Re: [agi] AGI Consortium
Well, if one of us becomes extremely successful biz-wise, but the other has made some deep AI success, the one can always buy the other's company ;-) Hey! If I become both extremely successful biz-wise *and* make some deep AI success, can I give you the company and just make you pay me some decent residuals while I play with some more ideas? - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
RE: [agi] AGI Consortium
Certainly there are many ways to slay the beast. And the beast has many definitions. For an open source AGI you'd have to not throw in the kitchen sink, come up was a very basic design and maybe not tout how the thing is going to trigger a singularity? Maybe not try to replicate human brain functionality? I would do AGI as generalized AI. And I personally think that it is a model of specialized mathematics put into code to achieve this. But yeah there needs to be design impetus and if the design is not there at the beginning or is not adequate the project could churn indefinitely or just sputter. I do think OS AGI is doable and it may take fulltime funded individuals to do it. Software AGI will be larger than most OS applications. Seems to be certain thresholds on lines of code in open source projects. Someone should do a study and graph the number of lines of code to see the distribution across all projects. Not too many over a certain size. Also there should be a study on the complexity of the code. Not too many that are extremely complex. AGI will be complex, if not the most complex. Though there are a few large and very complex projects. Building very complex large projects without funds is rare. To be able to concentrate for months and years while paying the bills is tough, and doing it part time is challenging as well. But a person may be employed by an organization and be able to work on a separate AGI project. I have been in this position. The organization might only be able to dedicate one person to do this and other organizations may be in the same situation. There are many ways to achieve resources... OS AGI or yes even commercial AGI - large potential for failure. May be more advisable to put efforts into a project already far down the road? John From: YKY (Yan King Yin) [mailto:[EMAIL PROTECTED] I'm sympathetic to opensource, but I'm afraid an OS AGI is unlikely to happen. You'd need 1 or 2 person who has the background knowledge (requiring at least a couple years of intensive study) and the determination to start it, but most people who have gone that far, are unlikely to give it out for free. Someone can try, but I'd wait and see. I have started an AGI project before, and we were undecided whether to go OS or not. What happened was: 1. about 80+ people signed up 2. 20-30 have talked only once, and never spoke again 3. ~5 are regularly talking, discussing things with me 4. no one except me have contributed code 5. the project is now inactive I think the reason is simply that AGI is very hard and it's not easy to get people to understand what the project is doing. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
On 6/13/07, Benjamin Goertzel [EMAIL PROTECTED] wrote: YKY, I think there are two better schemes for collaboration than the one you've proposed: -- the traditional for-profit company with equity and options based compensation determined by a committee of trusted individuals -- the open-source project and various combinations thereof... I am interested in innovative organizational structures, and I appreciate your efforts brainstorming in this regard, but I don't really think you've yet come up with anything as good as the more traditional approaches, let alone better. I'm aware that OSS has worked better than most people thought it would beforehand though -- so that skeptical predictions about new forms of organization can certainly be proved wrong! But still I gotta call it as I see it... What Mark W has suggested seems less innovative than what you're grasping toward, but more clearly workable (excepting possible taxation issues). He wants to basically set aside a huge pool of shares in his company to be allocated at a much later date based on retrospectively determined technical/scientific contributions. I see no fundamental problem with this, except that it requires a LOT of trust in the part of the participants in the trusted managers. And as pointed out it may alienate potential investors. I think Eliezer's point is partly the simple one that if we create an AGI that launches the Singularity, this is a hell of a lot more interesting and important than who gets how much money or how much status for playing a role in creating it. In the context of creating a Singularity-enabling AGI, **control** is a lot more important and interesting an aspect than financial or social credit. On the other hand, if one creates an AGI that does not launch a Singularity but does interesting sub-Singularity-level stuff, then credit of various sorts may of course be valuable. And in some AGI projects, in the early stages, it may not be entirely clear whether the project really has Singularity potential or just valuable-sub-Singularity-AGI potential. I think you've done a very nice re-cap of what we've said so far, thanks. The problem is that right now I'm not joining Novamente because I have some different AGI ideas that you may not be willing to accept. And I don't blame you for that. If I were to join NM, I'd like to make significant modifications to it, or at least branch out from yours and to explore my favorite ideas. And that, in the context of a conventional company, with the accountability to investors, etc, is just very unfeasible. Therefore I'm NOT blaming you. [ In fact I'm very grateful that you've given me a lot of opportunities even though I'm a nobody, but this is personal. ] Yet if I were to join NM using mostly your AGI ideas, I'd feel very unhappy and also feel that I'm not doing my best. That why I'm trying to find a way out of this. You may find my ideas naive, but there's also inflexibility on your part. You may think you're just doing your business and you don't need to care about an external person pesting you, but that's not optimal. A business is always operating within a business landscape with other competitors. It is really the job of a leader to deal with that landscape as a whole, not just the company itself. In other words, if you cannot efficiently organize the people, then those who cannot find a place in your company, like me, would end up being your competitors, simply because they have nothing better to do. And that's a senseless waste of resource, among other things. We're still brainstorming I guess... YKY - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
