EFTA00763967.pdf

From: on behalf of Ben Goertzel <ail To: Jeffrey Epstein Subject: Progress on my "How to Build AGI" book: good but still a bit more to be done; can you help? Date: Fri, 19 Feb 2010 18:35:52 +0000 Attachments: bbm_toc.pdf; book2.pdf; bookl.pdf Hi Jeffrey, As you may recall when you funded me with $50K in the early fall, I said I would use the money to finish my book on "how to build an AGI" within a 6 month period. I "wasted" a month of that time writing a research proposal and trying to sell it to "AI big shots", as you know... but even so I made pretty good progress. But III not done yet, and I could use some help during the last stages... Attached you'll find what I have so far. It's first-draft and hence NOT suitable to be sent to AI experts to read; III just sending it to you confidentially as an indication of my progress. I've attached a file with just the table of contents if you just want to see the scope of the effort. It's about 750 pages and has been split into 2 volumes. The first drafts of the two volumes are also attached, though I **don't** recommend you to read them yet, but rather to wait till a more final draft is available. (This draft probably still has some errors in equations, and lacks many needed figures, etc.) What needs to be done now is: my co-authors and I need to check over everything very carefully, correct errors and insert better explanations; and, I need to deal with all the figures and references, which is a huge job. I think I can bring the book to a reasonable state of completion in 3 more months of work. This may seem like a long time, but dealing with 750 pages of complex information is just plain time-consuming ... and stuff like "checking for mistakes" has to be done painstakingly, it can't really be hurried. So, unsurprisingly, III wondering if be willing to fund the completion of the book. My personal burn rate remains $8-9K/month as it has been for a long time, due to alimony and 2 kids in college, etc. III very appreciative of the funding you've provided so far and I hate to keep asking for more, but my everyday reality is what it is ;-p Now some comments on how the book relates to those evaluation letters that you didn't like... Those evaluation letters that I got were written by people who didn't really understand my ideas. They got the "general picture" of my AGI design and wrote the most positive letters they could based on that general picture. I thought that, given this situation, the letters were as positive as could be expected ... so I was happy with the letters; but very unfortunately it seems you had a different opinion ;-( The only really surefire way to convince "AI big shots" of the viability of my approach would be to build it and show them the working system. But of course, by that point, I wouldn't need the recommendations of those people anyway -- because a "working AGI system" would enable me to bring in $$ from all sorts of sourses. However, a second-rate way to try to convince these folks my approach is good would be to have them read the DETAILS of my approach, rather than just a brief high-level overview proposal. In other words: if I **finish** my book, and get these AI big shots to **read** it, then I have little doubt that their evaluation of my work will **improve** significantly. Whether it will improve significantly enough to motivate you to fund my work, remains unclear, of course. But at least, in this case the "big shots" would be offering opinions based on more understanding of what III suggesting. Of course, their intuitions EFTA00763967 about what kind of approach is best might STILL differ substantially from mine, though.... (After all, if these big-shot guys had the right intuition about AGI, they would have created one already, since they have a lock on all the US gov't AI funding... so it's kind of obvious from their (lack of) results that their AGI intuitions are somewhat flawed...) Of course I would MUCH MUCH MUCH rather see you actually fund my research project!!! The more I work on the book, the more frustrated I am that III not, instead, leading a team building the damn system!!!! But, if you aren't convinced enough to do that at present, it would be great if at least be willing to fund me to put in the few more months needed to finish the book.... Publishing the book via some major scientific publisher (Springer, World Scientific, whatever) shouldn't be a problem... I will finish the book even without your additional funding; but it might take 9-12 months rather than 3 months due to needing to spend my time on income generation instead of the book.... BTW III still debating moving to China but I haven't actually gotten the job offer from the Chinese university yet ... and even if that happens I wouldn't move till late August. Also I applied for some US university positions, but that option is not as appealing because I don't have funding to pay students to work on my AGI system here; whereas the Chinese gov't, if they offer me a job, will also offer funding for students. [China, unlike the US, realizes that it's really stupid to make top scientists spend 70% of their time writing proposals and fighting over money ;-p]. It seems likely I will make a decision about the "China option" in March or April. As always feel free to call me (240 505 6518) or arrange a F2F meeting if you want to discuss in more depth... I will continue to press on along the research path I KNOW is correct ... the risk is not that my approach is wrong, but rather that due to lack of funding I'll get old and die before I manage to complete the work ;-p Thanks.. Ben Ben Goertzel, PhD CEO, Novamente LLC and Biomind LLC Director of Applied Research, Singularity Institute for AI Chairman, Humanity+ External Research Professor, Xiamen University, China "Before something is a breakthrough, it's a crazy idea" -- Peter Diamandis EFTA00763968