Singularity Hub

Kevin Kelly, cofounder of Wired, gave a thought provoking, concise, and easy to grasp 20 minute talk at the 2007 EG Conference on the future of the internet.  Below I have posted the video of Kevin’s talk, followed by my own summary and thoughts of what Kevin said:

The internet is only about 5,000 days old, and in this time it has delivered to us an amazing list of capabilities and changes that nobody could have predicted beforehand (satellite photos of the entire world, instant access to almost any piece of information such as phone numbers, photos, sports scores, government forms, real estate listings, etc.)  Kevin points out that most people have accepted and adapted to these changes with very little sense of awe or wonder, as if it is no big deal.  It is amazing how easily mankind accepts the impossible once it becomes reality.

Kevin predicts that in the next 5000 days the internet is destined to evolve into one giant super computer that can store, share, and manipulate all of the world’s information.  All of the devices that we use (computers, handhelds, phones) are simply points from which we are able to access and interact with this global computer or mind…

Read the rest of this entry »

Singularity Hub is proud to deliver the web’s most comprehensive coverage and analysis of the Singularity Summit 2008. The Singularity Summit is the premier annual event for those that are interested in the singularity. Below you will find our high level summary, followed by a link to a much more detailed description with pictures.

On Saturday October 25, 2008 I attended the Singularity Summit at the Montgomery Theatre in San Jose, CA. An impressive lineup of speakers, including Ray Kurzweil (de facto singularity advocate), Peter Diamandis (Founder/Chairman of Xprize Foundation), Vernor Vinge (famous science fiction author), and Justin Rattner (CTO of Intel) were on showcase for the roughly 500 attendees. The summit was thought provoking, inspiring, and overall a success.

The summit began promptly at 9:00am and continued throughout the day until 6:00pm with a few breaks in between and a one and a half hour lunch break. Here are the Hub’s major takeaways from the event:

1. When people become believers in a near term singularity (a singularity that may come in their lifetimes) they radically change their behavior in terms of risk tolerance, eating habits, and investment horizon. If large numbers of people begin to believe in a near term singularity this poses the possibility of enormous and potentially dangerous upheavals for society.

2. Even if a true singularity is not reached within our lifetimes the singularity summit reinforces the vision that tremendous technological change beyond our imagining is coming in the next 40 years. In the next 5 years an explosion in interest about the singularity and the pace of accelerating technology may occur.

3. According to Ray Kurzweil, solar energy is an information technology that is experiencing exponential growth. Solar energy production has doubled every year for the last 20 years and is now only 8 doublings away (that is about 10 years!) from providing nearly all of the world’s energy needs. The implications of this trend are huge and warrant careful consideration for the environment, investment, politics, etc.

4. Peter Diamandis announced that the Singularity University (SU) will be launched in the near future. The Hub’s Keith Kleiner will be a founding member of SU and we will have much more to say about SU soon!

5. According to Intel CTO Justin Rattner Intel has a solid roadmap that will ensure that Moore’s law will continue for at least another 10 years, by which time computers will be at least 1,000 times more powerful than today’s computers

6. Virtual worlds will continue to gain traction and functionality as people continue to recognize and leverage the unique advantages that these worlds offer versus the physical world.

7. Computers may be able to beat humans at chess and air hockey, but they are still a long way off from emulating human emotion and social behavior. Demonstrations today of the cutting edge in computer emulation of emotion and social ability were downright pitiful. Of course it is possible that we will make big leaps in the coming years, but today’s demonstrations were not encouraging.

Below is a breakout of the entire Singularity Summit:

Read the rest of this entry »

The Short:

Researches across the globe are making daily advances towards the development of human level artificial intelligence, but sadly the algorithms and the software that represent these advances often remain hidden within researcher’s computer labs, out of reach for others to review and build upon.

Enter OpenCog, an entire website dedicated to the development and distribution of artificial intelligence tools, software, and resources that are open source and freely available for anyone to use and modify.  Opencog’s ultimate goal is to develop true human level artificial intelligence, but even if Opencog fails to achieve this most ambitious of goals the project still seems poised to bring many other significant benefits to the AI community.

Opencog was only recently launched in 2008 and at present appears to have a small following.  The project offers the promise of an open and novel approach to developing and sharing new advances in AI, and it will be exciting to see it evolve and gain momentum in the coming years.