On 6/14/07, John G. Rose [EMAIL PROTECTED] wrote: Even if YKY was to succeed in coming up with a new hybrid organizational structure, which could likely happen as there are definitely opportunities for innovation judging by existing structures, there still needs to be the traditional open source project that everybody and his brother(or sister) can go in and get a piece of without any strings attached. You go online, search for AGI, find it on SourceForge or similar, download and compile. If you like it enough you start working on it submitting changes and code improvements. With AGI there should be mechanisms i.e. forums and email lists, like this one, where non-developers contribute and work on it since a large amount of priceless input comes from learned individuals in various related fields. There does have to be a couple movers and shakers to get the thing started, just building up a skeleton and framework is a huge amount of work. Hi John, I'm sympathetic to opensource, but I'm afraid an OS AGI is unlikely to happen. You'd need 1 or 2 person who has the background knowledge (requiring at least a couple years of intensive study) and the determination to start it, but most people who have gone that far, are unlikely to give it out for free. Someone can try, but I'd wait and see. I have started an AGI project before, and we were undecided whether to go OS or not. What happened was: 1. about 80+ people signed up 2. 20-30 have talked only once, and never spoke again 3. ~5 are regularly talking, discussing things with me 4. no one except me have contributed code 5. the project is now inactive I think the reason is simply that AGI is very hard and it's not easy to get people to understand what the project is doing. YKY - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
Hi YKY, The problem is that right now I'm not joining Novamente because I have some different AGI ideas that you may not be willing to accept. And I don't blame you for that. If I were to join NM, I'd like to make significant modifications to it, or at least branch out from yours and to explore my favorite ideas. And that, in the context of a conventional company, with the accountability to investors, etc, is just very unfeasible. Therefore I'm NOT blaming you. Actually, for unpaid volunteers (working under non-D's) to fork the NM codebase in a different direction is not really problematic from NM's point of view. However, it may be problematic from the volunteers' point of view, because NM would still own the forked code. But if the forked code demonstrated superiority to the main branch, then the volunteers who created it would get a lot of NM options, and their forked branch would become the main branch! I don't think this is terribly likely to happen, but there are no psychological or organizational barriers to it on NM's side. Yet if I were to join NM using mostly your AGI ideas, I'd feel very unhappy and also feel that I'm not doing my best. That why I'm trying to find a way out of this. You may find my ideas naive, but there's also inflexibility on your part. You may think you're just doing your business and you don't need to care about an external person pesting you, but that's not optimal. I agree that the traditional corporate structure is not optimal. But I feel it's good enough to use as a vehicle for creating an AGI. The key thing is having a workable AGI design, not having the perfect organizational structure surrounding it. (Perfection and optimality are hard to come by in human systems, since we humans are so bloody imperfect!!) A business is always operating within a business landscape with other competitors. It is really the job of a leader to deal with that landscape as a whole, not just the company itself. Of course ... and NM's business path is all about trying to cope with the whole ecology of the modern biz world. We plan to proceed in large part via partnerships with other, more customer-focused firms, as evidenced by our current partnership with Electric Sheep Company. I really don't think competition among AI firms is a big problem at the moment. The market for AI is largely untapped and there is room for many players with solutions that have complementary strengths weaknesses. At this stage, I feel, if any AI firm succeeds, it benefits all AI firms, by increasing the reputation and legitimacy of AI in the marketplace. In other words, if you cannot efficiently organize the people, then those who cannot find a place in your company, like me, would end up being your competitors, simply because they have nothing better to do. And that's a senseless waste of resource, among other things. Well, if one of us becomes extremely successful biz-wise, but the other has made some deep AI success, the one can always buy the other's company ;-) -- Ben - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
I hardly think that's matter given that it's a truly a Singularity-class AI. Do you sit around calculating which of your grandparents deserves the most credit for bringing you into being? No, you take care of them as they need it. Thank you too, Josh -- maybe I was too cynical in thinking that that wouldn't be a major selling point (since anyone who truly got it would already be taking it as a given anyways). - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
A successful AI could do a superior job of dividing up the credit from available historical records. (Anyone who doesn't spot this is not thinking recursively.) Yay! Thank you! ( . . . and to think that last night I decided to give up on the topic. But don't worry, I'll still punt on it. It's reached the point of diminishing returns. :-) - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
