Ben, On Tue, Dec 25, 2012 at 11:46 PM, Ben Goertzel <b...@goertzel.org> wrote:
> Steve, > > > 1. Can you roll your ideas into a patent and make some new IP that might > > form the nucleus for a fundable entity? > > Software patents aren't usually defensible these days, unless you're a > big company with big bucks to defend them... > The trick is to widely publish the IP, expect widespread infringement, and partner with the second largest infringer to sue their infringing competition. > > And AI is extra tough to patent defensibly, since there are always > multiple ways to do any particular thing > This is where the art of claims drafting enters the picture. If there is nothing new and important in there, you should probably be working on something else. > > I do think that getting patents can be valuable as part of an overall > strategy for getting VC $$, though; since many VCs are still foolish > enough to be psyched about software patents ;p > Indeed, it would be straightforward to identify dozens of patentable > technologies within the OpenCog design. But most of these have been > published online for a few years, hence are probably not patentable > anymore... > Right, but I suspect that you have some new wrinkles that you haven't talked about yet. As you might now be realizing, publishing everything you come up with can doom your own efforts. BTW, you have ONE YEAR from first publication to patent, so you might consider what you have published in recent months. > > > 2. What is the simplest valuable application of your technology that > > could be sold, e.g. semi-intelligent avatars? Self-adapting street > lights? > > Etc? If you could produce a patent aimed at such an application, you > might > > be able to turn that into R&D money to create a better product, and use > that > > income to fund greater developments. > > I'm already trying this sort of thing. Actually I've been spending > the bulk of my time for the last year, working on a Hong Kong based > financial prediction application of various AI technologies (including > some OpenCog components, plus other proprietary stuff my colleagues &I > invented). You should read the recent "In re Bilski" decision, where a patented investment strategy was ruled as not patentable. This is likely to limit financial patents, though probably not as much as you might think. If this works well, then within a number of years I will > have enough personal $$ to fund OpenCog AGI myself. This financial > prediction enterprise is angel investor funded, and I'm one among > several founding equity holders. This has been a rather fun project. > (Also note, this is an area where one doesn't patent one's proprietary > stuff; one just retains it as trade secrets...) Why not. Business process patents are the big thing these days. It is AMAZING what, with some legal creativity, can be patented these days. > I am quite hopeful > for its success, but of course every business has its own > uncertainties... > Since only ~10% of startups succeed, I make up for this in numbers. > > The downside of this approach, from an AGI perspective, is: Due to > spending so much time in the last year on financial prediction, I > haven't been able to guide the OpenCog work nearly as closely as I > would like. So the OpenCog HK team has gone more slowly than would > have been the case, if I'd been able to devote more of my time to > OpenCog in the last year... > Then, stop wasting your time answering Matt's postings!!! > > > 3. OpenCog, etc., might be a valuable foil around which you could form a > > think tank to at least identify the bounds around the "secret sauce", > > identify alternative approaches, identify approaches to make the > structure > > more or less impervious to new developmen,ts, etc. With this, you might > be > > able to get the DoD, etc., to chip in some money. > > That's an interesting idea, but given my response to point 2 above, I > don't have time for it... > The trick here is to get OTHERS to do the work. > > > 4. IMHO you can't possibly succeed alone. What you are doing is WAY too > > big for one person to do, just like Babbage couldn't build his machine. > You > > need to accumulate some human resources, e.g. a cooperative patent > attorney, > > a PR guy, a CEO type, an angel investor to pay for patent applications, > > etc., who are willing to work for a piece of the action. This also means > > that YOU need to give away some of the future action. You might be > surprised > > how cooperative some of the naysayers become when they realize that if > they > > turn their negativity into positive help, that they might actually get a > > percent of the result. > > As noted, I'm not alone in OpenCog by any means. But I agree, OpenCog > could badly use something more than "a fraction of Ben" > on the business / publicity side... > > I have a phone call tomorrow with someone who wants to volunteer to > help with PR and fundraising for > OpenCog. So I am exploring options in these regards, with what > little time I can allocate to such > exploration... > Great. Start shedding some of the load onto others. The downside of this approach is that you become incredibly committed - people's futures then depend on your performance. When you hit a technical snag, you must put out Herculean efforts to quickly overcome them to keep things on track. Like ladies, partners tend to wander off if not tended to regularly. > > In short, your advice makes sense, but -- with the exception of the > patent approach -- it's stuff I'm already exploring..... But as > options for accelerating the work are explored, time passes, and > meanwhile skeptics complain that progress is sooooo slow ;p ;) .. > You need SOME approach to bootstrap from zero (where you now are) up to an 8-digit monetary place where you can do what you know you need to do. My approach is via creating salable IP, which seems to necessarily require patents. Sure there are other approaches, but they must be nearly airtight. I have told my story to many people, how I intend to capture the market despite large intractable corporate competition. I have attended to every identified hazard. If my plan fails, it will be for some reason that no one has anticipated, and I have been doing a lot of anticipating. The "problem" I see with your stuff is that it appears to be FAR from airtight - as is evidenced by the population of persistent naysayers - sometimes even including me. When I ran into naysayers a few months ago when I started discussing my stuff here, I entered into extensive off-line discussions until I satisfied THEM that my plans would work, and along the say I made many tweaks in my plans to address the real and imagined hazards. Now, everyone agrees that it should work, though some believe that it is dreadful, unethical, destructive of the Internet, etc. Yes, there are parts of my patent application that address what I believe are non-existent hazards, but if I am going to sell this thing, I must cover ALL hazards, both real and otherwise. Though difficult, I think this is possible with your stuff. However, if you keep giving everything away, then forget it. My own personal suspicion is that the MAJOR breakthrough that is or should be in your work is a means to debug it. Your whole thesis that things can be twiddled into operation depends on being able to understand what is happening, what should be happening, and the difference between these two things. Simply having THAT secured would go a long way to making your plans more believable. Often, the real value isn't in the main approach, but in clever tools, debugging aids, validation methods, etc. Steve ------------------------------------------- AGI Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657 Powered by Listbox: http://www.listbox.com
