36 replies to this topic

[brokenportal]

Mind interviewed Ben Goertzl in the sunday evening update again yesterday. It occurs to me today, we have a way to get people to help with exposure for the cause through imminst and we have a way to get people to help with the science through MFURI. How though for agi? Why dont we help get an NMURI program going, "Nova Mente Undergraduate Research Initiative." Yes, no, would it work, would it not work?

[modelcadet]

Mind interviewed Ben Goertzl in the sunday evening update again yesterday. It occurs to me today, we have a way to get people to help with exposure for the cause through imminst and we have a way to get people to help with the science through MFURI. How though for agi? Why dont we help get an NMURI program going, "Nova Mente Undergraduate Research Initiative." Yes, no, would it work, would it not work?

I think this is a great idea. I would love to work through such a project; even if there weren't money involved. As Ben said though, a lot of the programming work requires 1337 coders. Undergrads (except maybe Joseph) can't really do that. Most programmers can't.

I think Undergrads can do quite a bit though. Marketing, non-profit and business administration, political advocacy, etc etc... And I believe undergrads can work on other, more technical areas, like friendliness, as I am doing... but all these things require beastly brains.

But yeah... I personally would love more access to luminaries like Ben and Eliezer. Well, Eliezer not quite so much... but still, I'd love to not have to complete mathematical feats in order to access him. I know they are all very busy. Perhaps we could start something with the help of Michael Vassar...

[brokenportal]

Maybe we could get Micheal Vassar on the sunday evening update. Any dialogue in this regard would be helpful in sparking this idea somewhere. Another thing is, why not an AGI prize like the mouse prize and x prize?

[Connor MacLeod]

Mind interviewed Ben Goertzl in the sunday evening update again yesterday. It occurs to me today, we have a way to get people to help with exposure for the cause through imminst and we have a way to get people to help with the science through MFURI. How though for agi? Why dont we help get an NMURI program going, "Nova Mente Undergraduate Research Initiative." Yes, no, would it work, would it not work?

I think this is a great idea. I would love to work through such a project; even if there weren't money involved. As Ben said though, a lot of the programming work requires 1337 coders.

I guess I'm pretty 1337.

But yeah... I personally would love more access to luminaries like Ben and Eliezer. Well, Eliezer not quite so much... but still, I'd love to not have to complete mathematical feats in order to access him.

What, does he require you to compute a posterior probability before answering a query?

Heh!!! After glancing at his website it appears that this is more or less what he wants:

I am not interested in your amazing new theory of Artificial Intelligence unless you can derive the formula for propagating updates in a Bayesian network.

That's seems to be a pretty low bar.

You do not have psychic powers.

No, but occasionally God lets me make inferences based on His prior...beats the heck out of Jeffrey's every time.

Edited by Connor MacLeod, 15 December 2008 - 08:00 AM.

[Heliotrope]

mfuri does biology, nfuri does computer sci , nice combo

[brokenportal]

Would you help support an agi undergraduate research initiative?

[Heliotrope]

Would you help support an agi undergraduate research initiative?

yes. like i said: "mfuri does biology, nfuri does computer sci , nice combo "


that's a 1-2 killer punch, cool boxing combo that deals a blow to our blight

[brokenportal]

I was writing that to connor at the same time that you were writing that response. In any case the question works well for anybody. So you would support it then, like, you would help contact universities and tell them about such a program and try to get them to get students in their clubs and stuff to sign up?

It would be a killer punch. Through chatting with so many people Im realizing there are 4 main paths that people choose from to help with the cause, agi, science, exposure and philanthropy. I was then thinking how and undergraduate research initiative and a prize would work well with most all of them wouldnt it?

There could also be an agi prize like the m and x prize.

Then there could also be an undergraduate research initiative for the exposure too right? And then more prizes could be offered for say like, commercials. We could put up something like a 500 dollar prize to whoever can make the best commercial, and that kind of thing.

Edited by brokenportal, 15 December 2008 - 08:17 AM.

[Heliotrope]

Exposure, publicity, no-secret little known org type, ads, philanthropy/donation too, yes

you can make initiative w/ millionaires donating


and news expose in good ways , ads etc

[brokenportal]

Im a little confused, could you rephrase that?

