Interestingly enough, VentureBeat was inspired by the launch of the OpenCogPrime wikibook to do an article on Novamente and OpenCog generally:

http://venturebeat.com/2008/08/05/can-novamentes-virtual-dogs-make-a-nintendog-look-dumb/

I talked to the writer of the article a while, and while he had to keep the article fairly short and to the point, he was actually quite savvy on issues related to Narrow AI versus General AI, the possibility of a technological Singularity, and so forth. Which is great: the more futurist and AGI memes spread through the general techo-savvy population, the easier the job of marketing advanced technologies and products based thereon becomes.

It’s also encouraging that the launch of an open-source project would inspire a VC-focused media outlet to become enthused about a company. I think this is correct, of course — in the current era, the open-source methodology can be a powerful way to move a company ahead, which is one of the reasons Novamente decided to work with SIAI and others to launch the OpenCog project. But not all correct things are embraced by the media and the VC community … I think it’s great that the OSS approach has reached such a level of legitimacy in the business context. (Although, it still must be pursued with subtlety — we have been quite careful about choosing which aspects of our technology to keep proprietary, based on our smart-NPC’s business focus.)

Onward!

Ben

It’s been a while since I posted anything here, and I won’t try to do justice to everything that has happened for Novamente LLC in the last months. Instead, this blog post just highlights a few of the many exciting pieces of news in the Novamente universe.

First of all, we’re going to be present at the Virtual Worlds 2008 conference in Los Angeles in 2008, showing off our virtual pets in the Multiverse virtual world.

Next, among a variety of other AI consulting projects for various commercial customers and government agencies, we’ve recently undertaken two interesting customer-focused projects specifically oriented toward AI in virtual worlds:

• a collaboration with RealXTend, aimed at putting our smart virtual pets in their virtual world
• a project for a government agency, creating a simulation game in multiverse in which human players and AI’s cooperate, with a view toward studying variations in the patterns of game play among humans of different cultural backgrounds

Finally, Novamente LLC has launched an open-source AGI project, the OpenCog project at opencog.org. Novamente has contributed a significant amount of AI software code to the OpenCog project, and has also contributed an extensive AGI design called OpenCogPrime. The aim of this project launch is to accelerate progress toward powerful artificial general intelligence by bringing additional collaborators into the process of advancing the AGI code and ideas developed at Novamente LLC during 2001-2008; and also to allow Novamente LLC to focus more closely on its business of creating the world’s most intelligent virtually agents for online virtual worlds and games.

In this era the world has seen a long list of examples of commercial firms profiting considerably from their work with open-source code-bases, and we fully expect Novamente LLC to form yet another example on this list. However, I hasten to add that Novamente’s virtual-agent product code remains closed-source and proprietary, along with certain business-critical AI components. Furthermore, as Novamente moves to launch increasingly intelligent virtual agents in online worlds and games, the knowledge inherent in these agents’ minds — learned via interactions with human teachers and other agents — will itself form an increasingly valuable asset.

We are excited about the power the free and open source development methodology has for accelerating AGI progress, and look forward to collaborating with academia, industry and independent researchers to further advance this very important technology, for mutual benefit.

(On the Merits of Parrots … or: “The Wisdom of Crowds” as a Strategy for Educating Young AI’s)

In this blog post I’ll enlarge upon a point I made during my recent talk at TransVision 2007 (see a recent blog post by Bruce discussing this talk), regarding the potential of virtual worlds to help in accelerating the path of AI’s toward mastery of human language.

(This is just an example of a more general point: The more I think about the direction Novamente has chosen to take over the next few years — seeking to roll out intelligent virtual agents widely throughout 3D and 2D virtual worlds and MMOG’s — the more I become convinced that it’s a very positive direction from a pure-AGI perspective as well as from a business perspective.)

As a specific example, one vision that’s been haunting me lately is a virtual talking parrot. A simple idea, of course — but very powerful in its AI implications. Imagine millions of talking parrots spread across different online virtual worlds — all communicating in simple English. Each parrot has its own local memories, its own individual knowledge and habits and likes and dislikes — but there’s also a common knowledge-base underlying all the parrots, which includes a common knowledge of English.