The Long:

Opencog is led by prominent AI scientist Ben Goertzel who will be presenting the opencog project at the Singularity Summit 2008.  Opencog is a multifaceted project with multiple goals and initiatives.  One of the most prominent components of the project is called the opencog framework.  From the opencog website:

The OpenCog Framework, which is part of the The OpenCog Project (a platform to build and share artificial intelligence programs), includes components for procedural and declarative knowledge representation (AtomTable), task scheduling (CogServer), AI algorithm containers (MindAgents), XML data import and export, connectors to instant messaging and virtual world systems, and other facilities. MindAgents and other add-ons under active development explore a wide variety of AI techniques including evolutionary program learning, probabilistic inference and reasoning, natural language processing, and others.

The OpenCog Framework will be released under an OSI-approved free software license in Summer 2008. To get involved, please post to the OpenCog general mailing list at http://groups.google.com/group/opencog/.

Interested C++ developers or other technically inclined individuals, please drop in at #opencog on IRC.freenode.net.

Here is an image showing where the the opencog framework sits on the computing stack:

Opencog is also developing a collection of specific AI tools (e.g. learning and reasoning components) to be used in creating AI systems.  From the site:

The key tools initially provided as part of OpenCog are

1. The RelEx natural language comprehension system
2. The MOSES probabilistic evolutionary learning system
3. The PLN (Probabilistic Logic Networks) framework
4. A system for Economic Attention Allocation and associative memory

Yesterday Intel CTO Justin Rattner delivered his keynote presentation about the future of technology and the coming of the singularity at Intel’s annual Intel Developer Forum (IDF). Although Rattner could use some tips on giving a more entertaining presentation, some of the ideas and technologies presented during the keynote were truly fantastic and we give the keynote a mild thumbs up.  Below is the Hub’s take on this exciting event:

The first thing that must be acknowledged about Rattner’s keynote is that it represents a symbolic endorsement of the singularity from a major player in the technology industry. Believers in the singularity are sometimes seen as crazy or over optimistic, but with Rattner’s keynote it is clear that the idea of a singularity occurring in this century is gaining serious momentum.

Rattner’s keynote began with a roughly 5 minute introductory video from singularity proponent Ray Kurzweil. During the rest of the keynote Rattner brought out a series of individuals to demonstrate futuristic technologies that are on the horizon. Some of these demonstrations were really cool…others not so much. Here is the breakdown:

1. Intel Scientist Dr. Mike Garner Presents the Future of Moore’s Law
Our Take: Nothing new here

In this segment Rattner discusses with Dr. Garner how the industry will be able to continue the doubling of computing capacity every year as it has done for the last 40 years. With current technology at the 32 nanometer scale, we are fast approaching limits on how small we can get with current semiconductor technology. Dr. Garner suggests a number of technologies and ideas for overcoming this hurdle (trigate transistors, quantum computing, carbon nanotubes, etc.), but in our opinion none of these ideas are new and no new light is shed on how and when any of these technologies will become a reality.

2. UCSB phD Brian Koch Presents Silicon Photonics
Our Take: pretty cool

Brian discusses a collaboration between Intel and UCSB to combine the speed and bandwidth offered by light (photonics) with the versatility and mass production capabilities of silicon. This marriage between silicon and photonics could offer a host of new capabilities and advantages in the future including huge bandwidth, decreased cost, and decreased energy consumption. Intel has a pretty good website dedicated to this entire technology here.

3. UCSB Professor Jan Rabaey Presents Massively Wireless Communication
Our Take: Interesting, but not revolutionary

Dr. Rabaey predicts that we will have 1000 radios per person in 10 years! This was the most interesting idea offered during Dr. Rabaey’s presentation and is definitely an exciting vision of what the future may hold. Dr. Rabaey stated that the three major problems with a massively wireless world are limited radio spectrum, limited energy for radio devices, and too many standards. Dr. Rabaey offers intresting ideas for overcoming these problems, but from our view this is simply a focus on improving the efficiency and the intelligence of how radios communicate. What we fail to see in this presentation is any revolutionary technology or capability for the future.