[Heliotrope]

Im a little confused, could you rephrase that?

Basically this is another nice combo that can be planned out.

I like the 4 things you said.

1. mfuri
2. nfuri
3. ad
4. money

ask for ad and donation too. new initiatives

Edited by HYP86, 15 December 2008 - 08:53 AM.

[modelcadet]

That's seems to be a pretty low bar.

Yeah, I know. But I was moreso implying that I can use the credibility I have as a member of this community, a sister to the SIAI. I'm sure he gets a lot of emails, but he's pretty conspicuous about his filtering mechanisms, which include but are not limited to the above example: It's intimidating to the more Platonic among us budding Jupiter Brains. I lurk on Ben's email lists, but I think I've only ever sent an email once or twice. I really enjoy using ImmInst as an interface with these researchers, though, I think because of the community. No matter how big SIAI gets, ImmInst will always be there, like the older brother who, after having turned to alcoholism, receives a check every month out of duty to family...

ImmInst should work with SingInst to create a profitable project in biofeedback sensors and services. We should get some people who can program for android and that POS the iPhone, some who can do some computational biology black majik, and some of the more classic medicine, mechanical engineering, biology, economics, marketing, etc etc. to start outputting a suite of really cheap biofeedback devices and sensors. I know we have the technology today to make bluetooth heart moniters that connect to your smartphone and send data real-time about your health and location (yay GPS!). Perhaps open instructionals for different things, with open code for the software side...

Anyway, we were talking about starting projects. I like that ImmInst has projects. SingInst needs projects. The thing... SingInst has no open boards. I feel like they're doing this intentionally, and it's part of the reason why they are so successful (besides the fascinatingly perplexing manner in which Ray Kurzweil reinvigorates the idea of economic growth). There are a lot of us eccentrics interested in these new transcendental projects, and I'm sure we can be quite offputting (I know most people look at me and think I'm crazy when I try to talk to them on the level; I feel sometimes like I'm trying to teach math to kid whose brain hasn't yet grown enough to learn the concept). If SingInst sanitizes itself of the general transhumanist movement to gain credibility, should we take that as a compliment if they do attain success in spreading memes and advancing temes?

All I know is this: I'm prolly gonna have about $10,000 in debt when I graduate, but as long as I can hack that, have a roof over my head, food in my belly, and access to nootropics, I'm willing to do just about anything, from telemarketing (shudder) to hoboing it up outside Google with a sign saying something new every day until they hire me. I'm sure there are more young people like me who understand the magnitude of the subject, and would be willing to do about anything for even a tiny bit of push in that next giant leap for mankind.

[I just decided I will creative commons or science commons my hobo signs]

Edited by modelcadet, 15 December 2008 - 02:57 PM.

[brokenportal]

Well, Im thinking that if an interested and qualified person wanted to do it, then they could talk to Ben and Bruce, and then get some guidance from Kelsey about how he has been doing it for MFURI, and then start setting it up. You seem like you might have the insight and tenacity to do it, if it turns out that you do then could you do it, and or do you know of other people you might recomend to do it and or help with the project?

[Mind]

Here is how Ben said we could combine narrow AI (like biomind) and life extension research:

We need some biology people t go through and standardize some large data sets and format it so that biomind (a narrow AI developed by Novamente) can analyze the data and derive new relationships between genetics and metabolism. It's boring work, but could be done with a couple seasoned biology/chem people leading a group of undergrads - undergrads do most of the boring stuff of course, combing through reems of data, leaders do QC. Anyway, that is Ben's vision of how to accelerate anti-aging research - identify complex genetic/metabolic relationships that are hidden in the data and then use that to focus on new highly targeted research.

[brokenportal]

Here is how Ben said we could combine narrow AI (like biomind) and life extension research:

We need some biology people t go through and standardize some large data sets and format it so that biomind (a narrow AI developed by Novamente) can analyze the data and derive new relationships between genetics and metabolism. It's boring work, but could be done with a couple seasoned biology/chem people leading a group of undergrads - undergrads do most of the boring stuff of course, combing through reems of data, leaders do QC. Anyway, that is Ben's vision of how to accelerate anti-aging research - identify complex genetic/metabolic relationships that are hidden in the data and then use that to focus on new highly targeted research.