YKY, I think that I'm going to take this opportunity to give up on this conversation for the following reasons: Come on, there're no obvious reasons for this complex issue. I have to disagree. There *ARE* certain things that really should be obvious if you get it. To put it another way, I'm not going to be the boss under this system of collaboration. I'm just going to be just another contributor among many. This is a scheme where we can work together without someone dominating anyone else a priori. You clearly also don't get that *no one* with a real clue desires to be the boss. Being the boss (when done correctly) is a duty, not a personal benefit. The only time I want to be the boss is a) to prevent bad things from happening and b) when everyone else has fulfilled -- or is going to fulfill -- their responsibility and it's my turn to do my share. Given my personal druthers, I'd just as soon be an average worker without the added responsibilities. And if you have to be the boss to get enough money to make you happy, then that just proves that you really don't have a clue To organize average people to work together you have to give rewards. Finally, I really, *really* don't believe this either (unless you want to insist that the satisfaction of a challenge met or a job well done -- or the warm fuzzy that you get when you help someone -- are rewards). You don't do much charity, do you . . . . - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
If you're not proposing a better scheme for collaboration, and you criticize my scheme in a non-constructive way, then effectively you're just saying that you're not interested in collaborating at all. And that's kind of sad, given that we're still so far from AGI. YKY, I think there are two better schemes for collaboration than the one you've proposed: -- the traditional for-profit company with equity and options based compensation determined by a committee of trusted individuals -- the open-source project and various combinations thereof... I am interested in innovative organizational structures, and I appreciate your efforts brainstorming in this regard, but I don't really think you've yet come up with anything as good as the more traditional approaches, let alone better. I'm aware that OSS has worked better than most people thought it would beforehand though -- so that skeptical predictions about new forms of organization can certainly be proved wrong! But still I gotta call it as I see it... What Mark W has suggested seems less innovative than what you're grasping toward, but more clearly workable (excepting possible taxation issues). He wants to basically set aside a huge pool of shares in his company to be allocated at a much later date based on retrospectively determined technical/scientific contributions. I see no fundamental problem with this, except that it requires a LOT of trust in the part of the participants in the trusted managers. And as pointed out it may alienate potential investors. I think Eliezer's point is partly the simple one that if we create an AGI that launches the Singularity, this is a hell of a lot more interesting and important than who gets how much money or how much status for playing a role in creating it. In the context of creating a Singularity-enabling AGI, **control** is a lot more important and interesting an aspect than financial or social credit. On the other hand, if one creates an AGI that does not launch a Singularity but does interesting sub-Singularity-level stuff, then credit of various sorts may of course be valuable. And in some AGI projects, in the early stages, it may not be entirely clear whether the project really has Singularity potential or just valuable-sub-Singularity-AGI potential. -- Ben G - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
To organize average people to work together you have to give rewards. Finally, I really, *really* don't believe this either (unless you want to insist that the satisfaction of a challenge met or a job well done -- or the warm fuzzy that you get when you help someone -- are rewards). You don't do much charity, do you . . . . And of course: To create an AGI using a team average people is very likely to be a losing proposition!!! It's hard enough to create a database-back-end/web-front-end standard software system using average programmers, who are smarter than average people !!! I do believe that creating an AGi requires a small core time of highly gifted and dedicated and cooperative people... that is what I have tried to build at Novamente... -- Ben - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
Overall measures are per-module as well, so a basic DB-access module would only get 1% to distrubute to all its lines of code, as it has little originiailty and is well known code. Samantha Atkins [EMAIL PROTECTED] wrote: No LOC based credit please. That measure is totally bogus. Ten lines of beautifully crafted spot-on code can be more important than a 1000 lines of more ordinary code. The real measures are pretty subjective and the quality of the measure is utterly dependent on the quality and insight of the measurer. Any sort of averaging out of the measuring/measurers will average out the quality of the measure. - samantha On 6/11/07, James Ratcliff [EMAIL PROTECTED] wrote:Even if they received credit for the 7,000 lines, it would be worth very little in the overall scheme, and any code that was not good could be marked as too be fixed or optimized fairly easily, (similar again to the Wiki markups) to where that credit could be diminished... and any obvious spam or dragging out fo larger code would be removed. Also a time delay could be in place, so no credit is applied until 3-5 people have looked over the code, and a month has passed by, so any new spammy code would fall thru the cracks. James - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?; ___ James Ratcliff - http://falazar.com Looking for something... - Building a website is a piece of cake. Yahoo! Small Business gives you all the tools to get online. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
RE: [agi] AGI Consortium
Even if YKY was to succeed in coming up with a new hybrid organizational structure, which could likely happen as there are definitely opportunities for innovation judging by existing structures, there still needs to be the traditional open source project that everybody and his brother(or sister) can go in and get a piece of without any strings attached. You go online, search for AGI, find it on SourceForge or similar, download and compile. If you like it enough you start working on it submitting changes and code improvements. With AGI there should be mechanisms i.e. forums and email lists, like this one, where non-developers contribute and work on it since a large amount of priceless input comes from learned individuals in various related fields. There does have to be a couple movers and shakers to get the thing started, just building up a skeleton and framework is a huge amount of work. John From: Benjamin Goertzel [mailto:[EMAIL PROTECTED] If you're not proposing a better scheme for collaboration, and you criticize my scheme in a non-constructive way, then effectively you're just saying that you're not interested in collaborating at all. And that's kind of sad, given that we're still so far from AGI. YKY, I think there are two better schemes for collaboration than the one you've proposed: -- the traditional for-profit company with equity and options based compensation determined by a committee of trusted individuals -- the open-source project and various combinations thereof... - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