4. Alanson Sample from the University of Washington Presents Wireless Power Transmission
Our Take: Awesome!

Wireless power transmission technology debuted about a year or so ago, but nonetheless it is a stunning concept and the demo was great to see. Alanson was able to send 60W of energy wirelessly across a distance of two feet to power up a light bulb. Approximately 25% of the energy is lost during transmission, which is pretty good (Alanson claims this is more efficient than many wall warts)! From the standpoint of the singularity one of the more interesting applications of this technology would be the ability to recharge or power up devices that have been implanted into the human body. Jordan Robertson from the Associated Press in a recent article gives us more in depth information about the technology, and reveals that 90% efficiency of power transmission has actually been achieved at a distance of 3 feet.

5. Dave Ferguson and Siddhartha Srinivasa Present Autonomous Robots
Our Take: Not bad, but nothing revolutionary

David and Sidd demonstrate a robot named Herb that can enter an environment and autonomously find, navigate to, and then pickup any coffee cups in the vicinity. Many people underestimate the difficulty in such a task, but in truth it is actually quite difficult for robots to replicate the human ability to navigate a new and unknown environment that is also changing as people and other objects within it move. Herb did a commendable job of performing his task, but overall this demonstration is similar to many other projects occurring all over the world, most notably the DARPA Urban Challenge which is referenced during the presentation.

6. Joshua Smith from Intel Presents a Sixth Sense Called Pretouch
Our Take: Now This is Revolutionary!

Josh presents to us something completely novel, a sixth sense called pretouch in which a robot uses electric fields to sense its environment. The inspiration for this technology comes from certain species of fish that have this same capability. In the demonstration a robotic hand was able to use pretouch to create a 3 dimensional spatial model of an apple that it was presented with. The robotic hand was then able to grab the apple based upon this dynamically generated model. The entire task was performed without the aid of cameras (for vision) or any other conventional senses. MIT Technology Review wrote an article about pretouch nearly a year ago and it is a good place to start if you want to learn more. Pretouch gives us a glimpse of a future in which humans and robots will have capabilities that are completely different and novel from those that we have today.

7. Tan Le From Emotiv Presents Noninvasive Neural Interfaces
Our Take: Cool, but invasive neural interfaces are the real future

Emotiv demonstrates one of the most practical and well developed noninvasive neural interface applications we have seen. In the demonstration a human plays a game in which he is able to scare creatures away, change colors of the game, and even move objects (though with noticeable difficulty) simply by thinking alone. The gamer is wearing a helmet that detects electric field fluctuations emanating from his skull and these fluctuations are translated into actions within the game. Emotiv shows us a really cool possibility for enhancing gaming interfaces and a compelling pathway to many other novel applications. Despite this, we are only marginally excited about this technology because noninvasive neural interfacing is a poor substitute for the capabilities offered by truly invasive, direct neural interfacing. Noninvasive neural interfacing is error prone and limited in its ability to extract human intention. Lets go straight to the source and focus on direct neural interfaces into the brain…that is where the true link to human intention will be revealed.

8. Jason Campbell from Intel Presents Shape Shifting Materials
Our Take: This Rocks!

Jason Campbell presents an Intel initiative to build shape shifting materials composed of building blocks called catoms. Intel’s latest catom prototypes are currently 1/10 of millimeter in diameter nanobots with their own microprocessors and actuators that are capable of changing their color and other properties. Jason presents us with the vision of a matchbox sized cell phone composed of thousands or even millions of catoms that can morph on demand into a nearly full sized keyboard. Catom based shape shifting technology is still in the very early research phase and there does not seem to be any real world example of this technology to date. Nonetheless, it was fascinating to see real progress being made in a technology that has captured the imagination of science fiction writers for decades.

New Scientist released an awesome article describing how researchers are using rat brains to control a robot. Approximately 300,000 neurons from a rat fetus are deposited onto a sheet of nutrients and electrodes (and MEA, or multi-electrode array) and the neurons immediately begin to build connections to each other and to the electrodes.

The robot can send signals to the neurons via these electrodes and the neurons will over time fire off in predictable patterns. These patterns can be connected to output electrodes that can send signals back to the robot and cause it to react to its environment, such as avoiding a wall. The video is simply awesome so check it out:

Add Intel’s CTO and senior fellow Justin Rattner to the growing list of singularity proponents according to an article from computerworld. According to Rattner:

“…perhaps as early as 2012 we’ll see the lines between human and machine intelligence begin to blur. Nanoscale chips or machines will move through our bodies, fixing deteriorating organs or unclogging arteries.  Sensors will float around our internal systems monitoring our blood sugar levels and heart rates, and alerting doctors to potential health problems.”