Would that be in any way like how folding at home figures out proteins on our computers? Helping solve sens more directly through distributed computing would, well, the very thought of it has got butterflys in my stomach. Or dragonflys I guess rather. Apparently dragonflys are some kind of symbol for immortality.

So then would an undergrad initiative be a good thing to set up for this? Ive invited Bruce and Ben to this thread. Hopefully they let us know.

Does anybody know if Novamente has any money and if or if not would it be worth funding undergrad research like that through imminst? Seems like that would be a huge green light to me, especially since Ben in those quotes there seems to be looking for a good way to get it going.

[JohnDoe1234]

I like the ideas presented here!

Does anybody know if Novamente has any money and if or if not would it be worth funding undergrad research like that through imminst?

That would be an interesting development. Ben will speak for himself, but I think that would be a good partnership! (Assuming the Novamente project isn't at a position where extra funds would only give diminishing returns)

My two cents:

One thing that I've figured out is that I can disguise my AGI research/development as more conventional projects. For instance, I just got permission at my university to build a Semi-Stochastic Time-Series Pattern Detector (for stock market prediction) and receive credit as if it were a regular class (essentially just an independent study). But here's the kicker.... it's a required subsystem in my AGI design, so I'm killing two projects with one keyboard.

I would suggest anyone with AGI ambitions to try to do the same... I know it saves me time, so it might be something to try (Although it's not exactly a PR campaign as you all are discussing).

EDIT: Grammar

Edited by Joseph, 16 December 2008 - 07:33 AM.

[modelcadet]

Joseph, I think what you're doing is brilliant. Perhaps we, as a community, could brainstorm a set of more mainstream sugary coatings for our necessary set of projects. I for one am using historical economic studies to do work on the Friendliness problem. I've always wanted to create a wiki of dissertation ideas... not that I have the time to complete too many myself...

[brokenportal]

(Although it's not exactly a PR campaign as you all are discussing).

Well, it would be a PR campaign and a way to bolster the science. I mean if something like that got even just 5 more students in the whole world to work on AGI projects more and one of them came out with some amazing breakthroughs, then that could mean the difference between a grave yard and having the chance to see indefinity and all of existence pioneered.

[JohnDoe1234]

Perhaps we, as a community, could brainstorm a set of more mainstream sugary coatings for our necessary set of projects

I think this would definitely be a good use of our time.

I for one am using historical economic studies to do work on the Friendliness problem.

I'm interested in learning a some about this, any interesting findings so far?

Well, it would be a PR campaign and a way to bolster the science.

I didn't look at it that way, but you're right! Back in the early 60's AI research shattered into a thousand pieces like a broken window, and we've been slowly sweeping them into a single pile for the past 50 years. A general theory of intelligence and thus a general theory of artificial intelligence is getting closer, and is seeming more like a realistic goal than ever. Even if we did nothing to push the concept of AGI into the mainstream, it would still find its way in within the next decade.

The most important thing for AGI research at this point is to have a place for open dialogue and discussion and a collection of very intelligent people, which luckily we have.

[dhart]

Several undergraduate students are working informally now on OpenCog (http://www.opencog.org), which is an open source project based on the Novamente Cognition Engine.

Earlier this year Google funded undergraduate OpenCog research (http://google-opensource.blogspot.com/2008...g-and-gsoc.html).

On the subject of sensors & services, check out http://www.opensourcesensing.org/, a project sponsored by the Foresight Institute.

[brokenportal]

Several undergraduate students are working informally now on OpenCog (http://www.opencog.org), which is an open source project based on the Novamente Cognition Engine.

Earlier this year Google funded undergraduate OpenCog research (http://google-openso...g-and-gsoc.html).

On the subject of sensors & services, check out http://www.opensourcesensing.org/, a project sponsored by the Foresight Institute.

Interesting, now if we could form an ongoing permanent front for this, like an AGI URI or NMURI or whatever it would be called, then we could try to get places like google to give in the name of it to increase awareness for the ongoing project too. May as well get it all under one permanant roof right? I could be wrong Im not sure, just saying, you tell me? Then imminst could consider say, a $2,000 donatation to it also, and we could start petitioning transhumanism sites etc.. Ill help keep the effort going if somebody wants to help. The idea is on line 46 of LEEEP. If you have ideas and interest in this then ad to that. Only ad the good ideas, if you need help brainstorming and discussing them then go to the imminst broadcast channels chat room, theres life extensionists in there almost 24/7.