On 6/12/07, Mark Waser [EMAIL PROTECTED] wrote: If you think my scheme cannot be fair then the alternative of traditional management can only be worse (in terms of fairness, which in turn affects the quality of work being done). The situation is quite analogous to that between a state-command economy and a free market (or actually identical?) I don't find the situation analogous at all and once again you haven't answered my direct questions about your proposed managerial board -- How do you intend to select this board? How do you intend to keep them honest? How is this truly different from my trustworthy owners $ Board members will be nominated and elected by the entire group, and hopefully we can find some academics who have reputation in certain areas of AI, and are not contributors themselves. I tend to think that they will be more judicious than other types of people. Also can you explain: 1. why you think that my scheme will lead to systematically incorrect estimates of contribution values? 2. why you don't see the analogy between a peer-estimated attribution system and a free market, versus a state-command economy and a CEO-directed company? YKY - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
Board members will be nominated and elected by the entire group, and hopefully we can find some academics who have reputation in certain areas of AI, and are not contributors themselves. I tend to think that they will be more judicious than other types of people. Again, how is that different from selecting trustworthy owners? 1. why you think that my scheme will lead to systematically incorrect estimates of contribution values? I think that it will lead to consistently (but variably) incorrect estimates (as opposed to systematically incorrect) because I don't see a group of humans being able to correctly assess the value of a contributions -- particularly before the results are in. Or to put it more simply, I don't see you as being qualified to judge Ben's code. 2. why you don't see the analogy between a peer-estimated attribution system and a free market, versus a state-command economy and a CEO-directed company? You keep putting up this strawman of a CEO-directed company. First off, CEO don't generally direct as thoroughly as you seem to believe. Second, CEOs certainly don't do every single evaluation by themselves -- normally they don't even do one so you're trying to draw an analogy with a single point of control system is nonsense. Finally, I'm not proposing a CEO-directed company so I don't see why you keep throwing this up. - Let me rephrase it this way -- There is *NO* difference between what you are proposing for a managerial board and what I'm proposing *except* that you could have a successful uprising against the managerial board and my set-up prevents that. For the purposes of day-to-day operations, they are IDENTICAL. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
On 6/13/07, Eliezer S. Yudkowsky [EMAIL PROTECTED] wrote: A successful AI could do a superior job of dividing up the credit from available historical records. (Anyone who doesn't spot this is not thinking recursively.) During the pre-AGI interim, people have got to make money and to enjoy life to various degrees (which depends on the individuals). Right now you're not solving the problem of organizing people to build an AGI and rewarding them -- without interim-term rewards, people's incentive may diminish, so that'd not be an optimal way of organizing people. Also, a future AGI cannot go back in time to reward us during the interim -- and we're talking about anything from 5 years (not very likely) to decades. Anyone who doesn't spot this is doing time travels. YKY - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
On 6/13/07, Eliezer S. Yudkowsky [EMAIL PROTECTED] wrote: I wouldn't bother working with anyone who was seriously worried over who got the credit for building a Singularity-class AI - no other kind matters. There are two reasons for this, not just the obvious one. Come on, there're no obvious reasons for this complex issue. And Singularity-grade is not a well-defined term. What I'm proposing is a way for *us* to collaborate and to bring about progress faster. I admit that my scheme is not perfect, but it seems that no one else is proposing a better one. If you're not proposing a better scheme for collaboration, and you criticize my scheme in a non-constructive way, then effectively you're just saying that you're not interested in collaborating at all. And that's kind of sad, given that we're still so far from AGI. To put it another way, I'm not going to be the boss under this system of collaboration. I'm just going to be just another contributor among many. This is a scheme where we can work together without someone dominating anyone else *a priori*. That said, let me add that your non-profit model is also feasible -- as charity and altruism will always be welcomed by everyone. But the problem is that not everyone is highly altruistic. To organize average people to work together you have to give rewards. YKY - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
Has anyone tried a test of something as simple as per line of code / function? Meaning that each function or module could have a % value associated with it (set by many users average rating) And then simply giving credit by line of code input. Anyone writing cruddy long code would initially have some credits, but as someone else rewrote it in better code the inflated original code and credit would disappear, and more credit would go to more worthy modules of the code. With a Wiki-style ability to easily see the history and changes, you could prevent abuse of the system as well. People could easily look over the code itself, and the modules, the style, without ever knowing exactly who did it, and rate it without rating the person behind it. James Ratcliff YKY (Yan King Yin) [EMAIL PROTECTED] wrote: On 6/11/07, Mark Waser [EMAIL PROTECTED] wrote: I'm going to temporarily ignore my doubts about accurate assessments to try to get my initial question answered yet again. Why wouldn't it be to my advantage to exaggerate my contributions? I'm sorry about the confusion. Let me correct by saying: it *is* to your advantage to exaggerate your contributions, but your peers won't allow it. The origin of the confusion is re point (B), which is not essential. Can I skip the explanation of that? 'Cause I don't want to make the scheme sound overly complicated. YKY - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?; ___ James Ratcliff - http://falazar.com Looking for something... - Boardwalk for $500? In 2007? Ha! Play Monopoly Here and Now (it's updated for today's economy) at Yahoo! Games. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
