Not that I'm especially excited by the internal machinations of SIAI, but I recently received an email from David Hart, who as far as I'm aware either is, or was, the project manager for the OpenCog AGI system. It sounds as if he got fired, demoted or otherwise excommunicated from the inner sanctum of SIAI after seeking funding for OpenCog from various big spenders - presumably going above his superiors (whoever they may be).
[email comments removed by request]
For a number of years I laboured under what transpired to be a false belief that SIAI was about promoting AI research and trying to foster the sort of "blue sky" AGI projects which probably wouldn't be possible in academia due to their scope and likely duration (hence the "for AI" part of the acronym).
I think I completely misjudged them, and my current understanding is that it's more of a theological institution with a particular agenda which doesn't have much to do with AI or making any serious attempt to advance the sum of knowledge in this area. Instead their focus seems to be upon an inflexible kind of AGI design of a very theoretical kind which, at least in my opinion, is not a helpful research direction. It looks like they've chosen to focus their attention upon issues which are sufficiently open ended that they can be indefinitely pontificated over, whilst ignoring more pertinent AI issues which are likely to genuinely touch people's lives in the coming decades. As I remember Steve Grand saying in one of the Biota podcasts, the whole singularity movement and its titanic marketing machine is pretty much a public relations disaster for anyone trying to do real AI research, simply adding to the already substantially confused way in which the field is popularly portrayed.
I think I completely misjudged them, and my current understanding is that it's more of a theological institution with a particular agenda which doesn't have much to do with AI or making any serious attempt to advance the sum of knowledge in this area. Instead their focus seems to be upon an inflexible kind of AGI design of a very theoretical kind which, at least in my opinion, is not a helpful research direction. It looks like they've chosen to focus their attention upon issues which are sufficiently open ended that they can be indefinitely pontificated over, whilst ignoring more pertinent AI issues which are likely to genuinely touch people's lives in the coming decades. As I remember Steve Grand saying in one of the Biota podcasts, the whole singularity movement and its titanic marketing machine is pretty much a public relations disaster for anyone trying to do real AI research, simply adding to the already substantially confused way in which the field is popularly portrayed.
My advice to David would be not to waste any more intellectual effort on SIAI and instead just concentrate his skills on trying to push the envelope and keep OpenCog going, if that's his primary concern. I'm certainly in favour of applying open source methodology to AI and robotics problems, and the future holds potential for some exciting developments. On the other hand, theological debates and egotistical power politics are not helpful, and it's always worth remembering that those who claim to be the high priests of rationalist utilitarianism may actually be false prophets trying to lure you away from the critical path.
A good direction for OpenCog to go in would be to try to link up with what Willow Garage have been doing, especially in terms of reasoning about maps and objects in the robot's field of view, or maybe reasoning about possible intentions of people in the environment. If they could acquire one of the PR2 robots and put it online in a test area for project developers to use that might result in some progress being made.
17 comments:
I'm sorry that this little kerfluffle involving David has made its way onto a public blog!
In the big picture this is not a big deal and there are a lot of more interesting things to worry about...
My own biased view is that OpenCog is a great AGI project with a real chance at creating a superhuman ethical thinking machine ... and SIAI is a excellent futurist thinktank, which has put on great conferences and come up with some really cool ideas and may play a major role in shaping the dialogue about AGI and Singularity as things move forward...
Human personalities don't always get along. Let's move on to something else ;p
Oh, just in case some readers don't know: I'm the leader of the OpenCog project and also the Director of Applied Research of SIAI...
Agreed, and the title of the post does imply that this is not an earth shattering event. It may only be of interest to a small clique of AIphiles. I'm sure that people at Google must get a lot of emails either about GSoC or just general requests for funding. Sending an email to Peter Norvig isn't usually a sacking offense though.
David's publishing the email messages without obvious sign of consent seems a bit rude.
Summarizing makes it look as though you are giving your version of the story - but that's better than publishing email messages without the sender's consent.
The messages have multiple email addresses in the clear, nontheless :-(
Oh, hang on - it looks as though that's YOU doing that, Bob! :-(
Well that's why I've linked to the full email in case you think I'm taking it out of context. If someone sends me an email and I think it's publishable (i.e. not too impolite) and might be of interest to a wider audience, I'll publish it.