[brokenportal]

Alright, so we've got a few things moving on this project on line 46 of LEEEP there so far. The next step Im looking to write in is listing at least a few direct projects that undergrads can read through and decide to work on or not. So far we've got mentioned here in this topic, standardizing large data sets for biomind, and work with opencog. If somebody can find out a few exact more specific projects that people can do involving those two things or additional things then please list them here. I'll be contacting people like Ben and Eleizer as I get time too but it would be better if somebody who is more familiar with the lingo did it first. Ill then write the info into the projects development and move on to another step.

[MichaelAnissimov]

My recommendation on this is that someone should talk directly to Ben (if this is about Novamente) or directly to Michael Vassar (is this is about SIAI). It sounds like a good idea, as long as someone can be found to be paid to manage it, and the organizations in question have the time/money to pay that person. Otherwise, maybe ImmInst could pay them. I'd volunteer to lead an undergraduate program on this, as long as it was a paid position.

I disagree that "The most important thing for AGI research at this point is to have a place for open dialogue and discussion and a collection of very intelligent people, which luckily we have." The most important thing is to get more funding for the people that the Singularity Institute already has available.

[brokenportal]

My recommendation on this is that someone should talk directly to Ben (if this is about Novamente) or directly to Michael Vassar (is this is about SIAI). It sounds like a good idea, as long as someone can be found to be paid to manage it, and the organizations in question have the time/money to pay that person. Otherwise, maybe ImmInst could pay them. I'd volunteer to lead an undergraduate program on this, as long as it was a paid position.

I disagree that "The most important thing for AGI research at this point is to have a place for open dialogue and discussion and a collection of very intelligent people, which luckily we have." The most important thing is to get more funding for the people that the Singularity Institute already has available.

We've begun outlining this a little bit on line 46 of this google doc, and the next step we are at now is contacting AGI research people. ModelCadet is talking to some people at his school and some people around the internet, and Ive written to the main web address at Bens BioMind site, now we are waiting on that. Ill keep Vassar in mind and probably email him too. This will probably be for all of the AGI projects that are specificly for research for indefinite life extension, rather than AGI in general, or toward the singularity or something like that. I would rather have somebody that is already working more directly with this do these info gathering missions but not many people have had enough time to step forward so far. We are also looking for more ImmInst AGI people to get in on this in general.

Im hoping we can get a few volunteers to head this up and become the face, the front for it, but pay is always a consideration too. I figure the pay is in the indefinite life extension but if we have to incentivize people with money I understand that too and think it would be well worth it. Anissimov, or anybody, what is a money figure and an amount of hours you might consider to help along the way and become the main guy for this in the end when its all set up?

We need an AGI branch, but it has occured to me that we might want to set this up so its a broader undergrad research fund. Such as, include nanotech projects, and other biological aging research like for example, stem cells research and other strategies for combating the SENS 7 besides the ones given in SENS. We would then collect around 5 projects and project mentors from each section for students to choose from, to keep it simple in the beginning so we can get it moving. If we end up going that way then we will keep developing the AGI discussion here and then start topics like this in the other sub forums for the other potential parts of the undergrad research fund.

Edited by brokenportal, 06 February 2009 - 02:31 AM.

[MichaelAnissimov]

The thing about only supporting AGI projects working towards AGI for extreme life extension and not "AGI in general" or the "singularity" is that there is no real distinction between these three things. AGI = Singularity = Extreme Life Extension. Part of the point of a friendly singularity would be radical life extension, and any progress in AGI in general could be applied to life extension.

The other thing about AGI is that I think the field may be so advanced, sophisticated, and challenging that undergraduates would have little to contribute to it. Remember that the world's geniuses have been attempting AI for decades and consistently failing.

AGI is also a delicate topic because pursuing that research could lead to bad outcomes for humanity as well as good outcomes. Say that we decided to support an AGI project that makes a big deal about life extension in their talks and writing online, but they don't actually go through the hard work of writing a goal system that includes a complete morality which is preserved over iterations of self-improvement. Then we'd be contributing to dangerous scenarios.

