There's also the LGPL, which *may* be linked against or depended upon
by non-GPL programs, but may *not* be extended or used to create
non-GPL derivative works.

And there's AGPL, which states that you may not use the software in
network programs unless all those who use the software over a network
*also* have access to the source. Ie, it's like GPL, with a broader
definition of "distribution".

The MIT and BSD licenses are much more liberal, and as far as I've
ever seen, they're pretty much equivalent. They are not copyleft, and
not viral. GPL is much less popular in the node community than MIT
and BSD. The Apache license is similar in intent to MIT and BSD, but
with additional language regarding copyrights an patents.

All I'm saying is: saying "the point of open source" is like saying
"the goal of american politics". There are a lot of different
conflicting goals, and groups, and ideologies. Lumping it all into a
single bucket loses a lot of details. In fact, many in the Free
Software movement would object to even being associated with the term
Open Source at all, since to them, it's more about freedom to modify
and extend than having access to code, which is viewed as merely a
means to an end.

> I've already been told more than once to rethink/change my approach.
>  Seriously?  I mean, let's get serious here.  The only time I get snarky is
> when people get pushy with their demands to just hand out something that has
> resulted from years of torment.  This just is not any old program.  It just
> isn't.  Period.

There's a pattern you're matching here, which I'm sure many of us have
seen many times in software communities.

"I have a really awesome killer idea. This is going to change
everything! No, you can't see it. But it's going to be amazing.
I've put *so much* time and effort into it. I've been slaving away
for years. If you come join me (which you will have to be very
motivated, lucky, and exceptional for me to allow you to do), then
untold riches will be yours to share!"

The pattern is:

1. Vague promises of paradigm shifting software of epic proportions.
(The "Noble Ambition")
2. Self-congratulatory claims of effort and sacrifice that went into
its production. (The "Quest")
3. A promise to share with a special few who are up to the challenge.
4. An resistance to sharing *any* relevant details. (When pressed,
reacting with indignant objections regarding principles and
propriety.)

I've taken to think of this as the "Genius Martyr" syndrome. It
doesn't always present with crippling insanity, but that is a common
pathology. It isn't always a technical idea, either; sometimes it's a
new way to handle governance in the group, or a revolutionary business
plan (where they're refusing to take any investment), etc.

It may be that there may come a time when some GM exhibits this
pattern in a technical community, and does in fact have some awesome
thing. Perhaps a few GMs have ideas which are actually worth more
than the paper they refuse to print them on.

If that time comes, if that idea shows up, and it's presented this
way, I'll probably miss out on it. Why? Because in my years in
software communities, 100% of the time that this pattern has presented
(and it is sadly not rare at all), the result has been the same.

A few people react with interest, which then of course leads to more
hinting. Finally people start pushing for proof, at which time the GM
lashes out defensively, claiming that the open source community is a
bunch of greedy hippies who want something for nothing, or questioning
the commitment or competence of those in the community to actually
contribute to the Noble Ambition.

In most cases, a flame war erupts and then the GM goes away
eventually, or sticks around making trouble until they're eventually
banned (or the community just gots to crap, which is sadly very
common.) In a few cases, someone in the group will bait the GM by
feigning interest, attempting to draw out the crazy for their
entertainment.

I'm sure that many of us have *been* a GM at one point or another in
our lives. I know I have. It was embarrassing. I survived.


As a person heavily invested in the Node.js community, I really don't
want to see it devolve into pointless bickering and personal insults.
We just can't have that. It's not good for anyone. It doesn't lead
to creativity or good will, and it tends to keep out the most
productive potential members.

So, Dennis, if you are serious, and you actually have a great idea,
you surely must realize that you need to go create something
*tangible* and *visible* before anyone will take it seriously.
Because, regardless of how talented, intelligent, insightful,
passionate, driven, or *right* you may in fact be, as long as you
*look* like a crazy person, people will think you're crazy. The more
you scream "TAKE ME SERIOUSLY", the less they will. That's just how
it works.

If your goal is to generate publicity for your project, or to recruit
others to your cause, then I think that the facts speak for
themselves, that this is not effective. You can curse the world for
failing to recognize your genius, but that won't make the world
recognize your genius.

> The only time I get snarky is
> when people get pushy with their demands to just hand out something that has
> resulted from years of torment.  This just is not any old program.  It just
> isn't.  Period.

If that's the only time you get snarky, then you should not
participate on this mailing list, because it is common practice here
to demand and expect that bold claims come with compelling evidence,
and excessive snark is not acceptable. No one cares how much torment
went into your program. They care how much value comes out of it.
Period.

[Duncan Gmail]

+1

[Dennis Kane]

Jesus, dude, thanks for the psychoanalysis. Anyone reading this
thread can see what turned into a slow moving train wreck. I posted
my thing and some people said: Cool! In the first post, I said
something about eventually doing an open source thing. Someone said:
"I want the source!" I said, "Haha you probably don't really want
it". People started getting serious about the holy node ecosystem and
chastising me for my tone. I started getting bitchy cause I found
myself having to defend myself about my politics (or rather my lack of
caring) rather than having fun with the issues that I really enjoy.

Haha whatevs...

Anyway Isaac, tl;dr.

Yes I'm trying to do something big. Big things are the only ones that
interest me. I'm trying to change the world, sorry. I have
psychological issues, so what? But I can program, dammit! Anyone
that wants to learn and help out is welcome. You have my site and my
email is on there.

I'm assuming that no one here is interested in NatLang issues or how
to implement them in Javascript?!


On Apr 28, 9:31 am, Isaac Schlueter <i...@izs.me> wrote:

[Isaac Schlueter]

On Sat, Apr 28, 2012 at 10:32, Dennis Kane <dka...@gmail.com> wrote:
> I'm assuming that no one here is interested in NatLang issues or how
> to implement them in Javascript?!

On the contrary, everyone's pretty interested. That's why they asked
you to show it off.

As you said yourself, previous explorations in this area have been
somewhat underwhelming. If you're not going to share anything about
your approach that differentiates it, from any others, then why should
anyone think it's any different?

A lot of people say they are going to change the world. Almost none
of them do. We're just playing the odds here.

I wasn't trying to psychoanalyze you, or claim that you're crazy. I
don't have nearly enough data or expertise to do that. I'm just
telling you that what you're doing matches a pattern, so it's a fair
guess that that's how people are interpreting it, accurately or not.

So, break the pattern. Do a different thing. Back up your bold
claims with compelling evidence. Don't ask the world to come to you,
go out and make something that *proves* you're worth taking seriously.
Don't be just another martyred genius.