Rattner will be presenting his talk entitled “Crossing the Chasm Between Humans and Machines” at the Intel Developer Forum Aug 21, 2008. Come back here to the hub in August to see our review!

Just found a nice video posted below of a group of Nao robots doing a synchronized dance. The Nao represents the next generation of “affordable, autonomous, and easily programmable humanoid robots”. Created by Aldebaran Robotics, the robot has succeeded the Sony Aibo as the official robot for this year’s RoboCup, in July 2008 in Suzhou in China. In late 2008 Aldebaran Robotics hopes to make the Nao commercially available.

Nao SmoothMove Dance

The Feelix Growing Project is a $2.5 million euro initiative funded by the European Union to foster research in the development of robots that can detect and mimic human emotion. The project started in December 2006 and has already been cited by numerous outlets such as Gizmodo.

Euronews, a major European news outlet, recently aired this interesting 8 minute video clip documenting the project:

euronews - futuris - Robots learn to express emotions

From the IEEE report on the singularity I found an interesting piece of information from the article titled “Can Can Machines Be Conscious?” by Christof Koch and Giulio Tononi. It would seem mankind’s attempt to reverse engineer the brain is quite a ways off if we can’t even reverse engineer the 302 neuron roundworm! From the article:

“What is the best way to build a conscious machine? Two complementary strategies come to mind: either copying the mammalian brain or evolving a machine. Research groups worldwide are already pursuing both strategies, though not necessarily with the explicit goal of creating machine consciousness.

Though both of us work with detailed biophysical computer simulations of the cortex, we are not optimistic that modeling the brain will provide the insights needed to construct a conscious machine in the next few decades. Consider this sobering lesson: the roundworm Caenorhabditis elegans is a tiny creature whose brain has 302 nerve cells. Back in 1986, scientists used electron microscopy to painstakingly map its roughly 6000 chemical synapses and its complete wiring diagram. Yet more than two decades later, there is still no working model of how this minimal nervous system functions.

Now scale that up to a human brain with its 100 billion or so neurons and a couple hundred trillion synapses. Tracing all those synapses one by one is close to impossible, and it is not even clear whether it would be particularly useful, because the brain is astoundingly plastic, and the connection strengths of synapses are in constant flux. Simulating such a gigantic neural network model in the hope of seeing consciousness emerge, with millions of parameters whose values are only vaguely known, will not happen in the foreseeable future.

A more plausible alternative is to start with a suitably abstracted mammal-like architecture and evolve it into a conscious entity.”

picture of c. elegans from newsdesk

IEEE produced a special report on the singularity in its June 2008 issue located here:

http://www.spectrum.ieee.org/singularity

This is a comprehensive report, representing many diverse views about the singularity from a selection of tech luminaries and scholars in singularity related fields. It is a must read if you are at all interested in the singularity. In a few separate posts I will be highlighting some of the cool insights to be found within this report.

The Short:
The EE Times reports that a computer has been built that can beat the world’s best humans at the game of air hockey.  Check out this video of the air hockey robot putting another wimpy human in his place:

The Long:
It is interesting to note how easy it was to build a robot that could outperform a human in air hockey because of the increasing capability of modern processors. The designers admit that they intentionally did not optimize the code when they could have in order to demonstrate the effect of going from an 8-bit to a 32-bit processor. The designers state:

“We could have optimized the 8-bit code, but we wanted to run the exact same program to show the difference in processors.”

And here are a few good quotes:

“So far, the robot has defeated every human opponent running in 32-bit mode, averaging three times as many goals as human players”

“A video system that tracks the puck’s position sends coordinates to the
board every 10 milliseconds. In 32-bit mode, the recalculation time
could keep up with the puck position. In 8-bit mode, the longer
recalculation time caused the robot to miss the puck.”

The day when computers equal and then surpass us is not as far away as many think…

Decent piece in BusinessWeek profiling Jeff Hawkins’ startup Numenta. Numenta is building artificial intelligence that attempts to replicate the function of the human neocortex to solve hard problems. The software specializes in recognizing patterns within massive streams of data.

I am a big fan of Jeff Hawkins…here is a short plug about him:
After inventing the Palm Pilot Jeff Hawkins left the tech industry to pursue his dream to unlock the secrets of the brain.  His book On Intelligence is a fabulous read for those looking for a straightforward introduction to the inner workings of the brain that doesn’t require a neuroscience degree to understand.