Something that might be more practical would be to encourage undergrads to write about how AGI could contribute to life extension in the long term.

I think it would be premature to lay out time commitments/etc until we have something much more solid... I can only say that my decision on that would have to be based on much more information about what the initiative would be about.

[brokenportal]

The thing about only supporting AGI projects working towards AGI for extreme life extension and not "AGI in general" or the "singularity" is that there is no real distinction between these three things. AGI = Singularity = Extreme Life Extension. Part of the point of a friendly singularity would be radical life extension, and any progress in AGI in general could be applied to life extension.

Is there soemthing that can be done thats even more directly for life extension now, such as a project for interfacing the human genome project with biomind or something like that? Maybe collating some data sets that have to do with bioengineering of cellular aging problems, maybe interfacing Stanfords protein information with biomind and the genome and some bio engineering info or something like that? I mean, the world works toward these ethically friendly robots, but we may never get there. In the mean time we can still benefit life extension science poignantly toward goals for bio or nano or computer interface engineering of negligible senescence and anti aging projects like that right?

The other thing about AGI is that I think the field may be so advanced, sophisticated, and challenging that undergraduates would have little to contribute to it. Remember that the world's geniuses have been attempting AI for decades and consistently failing.

Yes, but we wouldnt get just any old undergrad, we would get the ones with good grades that are enthusiastic whos thoughts are already running wild on this stuff, and they would have to be interviewed first. Something Ive always thought about the worlds geniuses is that they concern themselves with more details on a subject than most people, but that once the details are collected, it makes it easier for more people to be a "genius" on the subject. For example Einstein was a revolutionary thinker and a genius in his day, but now most students who have a good grasp on physics and math can read all about it and know as much or more then Einstein did. Some people are super geniuses, thats true and thats great, but we can use all the regular geniuses we can get too.

AGI is also a delicate topic because pursuing that research could lead to bad outcomes for humanity as well as good outcomes. Say that we decided to support an AGI project that makes a big deal about life extension in their talks and writing online, but they don't actually go through the hard work of writing a goal system that includes a complete morality which is preserved over iterations of self-improvement. Then we'd be contributing to dangerous scenarios.

Well, thats true. The students though would be working with mentors who would want to include those kinds of things in the projects.

Something that might be more practical would be to encourage undergrads to write about how AGI could contribute to life extension in the long term.

Good idea, can you link an example of a paper or two like this?

In addition to this Im hoping that we can come up with at least a good 5 projects to start off with that students can do to help collate and interface the kinds of data that indefinite healthy life extension already works most directly with. I understand that other projects for AGI would make sorting through and understanding the data altogether easier, but Im thinking we want to focus in on a few more direct things and see if we can make some headlines and spread the bio engineering meme, try to make work with the bio engineering data easier to work with, and get some more undergrads a way to get directly involved, and stuff like that. Im hoping we can do a small AGI undergrad research fund like this through imminst and then if you guys want to continue it then build a bigger more broad AGI undergrad fund through NovaMente, or SingInst, or a new web page, or someplace also. These are just suggestions though. I may be way off. I look forward to seeing how this progresses as more and more people like you discuss this.

Edited by brokenportal, 06 February 2009 - 08:35 PM.

[JohnDoe1234]

]My recommendation on this is that someone should talk directly to Ben (if this is about Novamente) or directly to Michael Vassar (is this is about SIAI). It sounds like a good idea, as long as someone can be found to be paid to manage it, and the organizations in question have the time/money to pay that person. Otherwise, maybe ImmInst could pay them.

He's right guys, talking directly to them would be your best bet.

I disagree that "The most important thing for AGI research at this point is to have a place for open dialogue and discussion and a collection of very intelligent people, which luckily we have." The most important thing is to get more funding for the people that the Singularity Institute already has available.

I will agree with you in a few years. However, judging from the discussions on the agi/sl4 lists and what the various folks at SIAI are saying, the projects still need some more brainstorming and project polishing before the funding will make much of a difference.