[Bruce Klein w/ parrot at Novamente’s Second Life HQ]

Now, suppose that an adaptive language learning algorithm is set up (based on, oh, say, the Novamente Cognition Engine), so that the parrot-collective may continually improve its language understanding based on interactions with users. If things go well, then the parrots will get smarter and smarter at using language, as time goes on. And, of course, with better language capability, will come greater user appeal.

The idea of having an AI’s brain filled up with linguistic knowledge via continual interaction with a vast number of humans, is very much in the spirit of the modern Web. Wikipedia is an obvious example of how the “wisdom of crowds” — when properly channeled — can result in impressive collective intelligence. Google is ultimately an even better example, I think — the PageRank algorithm at the core of Google’s technical success in search, is based on combining information from the Web links created by multi-millions of Website creators. And the intelligent targeted advertising engine that makes Google its billions of dollars is based on mining data created by the pointing and clicking behavior of the one billion Web users on the planet today. Like Wikipedia and Google, the mind of a talking-parrot tribe instructed by masses of virtual-world residents will embody knowledge implicit in the combination of many, many peoples’ interactions with the parrots.

Another thing that’s fascinating about virtual-world embodiment for language learning is the powerful possibilities it provides for disambiguation of linguistic constructs, and contextual learning of language rules.

Michael Tomassello, in his excellent book Constructing a Language, has given a very clear summary of the value of social interaction and embodiment for language learning in human children.

For a virtual parrot, the test of whether it has used English correctly, in a given instance, will come down to whether its human friends have rewarded it, and whether it has gotten what it wanted. If a parrot asks for food incoherently, it’s less likely to get food — and since the virtual parrots will be programmed to want food, they will have motivation to learn to speak correctly. If a parrot interpret a human-controlled avatar’s request “Fetch my hat please” incorrectly, then it won’t get positive feedback from the avatar — and it will be programmed to want positive feedback.

Yes, humans interacting with parrots in virtual worlds can be expected to try to teach the parrots ridiculous things, obscene things, and so forth. But still, when it comes down to it, even pranksters and jokesters will have more fun with a parrot that can communicate better, and will prefer a parrot whose statements are comprehensible.

What it comes down to is: A virtual parrot, learning language, will have lots of teachers, and that’s a good thing. The more customers we get for the parrot, the more teachers the AI underlying the parrot will have.

A baby AI has a lot of disadvantages compared to a baby human being: it lacks the intricate set of inductive biases built into the human brain, and it also lacks a set of teachers with a similar form and psyche to it … and for that matter, it lacks a really rich body and world.

However, the presence of thousands to millions of teachers constitutes a large advantage for the AI over human babies. And a flexible AGI framework, like the Novamente Cognition Engine, will be able to effectively exploit this advantage.

On a more theoretical level, this community of individually-acting yet collaboratively-learning parrots may be considered an example of a mindplex, a term I introduced in this essay a couple years back, referring to a collection of minds in which

• each individual mind has its own declarative and procedural memories, and sense of self
• there is also a collective declarative and procedural memory, and a collective sense of self

Mindplexes tie in interestingly with the notion of the emerging global brain, which I discussed extensively in my 2002 book Creating Internet Intelligence.

Getting back to practicalities: The rate of progress of Novamente LLC in our new business direction is difficult to estimate, as it depends on funding and other related issues. But, if all goes as we’re hoping, we may well be able to release a parrot-that-talks-and-adaptively- learns-to-talk-better sometime before the end of 2008. And that will be pretty exciting!

And of course parrots are not the end of the story. Once the collective wisdom of throngs of human teachers has induced powerful language understanding in the collective bird-brain, this language understanding (and the commonsense understanding coming along with it) will be useful for other purposes as well. Humanoid avatars — both human-baby avatars that may serve as more rewarding virtual companions than parrots or other virtual animals; and language-savvy human-adult avatars serving various useful and entertaining functions in online virtual worlds and games. Once AI’s have learned enough that they can flexibly and adaptively explore online virtual worlds (and the Internet generally) and gather information according to their own goals using their linguistic facilities, it’s hard to see limits to their growth and understanding. (And this leads to various deep and critical ethical concerns, such as I’m exploring with my colleagues at the Singularity Institute for AI.)