Has anyone tried a test of something as simple as per line of code / function? My first official programming course was a Master's level course at an Ivy League college. The course project was a full-up LISP interpreter. My program was ~800-900 lines and passed all testing with flying colors. The next smallest program was in excess of 7,000 lines with a number of people in the 10,000 to 13,000 range -- most of whom were not able to debug their problems with properly maintaining their environments. I believe that the key to truly effective programmers is that they know how to use levels of abstraction to minimize code (less code = less maintenance = less bugs = less mindshare, etc). The last thing that I want to do is *anything* that encourages people to write more code (even if it gets replaced later -- since it would still eat up mindshare until then). The only scheme that I'd possibly accept based on lines of code would be one where if someone else wrote a tighter program, the original writer would get negative credit (i.e. something like if they wrote 7,000 lines and I re-did it with 1,000 -- I get credit for half the difference for a total of 3,000 and they get credit for 1,000 minus half the difference for a total of minus 2,000 -- noting, of course, that if their initial code was relatively good and only 1,500 and I wrote 1,000, they would still get 750 while I only get 250). - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
On Monday 11 June 2007 12:12:26 pm Mark Waser wrote: ... The last thing that I want to do is *anything* that encourages people to write more code ... The classic apocryphal story is of the shop where they had this fellow who was an unbelievably productive programmer -- up until the day he discovered loops. Josh - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
Monday, June 11, 2007, Mark Waser wrote: MW The only scheme that I'd possibly accept based on lines of code MW would be one where if someone else wrote a tighter program, the original MW writer would get negative credit (i.e. something like MW if they wrote 7,000 lines and I re-did it with 1,000 -- I get credit MW for half the difference for a total of 3,000 and they get credit for MW 1,000 minus half the difference for a total of minus MW 2,000 -- noting, of course, that if their initial code was relatively MW good and only 1,500 and I wrote 1,000, they would still get 750 while I only get 250). It should be hopeless either way - flexibility of code may as well suffer from too tight implementation. Good balance just can't be estimated by a simple ruler :) -- Vladimir Nesovmailto:[EMAIL PROTECTED] - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
On 6/11/07, Mark Waser [EMAIL PROTECTED] wrote: I'm sorry about the confusion. Let me correct by saying: it *is* to your advantage to exaggerate your contributions, but your peers won't allow it. Cool. I'll then move back to my other point that is probably better phrased as I don't believe you (or any current human) can set up a system which is both systematic and reasonably fair under the constraints you've chosen. I also note that you've made several references to a managerial board. How do you intend to select this board? How do you intend to keep them honest? How is this truly different from my trustworthy owners (except from, of course, the manner in which they are selected -- or maybe not even then since maybe I'll just have a nomination and election process before I set up)? Programming is not the only skill needed for an AGI project, we need algorithm / architecture design, writing documentation, etc. All of these things can be termed mental work and their quality can *only* be judged by intelligent beings, which, pre-AGI, are human beings. In other words, the only choices we have for running a business are either to rely on peers or traditional management. If you think my scheme cannot be fair then the alternative of traditional management can only be worse (in terms of fairness, which in turn affects the quality of work being done). The situation is quite analogous to that between a state-command economy and a free market (or actually identical?) YKY - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
An additional idea: each member's vote could be weighted by the member's total amount of contributions. This way, we can establish a network of genuine contributors via self-organization, and protect against mischief-makers, nonsense, or sabotage, etc. YKY - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
Keep going ... won't be too long until you invent fungible tokens for your people that act as a medium of exchange, a store of value, and a unit of account. On Monday 11 June 2007 07:22:46 pm YKY (Yan King Yin) wrote: An additional idea: each member's vote could be weighted by the member's total amount of contributions. This way, we can establish a network of genuine contributors via self-organization, and protect against mischief-makers, nonsense, or sabotage, etc. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
On 6/9/07, Mark Waser [EMAIL PROTECTED] wrote: Think: if you have contributed something, it'd be in your best interest to give accurate estimates rather than exaggerate or depreciate them Why wouldn't it be to my advantage to exaggerate my contributions? But your peers in the network won't allow that. YKY - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
Obviously innovation comes from all walks of life, be they opensource or commercial people. But some entrepreneurs are more capable of appropriating their inventions, eg Edison did *not* invent the light bulb, but he got famous for commercializing and patenting it. Many people simply don't have the time and energy to commercialize their inventions, and I bet this happens in the software world too. Having a consortium may allow more people to get rewarded for their innovative ideas. YKY - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