It will still make a difference, but if you're concerned about the bang for your buck, I'd suggest waiting just a couple more years (maybe just 2 or 3). The smaller undergrad-oriented projects will have a greater impact in my opinion than funding the SIAI or the various projects on the aforementioned lists.

So for the time being, discussion is the most valuable thing we can have.

Unless of course the money given to one of the larger initiatives was used to create instructional courses much like Ben did with his OpenCog discussions, now that is a productive use of funds!

Hey, Eric do you have anyone with a project in mind we could maybe base the initiative around? (As a prototype?)

The other thing about AGI is that I think the field may be so advanced, sophisticated, and challenging that undergraduates would have little to contribute to it. Remember that the world's geniuses have been attempting AI for decades and consistently failing.

Agreed, but we must also consider the giant shed full of tools that we now posses, that they did not.

AGI is also a delicate topic because pursuing that research could lead to bad outcomes for humanity as well as good outcomes. Say that we decided to support an AGI project that makes a big deal about life extension in their talks and writing online, but they don't actually go through the hard work of writing a goal system that includes a complete morality which is preserved over iterations of self-improvement. Then we'd be contributing to dangerous scenarios.

Well, thats true. The students though would be working with mentors who would want to include those kinds of things in the projects.

Unfortunately even the brightest of human mentors currently don't have a grasp on this. And probably won't for a long time.

[brokenportal]

He's right guys, talking directly to them would be your best bet.

Yes, we are on that. Modelcadet and Mind and some others are in the process of asking some key people about this, and I have emailed a few people including the people at biomind. Some of us are trying to take this one step at a time through project #46 here. We have steps 1 through 3 done and are now working on step 4.

It will still make a difference, but if you're concerned about the bang for your buck, I'd suggest waiting just a couple more years (maybe just 2 or 3). The smaller undergrad-oriented projects will have a greater impact in my opinion than funding the SIAI or the various projects on the aforementioned lists.

If there are no projects that undergrads can work on that will make any worth while progress then we wont be able to do this but Im sure there are plenty of things they can be doing to help advance AGI and benefit life extension in the process that will be worth while. We just need to hear back from some people on these projects now.

Im hoping we can get your project in on this too. Could you direct me to an outline of what its all about? If it can directly benefit life extension along the way then we'll run it past some people and if you want to get some undergrads in on helping with it then we can look into that, now, or like you say, a year or two down the road.

Hey, Eric do you have anyone with a project in mind we could maybe base the initiative around? (As a prototype?)

Im thinking we should start off with a few main projects like biomind. We were chatting about it in the imminst broadcast channels chat room again today and we are thinking that these few projects may be built as a branch of a larger agi undergrad fund. They are thinking of calling the larger initiative "TEME" and so this branch with biomind and other projects might be called something like "The Indefinite Healthy Life Extension TEME". Please check in to the chat room whenever you have time. We all discuss this in there often. Im hoping to see discussion on it in there every day this week. I think Mind is going to schedule a meeting for it on thursdays also.

[modelcadet]

I'll post again when we have a time confirmed, but we are indeed planning on meeting Thursday afternoon (although we'll be likely meeting before then, and if you'd like to join our meetings, please PM me or Eric so we can add your name to our growing list of contacts). Mind mentioned asking Ben Goertzel to come and discuss projects involving Biomind as soon as this thursday, so please everyone keep brainstorming ideas for projects undergraduates can complete with such technology.

Here at UVa, we're working on the organization's constitution and the constitution for a first university chapter. We're reviewing this constitution with a contact at the Darden School of Business here at UVa, and will bring the constitutions to the meeting for review on Thursday. Mind stated he'd work on contacting Ben and on getting us webspace here, although he must go to the new Board of Directors.

I'm very excited about this idea, and I hope others are inspired to contribute to actualize this project. Again, please contact us for more information or offers to help! We need a critical mass of minds to actualize this project.

[Mind]

Ok, I got a hold of Ben and he said he had a couple of research ideas in mind through which we could use OpenBiomind. He said at least one of the project members would need to know command-line Java to get it done.

This is just one possible avenue to start out with, but it does not need to be the only possibility. It might help that we have a friend in Ben. If we can propose a focused project to start out with (perhaps with biomind), then we are more likely to garner support and be successful.