But, we need to get there one step at a time. What’s exciting about virtual parrots-that-talk — and the intelligent virtual agents space generally — is the way it poses an incremental path by which getting more and more customers for products is directly connected to making the AI underlying the products smarter and smarter (which in turn will attract more and more customers). This is exactly the kind of virtuous cycle one wants to see in an AI start-up company (in my never-very-humble and admittedly rather biased opinion!).

Ben Goertzel
CEO and Chief Scientist
Novamente LLC

[Photo credit: George P. Dvorsky’s Blog - Sentient Developments - Article: TransVision 2007: the good, the bad and the ugly]

Fun happenings this week!

During Transvision 2007, Ben Goertzel spoke (transcript-audio) on a panel w/ Marvin Minsky and Second Life’s Philip Rosedale. The following quote from Ronald Bailey’s Reason Online article — Would you give up your immortality to ensure the success of a posthuman world? Answering hard questions at the World Transhumanist conference — sums things up:

Finally, Rosedale mentioned the possibility of creating AI avatars that could learn from interacting with the avatars of humans in Second Life. “I find it very likely that any artificial intelligence we create will live first in a world like this,” said Rosedale.

Rosedale’s last observation flowed nicely into the next talk by Novamente AI researcher Ben Goertzel. Goertzel wants to create baby AI’s that can learn and insert them into virtual worlds where human avatars can teach them. He suggested creating them as virtual pets, perhaps a parrot or a cat, that would be embodied, reflective, and could use adaptive learning. People in virtual worlds like Second Life could teach AI avatars not only tricks, but also about space, objects and even to talk. Whenever any one of the AI avatars learned something new it could be transferred immediately to all of the other AI avatars. With millions of virtual world residents teaching AI avatars, they could rapidly acquire artificial general intelligence.

Towards a Joyous Singularity,

Bruce Klein
President, Novamente

Ben and I had the pleasure of visiting the sprightly Googleplex on May 30th, 2007 where Ben delivered a 60 minute Google TechTalk (Artificial General Intelligence: Now Is the Time) and was generously introduced by Google’s Director of Research, Peter Norvig (yes, that Peter Norvig). Ben recounted the occasion over at his blog, giving some fairly strong reasoning on why Google is probably not hot on the trail of creating AGI anytime soon. I made a couple of comments at my blog to remember this fun event as well.

Onward,

Bruce Klein
President, Novamente

Hi all,

This brief blog post is just to announce that Novamente LLC and Electric Sheep Company have started a collaboration on some projects applying AI to virtual worlds.

We are very excited about this collaboration: Electric Sheep is a great company; and their expertise in virtual worlds complements our expertise in AI. We anticipate building some great products together.

Watch this space for more information: Details will be released here when the time is right

Ben Goertzel
CEO and Chief Scientist
Novamente LLC

Hi all –

Ben Goertzel here, with some minor news: I have recently written two articles for kurzweilai.net, and the first one has just appeared!

It’s called “ARTIFICIAL GENERAL INTELLIGENCE: NOW IS THE TIME”, and can be viewed at

http://www.kurzweilai.net/meme/frame.html?main=/articles/art0701.html

It touches a bit on Novamente, and long-term applications thereof (artificial scientists, digital twins and such) but is mainly pretty general and high-level.

The second article focuses more specifically on the Novamente approach to AGI and should appear on kurzweilai.net fairly soon.

The theme of the first article may be inferred from the title: I make a case, familiar in essence to those who know me, that AGI at the human level or beyond could be achieved in a relatively brief period of time (say, 5-7 years … possibly less) with a serious, intensive, concerted effort by the right people.

Needless to add, this is one of the guiding aspirations of Novamente LLC. We aim to create progressively more and more powerful AGI that controls embodied agents in simulation worlds. In this way we will create AGIs that are not only extremely intelligent but have grown up in (online) human societies and learned how to think, act and be themselves via interacting with huumans.

We at Novamente are going as fast as we can toward powerful AGI. One point I make in my kurzweilai.net article is that progress could be even faster with a Manhattan Project type effort. However, assuming our progress continues to accelerate as it has been, we will get there even without that kind of large-scale support. We have a solid AGI design, a team competent to implement it, and a nifty plan for monetizing it. So: onward!