The way I read it the scandal began as the deletion of an email account and ended with sacking. As far as I can tell from the available information unless David said anything really inappropriate to Google's Big Cheeses it seems unreasonable to fire someone for sending an email. As a project manager you would expect him to be trying to promote his stuff, but I think the background is that he may have already had some differences with SIAI after they dropped funding for OpenCog on the belief that a "rogue AI" was being created.
Some people can be sensitive about that - e.g.:
"Other breaches of "netiquette" to avoid: Posting private email to a public forum or sharing private email with others. For example, if Lucy Leader sends a message only to Lisa Leader, Lisa should not share Lucy's message with others unless she receives Lisa's permission to do so."
- http://www.llli.org/llleaderweb/LV/LVAugSep97p84.html
Well this may be a culture clash. In the world of open source, of which OpenCog is a part, people don't generally ask for permission before sending emails to anyone. You get to see the entire process, warts and all, including rants and ideological/technical squabbles between developers.
Some may post emails from others publicly without asking permission. However many regard this practice as undesirable and a breach of netiquette. For instance:
"And please remember. Conversations in private emails are to remain private. It is very bad netiquette to post a private message to a public list without the writer’s permission. Although Internet law is still being written, this practice is commonly considered copyright infringement and at the very least makes you look extremely unprofessional."
- http://www.profnet.org/blog/category/giving-leads/
I am not aware of open source communities ignoring such guidelines any more than other communities do. I would generally expect such folk to be more aware of netiquette issues than most communities are.
Re: "[email comments removed by request]"
That's a big improvement. Aside from netiquette issues, publishing email addresses in the clear is also not recommended. If you look at http://singinst.org/aboutus/team - many of the addressees are making some minimal effort to avoid some of the more stupid spamspiders from obtaining their email addresses. I am pretty sure you don't want to contribute to the spam load of fellow researchers.
Re: singularity as public relations disaster
I do think the whole concept is a confused muddle - more likely to cause confusion than enlightenment. It is frustrating that there are still people promoting this terminology. It is not remotely scientific.
I am reminded of those who promoted the concept of a singularity in physics. They delighted and promoted the division by zero in their models. It let them weave fanciful stories about wormholes, about time and space exchanging places and "naked singularities" - where the laws of the universe broke down.
Real scientists just shake their heads at such nonsense. Dividing by zero just shows that your model is busted - and that you should return to the drawing board and try again.
Re: "A good direction for OpenCog to go in would be to try to link up with what Willow Garage have been doing"
It sounds like a bad plan to me. Robots have even worse disadvantages than interacting with humans in 3D virtual worlds does - most cripplingly, very slow build-test cycles, and the poor penetration of networked robots / virtual worlds.
It is practically bound to be better to put your sensors and actuators out on the internet - in much the way search oracles and stockmarket bots do.
It would be possible to have an internet based intelligence, but chances are that the type of intelligence which would emerge would be so different from ourselves that we might have difficulty in even recognizing it - rather like trying to detect the intelligence of cetaceans.
There are opportunities to study human communities, social interactions and economics within virtual worlds, in spite of their current limitations, and to devise agents which can learn to exist more or less independently in that sort of environment.
About the focus of OpenCog: we're actually taking both approaches mentioned, and believe they can be made to work together. At Xiamen University we're using OpenCog to control humanoid robots (Nao robots, but we'd like to use others as well); and we're also working on NLP with a view toward eventually having the system read the Web. Each approach has unique aspects to offer, and it seems the same AGI mind can encompass both...
Different intelligence - sure. That is a feature. Machines have progressed so far mostly by complementing human abilities - by being strong where we are weak. Being strong where we are strong puts you are in direct competition with human beings - which are fairly cheap anyway. It just doesn't make economic sense. One day, machines will be able to compete directly with us on our own turf - but we will get other pretty powerful machine intelligences before that.
Re: cetacean-like communication problems - we can talk to Google after a fashion in english today. Also, the creators will be able to probe and understand its brain. Then there are the economic effects. All rather different from cetaceans, I would say.
3D virtual worlds are the place to learn some things - but they are not where to learn how basic intelligence. For that you large volumes of sense data and actuator possibilities, with humans in the loop as little as possible and an economically significant application. Those requirements have internal conflicts - but IMO, simulated 3D is nowhere near the sweet spot.
Post a Comment