YKY Think: if you have contributed something, it'd be in your best interest to give accurate estimates rather than exaggerate or depreciate them MW Why wouldn't it be to my advantage to exaggerate my contributions? YKY But your peers in the network won't allow that. That is an entirely different argument (and one that I'm not willing to concede since I don't believe that the network can accurately prevent exaggeration without accidentally short-changing some decent percentage of contributors). Your statement was if you have contributed something, it'd be in your best interest to give accurate estimates rather than exaggerate or depreciate them. I would like an explanation of why part of your statement is true. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
On 6/10/07, Mark Waser [EMAIL PROTECTED] wrote: YKY Think: if you have contributed something, it'd be in your best interest to give accurate estimates rather than exaggerate or depreciate them MW Why wouldn't it be to my advantage to exaggerate my contributions? YKY But your peers in the network won't allow that. MW That is an entirely different argument (and one that I'm not willing to concede since I don't believe that the network can accurately prevent exaggeration without accidentally short-changing some decent percentage of contributors). Your statement was if you have contributed something, it'd be in your best interest to give accurate estimates rather than exaggerate or depreciate them. I would like an explanation of why part of your statement is true. YKY Sorry it's a bit complicated... A) First, there's the problem of estimating how much a *particular piece*of contribution is worth. As I've explained before, this will be done by self-rating + optional peer-rating + etc... and I think it will give acceptable estimates on average. You said: *I don't believe that the network can accurately prevent exaggeration without accidentally short-changing some decent percentage of contributors*. I don't understand what exactly you mean by short-changing. Can you give an example? Are you saying that a decent percentage of contributors would recieve *systematically incorrect* crediting? I wonder why. B) Secondly, there's the problem of someone wanting to check out while taking some other members' contributions along (note: one is always free to use one's *own* ideas / contributions elsewhere). In that case we need to estimate how much shares the new project owes the consortium. Maybe we'll rely on the managerial board for that. But the most important part is (A). My statement that you questioned above, is re estimating X% = $C / $C + $c, which is my earlier idea for solving (B). But it's not essential to the scheme, and it seems overly complicated, so we may simply ask the managerial board to make an estimate instead. YKY - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
I'm going to temporarily ignore my doubts about accurate assessments to try to get my initial question answered yet again. Why wouldn't it be to my advantage to exaggerate my contributions? - Original Message - From: YKY (Yan King Yin) To: agi@v2.listbox.com Sent: Sunday, June 10, 2007 4:13 PM Subject: Re: [agi] AGI Consortium On 6/10/07, Mark Waser [EMAIL PROTECTED] wrote: YKY Think: if you have contributed something, it'd be in your best interest to give accurate estimates rather than exaggerate or depreciate them MW Why wouldn't it be to my advantage to exaggerate my contributions? YKY But your peers in the network won't allow that. MW That is an entirely different argument (and one that I'm not willing to concede since I don't believe that the network can accurately prevent exaggeration without accidentally short-changing some decent percentage of contributors). Your statement was if you have contributed something, it'd be in your best interest to give accurate estimates rather than exaggerate or depreciate them. I would like an explanation of why part of your statement is true. YKY Sorry it's a bit complicated... A) First, there's the problem of estimating how much a particular piece of contribution is worth. As I've explained before, this will be done by self-rating + optional peer-rating + etc... and I think it will give acceptable estimates on average. You said: I don't believe that the network can accurately prevent exaggeration without accidentally short-changing some decent percentage of contributors . I don't understand what exactly you mean by short-changing. Can you give an example? Are you saying that a decent percentage of contributors would recieve *systematically incorrect* crediting? I wonder why. B) Secondly, there's the problem of someone wanting to check out while taking some other members' contributions along (note: one is always free to use one's *own* ideas / contributions elsewhere). In that case we need to estimate how much shares the new project owes the consortium. Maybe we'll rely on the managerial board for that. But the most important part is (A). My statement that you questioned above, is re estimating X% = $C / $C + $c, which is my earlier idea for solving (B). But it's not essential to the scheme, and it seems overly complicated, so we may simply ask the managerial board to make an estimate instead. YKY -- This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?; - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
On 6/11/07, Mark Waser [EMAIL PROTECTED] wrote: I'm going to temporarily ignore my doubts about accurate assessments to try to get my initial question answered yet again. Why wouldn't it be to my advantage to exaggerate my contributions? I'm sorry about the confusion. Let me correct by saying: it *is* to your advantage to exaggerate your contributions, but your peers won't allow it. The origin of the confusion is re point (B), which is not essential. Can I skip the explanation of that? 'Cause I don't want to make the scheme sound overly complicated. YKY - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