And, of course the Manhattan Project style approach would have its risks too. As AGI systems grow ever smarter, AGI ethics will become a major issue; and Novamente LLC’s concern for this issue is indicated by the fact that I and Novamente’s President Bruce Klein have become involved in SIAI, an organization devoted to AGI ethics (as Bruce discussed in his recent blog post here).

Ben Goertzel
CEO and Chief Scientist
Novamente LLC

Greetings, Novamente blog readers!

I have good news regarding a new collaboration between Ben Goertzel, myself and the Singularity Institute for Artificial Intelligence (SIAI). Over the past few years, SIAI has done a remarkable job of attracting top talent to focus on its mission. Be sure to read for yourself, SIAI’s mission statement:

In the coming decades, humanity will likely create a powerful AI. SIAI exists to confront this urgent challenge, both the opportunity and the risk.

SIAI is a nonprofit research institute based in Palo Alto, California, with three major goals: furthering the nascent science of safe, beneficial advanced artificial intelligence (self-improving systems) through research and development, research fellowships, research grants, and science education; furthering the understanding of its implications to society through the AI Impact Initiative and annual Singularity Summit; and furthering education among students to foster scientific research.

As such, Ben and I are happy to join forces with SIAI to help advance the Institute’s Outreach and R&D Program. We’ll certainly remain focused on Novamente. Indeed, our primary passion is to grow Novamente into a world leader in AGI. However, we feel rewarded by devoting a portion of our time to SIAI, which in turn can help us develop ethical AGI.

By the way, be sure to mark your calendar, Sept 8-9, for SIAI’s Singularity Summit in San Francisco, where Ben will be speaking, and I’ll be organizing. There’s already an excellent lineup of speakers!!

Regards,

Bruce Klein
President, Novamente

Hi all,

Welcome to the new Novamente blog!

Although it’s a “corporate blog” the idea is definitely not to have an impersonal corporate communication style — blog entries will be written by particular people within the company presenting their own perspectives on events and ideas of relevance to Novamente LLC. (For example, this entry is written by me, Ben Goertzel, Novamente LLC’s founder, CEO and Chief Scientist.)

Other than announcing this blog, the main purpose of this blog entry is to announce, as the title says, “A new direction for Novamente.”

No, we haven’t given up AGI and decided to reorient ourselves toward making the world’s shiniest paperclips (and, in spite of my daughter’s urging, I haven’t given up AI for bunny breeding either). However, we have made a substantial decision regarding our business direction, one which also has some impact on the technical direction of our AGI R&D.

Since its founding in 2001, Novamente LLC has operated as a combination AI software consulting firm, and AI R&D shop. Our R&D efforts have focused on the creation of the Novamente Cognition Engine AGI system, as well as on the creation of innovative software for natural language processing, data mining and (in collaboration with Biomind LLC) bioinformatics. Our consulting customers have included INSCOM, NIH, Zero Degrees, and other firms in the medical and financial industries. We haven’t grown to a huge size nor have we made a huge amount of profit, but through our consulting work we’ve supported our staff and managed to fund a modest amount of R&D.

But we need to move faster.

Business-wise, we have found that the kind of highly complex AI consulting work we’ve been doing is not a very scalable business model. We have been making a modest profit, but not a huge one.

And in terms of our AGI R&D plan, we have not been progressing toward the completion of the implementation of our Novamente Cognition Engine design, nearly as quickly as we’d like. The current, partial implementation is a good one, and can be used to do some interesting things — but still, it includes only around half of the NCE design. We would like the implementation of the next half to be a LOT faster than the implementation of the last half has been.

Some of our consulting projects have involved some of the Novamente Cognition Engine code, being used in specialized and application-specific ways. So there has been some synergy between the R&D and consulting sides of the company. But this synergy has not worked out as dramatically as was hoped. By and large, we have found, given any specific consulting project there are usually easier ways to meet customer needs on schedule and at low cost, than to use one’s in-development AGI system.

So — to cut finally to the chase — we have decided to re-orient the firm as a software product company, aimed at the Intelligent Virtual Agents space. Some information on our planned product offerings can be found on the Products page on this website:

http://www.novamente.net/product

In short, our intention is to focus our efforts on creating products that utilize the Novamente Cognition Engine to provide intelligent control for virtual agents embodied in simulation worlds, including MMOGs, consumer virtual worlds like Second Life, and training simulations as utilized in government and industry.

We hope to launch our first two products, “Novamente Intelligent Virtual Agents Server” (NIVA) and Novamente Virtual Pet, in 2008. Descriptions of these and other intended product offerings may be found on the Products page mentioned above.

For the time being, we will continue to do AI consulting projects in order to keep revenue coming in. But our goal is, over time, to transform the firm into an intelligent virtual agents software product firm — and to build out our Novamente Cognition Engine AGI design in the context of creating more and more intelligent embodied virtual agents residing in simulation worlds.

Long term, this doesn’t mean we won’t also one day pursue other sorts of opportunities. The ultimate capability of the Novamente Cognition Engine to out-Google Google is never far from my mind (it occurs to me every time I use Google and get frustrated with the whole archaic key word paradigm). I also remain extremely excited about the capability of AGI technology to accelerate progress toward human life extension via intelligent analysis of biological data and text; and our ongoing work in this area is bound to continue, via use of Novamente LLC technology by Biomind LLC staff. Among many other wonderful vistas.

But there’s an old saying, “Your strengths are your weaknesses.” Myself personally as a scientist, and Novamente LLC as a firm, have displayed the strength of diversity and “polymathy” — but this strength has also been a weakness in some contexts, because while there is a depth of understanding that comes from diverse experience, there is also an added efficiency that comes from narrower focus.

I have become convinced that what Novamente LLC needs to do right now is to set aside the protean, omni-potentially useful nature of its technology and concepts , and focus on building products in one particular area — so as to bring in the dramatic profits that will accelerate our progress toward more and more intelligent software (which in turn will bring in more and more profits, etc.).

We have also brought on some excellent advisors to help us with this transition, in the context of forming a formal Strategic Advisory Board for the firm, see

http://www.novamente.net/advisory

The choice of intelligent virtual agents as a focus area was not random, obviously. We went through a few bad choices before this one; and this is the first one that’s being announced publicly, because it’s the first one that really feels “right.”

One of the reasons it feels “right” is that it provides maximum harmony with the R&D goals that have been in the minds of myself and the Novamente LLC founders since the beginning. Controlling intelligent embodied agents in virtual worlds may or may not be the precisely best path to creating powerful AGI, but it’s certainly a reasonable and interesting path in that direction. As we progress down this path, advances in the Novamente Cognition Engine will lead fairly directly to improvements in the products we offer our customers.

Another reason it feels “right” is that it provides a mix of long-term grandiose business possibilities, with short-term concrete market opportunities.

The grandiose part: I do believe that as the Web unfolds over the next 5-10 years, it is going to gradually morph into a sort of “metaverse.” Second Life, World of Warcraft, multiverse.net and all the other virtual worlds coming out lately, are just the beginning. If we can become the dominant provider of intelligent agents for the metaverse as it emerges, that will be a pretty exciting business position.

Ultimately we would like to offer “digital twins” to users of Second Life and other virtual worlds, using Novamente cognition technology to enable your twin to act in your stead in various virtual-world situations. As a starting point, we intend to offer virtual pets for use by virtual world users. (There is a lot of funky learning code already working in the Novamente Cognition Engine, so it won’t be astoundingly hard to, say, make an online Aibo that can out-Aibo Aibo…)

The short-term and concrete part: There is also a healthy market right now in terms of training simulations for use within government and industry. Non-player characters in training sims and “serious games” are just as inflexible and unsatisfying as in entertainment games. Our business model going forward will be based on addressing this market, the MMOG market, and the nascent consumer virtual worlds market, in parallel.

Profit from selling NCE-powered virtual agents will fund R&D on the NCE, allowing us to scale up our R&D side beyond what it is now — and the smarter engines we will build will make our virtual agents yet more powerful and useful.

All in all, this business direction is what John Laird referred to long ago as the “killer app” for AI – a pun on the prevalence of scary murderous AI-controlled monsters in video games. Except, of course, monsters in video games are just a tiny part of it.

Anyway, we will keep you posted via this blog as things develop!

Yours,

Ben Goertzel
CEO and Chief Scientist
Novamente LLC