I'm sorry about the confusion. Let me correct by saying: it *is* to your advantage to exaggerate your contributions, but your peers won't allow it. Cool. I'll then move back to my other point that is probably better phrased as I don't believe you (or any current human) can set up a system which is both systematic and reasonably fair under the constraints you've chosen. I also note that you've made several references to a managerial board. How do you intend to select this board? How do you intend to keep them honest? How is this truly different from my trustworthy owners (except from, of course, the manner in which they are selected -- or maybe not even then since maybe I'll just have a nomination and election process before I set up)? - Original Message - From: YKY (Yan King Yin) To: agi@v2.listbox.com Sent: Sunday, June 10, 2007 4:52 PM Subject: Re: [agi] AGI Consortium On 6/11/07, Mark Waser [EMAIL PROTECTED] wrote: I'm going to temporarily ignore my doubts about accurate assessments to try to get my initial question answered yet again. Why wouldn't it be to my advantage to exaggerate my contributions? I'm sorry about the confusion. Let me correct by saying: it *is* to your advantage to exaggerate your contributions, but your peers won't allow it. The origin of the confusion is re point (B), which is not essential. Can I skip the explanation of that? 'Cause I don't want to make the scheme sound overly complicated. YKY -- This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?; - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
Think: if you have contributed something, it'd be in your best interest to give accurate estimates rather than exaggerate or depreciate them Why wouldn't it be to my advantage to exaggerate my contributions? - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
On 6/9/07, Mark Waser [EMAIL PROTECTED] wrote: ...Same goes for most software developed by this method–almost all the great open source apps are me-too knockoffs of innovative proprietary programs, and those that are original were almost always created under the watchful eye of a passionate, insightful overseer or organization. Obviously the author hasn't bothered looking at many open source projects. There are swaths of innovative usable open source projects. The thing is, they are often not noticed, because innovative does not necessarily mean useful to everyone who owns a computer. I'm also more convinced that the opposite is true: open source innovation leads to commercial knock-offs. e.g. iTunes is a piece of crap in comparison to amarok. J - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
On 6/8/07, Mark Waser [EMAIL PROTECTED] wrote: You are never going to see a painting by committee that is a great painting. And he's right. This was Sterling's indictment of Wikipedia–and to the wisdom of crowds fad sweeping the Web 2.0 pitch sessions of Silicon Valley–but it's also a fair assessment of what holds most (not all) open source enterprises back: Lack of vision. Every project has some developers recruitment policy; a smart mind is an integrated mind. The ideological divide goes between Open Knowledge and Source, and Closed Knowledge and Source. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
On 6/8/07, Mark Waser [EMAIL PROTECTED] wrote: Actually, it should be On 6/8/07, Mark Waser [EMAIL PROTECTED] quoted someone else as saying: I don't agree with Sterling's indictment of Wikipedia since I don't believe that a relatively unified vision is necessary for it. I do, however, agree with his belief that a kitchen sink mentality is extremely detrimental to most open sources enterprises -- and would likewise be a problem for the AGI consortium. I think that some sort of focus and *some* limitations on what is considered on-topic for the project is necessary for effectiveness. I'm tempted to assume that you agree with your a smart mind is an integrated mind comment. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
On Friday 08 June 2007 08:21:28 am Mark Waser wrote: Opening your project up to an unreliable parade of volunteer contributors allows for a great, lowest-common-denominator consensus product. That's fine for Wikipedia, but I wouldn't count on any grand intellectual discourse arising therein. Same goes for most software developed by this method-almost all the great open source apps are me-too knockoffs of innovative proprietary programs, and those that are original were almost always created under the watchful eye of a passionate, insightful overseer or organization. Firefox is actually Mozilla Firefox, after all. This is basically right. There are plenty of innovative Open Source programs out there, but they are typically some academic's thesis work. Being Open Source can allow them to be turned into solid usable applications, but it can't create them in the first place. Josh - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
On 6/8/07, J Storrs Hall, PhD [EMAIL PROTECTED] wrote: This is basically right. There are plenty of innovative Open Source programs out there, but they are typically some academic's thesis work. Being Open Source can allow them to be turned into solid usable applications, but it can't create them in the first place. Being Closed Source can't create them neither (just a note for the sake of completeness). - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
I noticed a serious problem with credit attribution and allowing members to branch outside of the mother project. For example, there may be a collection of contributions, from many members, that is worth $C in the consortium. Suppose someone decides to start an external project, then adding $c of new contributions to it, but where $c is unknown because it's outside the consortium's attributing system. Then, suppose the new project sells a million copies and earns $NetProfit. $NetProfit may be a very big amount because of leverage / nonlineae effects. The *fair* amount to pay back the consortium should be $PayBack = $NetProfit * ($C / ($C + $c)). Unfortunately, $c occurs outside the consortium and cannot be measured easily or consistently. Not being to estimate the $PayBack, the whole scheme is thrown into question. Also, using an approximate formula for $PayBack may not work since people will try to exploit the approximation to their advantage. It seems that allowing members to check out is very difficult, if not impossible, to manage. The only alternative that's left is to restrict members to participate and develop projects within the consortium *exclusively*... but this will turn off many people as they don't see why the consortium will win instead of other projects... and this is the problem faced by all AGI founders trying to recruit... YKY - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
Sure. Successful and innovative aren't the same thing -- in fact, they're often at odds. The best versions of something from the point of polish and usability generally come after lots of hard experience with its earlier versions. Bell Labs, where Unix came from originally, was very academe-like and Unix was the product of a small, focussed group. Windowing systems came from places like MIT-AI and Xerox PARC. Josh On Friday 08 June 2007 12:50:16 pm Samantha Atkins wrote: Apache and its various offshoots? Linux itself? KDE? JBoss and its subprojects? Hibernate? None of these came from some academic thesis work and all are wildly successful. So I do not agree with the characterization of Open Source. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
Really Open Source software projects almost never have a total open door policy on the contributions that are accepted. There is usually a small group that determines whether contributed changes are good enough and fit the overall project goals and architecture well enough. Wikipedia is one of the best innovations in information aggregation ever. I think many of us are very happy that it exists and use it extensively. It does work to filter wheat from chaff over time. Claiming most Open Source is me-too knock-offs is simply wrong. Apache and many of its subprojects took the market by storm because it is significantly better than the closed source solutions it replaced for one example among many. You understand that Mozilla is open source right? Most of the innovation we enjoy in Firefox today came long after Netscape days and long after lingering Netscape/AOL control. But I don't expect any great understanding about Open Source here. It is not the expertise or prime interest of the group. On 6/8/07, Mark Waser [EMAIL PROTECTED] wrote: from http://blogs.techrepublic.com.com/geekend/?p=696 Bruce Sterling: All blogs will die by 2018http://blogs.techrepublic.com.com/geekend/?p=696 - *Date*: June 5th, 2007 - *Blogger*: The Trivia Geek Security expert and tech curmudgeon Bruce Sterling famously quipped at this year's South-by-Southwest conference that I don't think there will be that many [blogs] around in 10 yearshttp://www.feed24.com/go/44185676. I think they are a passing thing. This got the blogosphere all a-twitterhttp://www.feed24.com/go/44185676(ahem), but I think enough time has passed that we can look past this ill-worded point from Sterling's SXSW rant and get to the real moneyline: You are never going to see a painting by committee that is a *great*painting. And he's right. This was Sterling's indictment of Wikipedia–and to the wisdom of crowds fad sweeping the Web 2.0 pitch sessions of Silicon Valley–but it's also a fair assessment of what holds most (not all) open source enterprises back: *Lack of vision*. Nearly all great innovation comes from a singular vision pursued doggedly until it achieves success. Apple is a great example of this, as the company didn't really resume its cutting-edge status (for better or worse) until Steve Jobs returned, and gave us the iMac and iPod (for better or worse). And say what you will about Microsoft, but it was Bill Gates singular vision for Windows and the software industry that drove his company to its excess…er, success. Opening your project up to an unreliable parade of volunteer contributors allows for a great, lowest-common-denominator consensus product. That's fine for Wikipedia, but I wouldn't count on any grand intellectual discourse arising therein. Same goes for most software developed by this method–almost all the great open source apps are me-too knockoffs of innovative proprietary programs, and those that are original were almost always created under the watchful eye of a passionate, insightful overseer or organization. Firefox is actually *Mozilla* Firefox, after all. -- This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?; - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
Well-said Samantha :-) On a different note: something YKY and Mark may want to read about a possible approach to running a new AGI consortium: eXtreme Research. A software methodology for applied research: eXtreme Researching vy Olivier Chirouze, David Cleary and George G. Mitchell (Software. Practice Experience 2005; 35:1441–1454 - try to get it from www.interscience.wiley.com). Some interesting ideas on building up research ideas prototypes systems from the ground up with a distributed group. Samantha Atkins [EMAIL PROTECTED] 06/08/07 7:01 PM But I don't expect any great understanding about Open Source here. It is not the expertise or prime interest of the group. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] AGI Consortium
On 6/8/07, YKY (Yan King Yin) [EMAIL PROTECTED] wrote: I noticed a serious problem with credit attribution and allowing members to branch outside of the mother project. For example, there may be a collection of contributions, from many members, that is worth $C in the consortium. Suppose someone decides to start an external project, then adding $c of new contributions to it, but where $c is unknown because it's outside the consortium's attributing system. Then, suppose the new project sells a million copies and earns $NetProfit. $NetProfit may be a very big amount because of leverage / nonlineae effects. The *fair* amount to pay back the consortium should be $PayBack = $NetProfit * ($C / ($C + $c)). Unfortunately, $c occurs outside the consortium and cannot be measured easily or consistently. Not being to estimate the $PayBack, the whole scheme is thrown into question. Also, using an approximate formula for $PayBack may not work since people will try to exploit the approximation to their advantage. It seems that allowing members to check out is very difficult, if not impossible, to manage. The only alternative that's left is to restrict members to participate and develop projects within the consortium *exclusively*... but this will turn off many people as they don't see why the consortium will win instead of other projects... and this is the problem faced by all AGI founders trying to recruit... After some thinking, there may be a solution to this problem, which is to let members estimate $c as well. In other words, let them estimate how much of a particular branch is completed as a percentage. So if someone wants to check out, he'll agree to pay with shares of his new project equal to that percentage. Being optimistic again, there're reasons to believe that most members will try to give fair and accurate estimations. (Think: if you have contributed something, it'd be in your best interest to give accurate estimates rather than exaggerate or depreciate them). Notice that this scheme is more likely to work with a high number of participants, but may fail when there are very few contributors to a particular branch -- but this is OK because *any* serious AGI project has to be large-scale anyway. YKY - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e