Singularity Summit Australia 2011 Wrap-up

Full shownotes and links: http://hive45.com/plus/singularity-summit-australia-2011-wrap-up

Our discussions:

James Newton-Thomas [07:38]
* "Advances in Science and Technology"
* Re-iterated recent trends in technology and robotics
* Quadrocopter - $300 RRP toy device with advanced microcomputer technology, could not be possible 10 years ago.

Hugo de Garis [8:37]
* "The politics of AI"
* Writing a new book with Ben Goertzel
* Constructing super human intelligence (Hugo calls them 'artilects')
* Global species war as artilects approach and surpass human intelligence
* Unabomber's Manifesto - Ted Kaczynski could have statues erected in the future as being the first Terran
* The uncanny valley of machine intelligence could spark the mainstream fear
* Laws of robotics, ethics in AI, AI nanny
* Anthropomorphizing AI
* Terrans have zero chance unless they strike first
* Is Hugo a Terran or a Cosmist?

Stelarc [18:08]
* "Human+"
* Biotechnology, transhumanism, surrogacy, art
* More distributed parts that make-up the conscious self
* "It's no longer meaningful to ask if it's biological or not, it's become a more effective operational system and that's what's important"
* First athlete with prosthetic legs qualifies for "real" Olympics

Steve Omohundro 1st talk [21:42]
* "Minds Making Minds: Artifical Intelligence and the Future of Humanity"
* If everything acts as an intelligent, rational agent they are sociopathic. Needs irrationality = lack of computational resources
* There's an equilibrium between rationality and available computing resources
* The ecology of AI
* Dangers of programming a utility function - e.g. AI with desire to play chess
* Chess grey goo!

Ray Jarvis [24:53]
* "The envy of Roboticists -- AI Operating in the Material World"
* Practical talk on what we have now in robotics
* Using spinning lasers to 3D map an environment for robots
* Google Earth3D map the entire world with drones - "burn the retinas, use the new Android Eye!"
* Role of robots for AI as a feedback mechanism for training

Ben Goertzel 1st talk [28:48]
* "Artificial General Intelligence -- Roadmaps"
* Artificial General Intelligence and the roadmap on how to achieve it
* OpenCog
* Operating in the Minecraft world to teach an AI agent
* Nanny AI as a possibility of halting progress in AI development until it matures
* Gradual transition and merge with machines
* The notion of AI being engineered or emerging?

Kevin B. Korb & Ann Nicholson [35:09]
* "Bayesian Artificial Intelligence"
* Bayesian probability and trying to workout utility function
* Utilitarian ethics in AI that considers all agents
* AI take 500 years to develop, got in an argument with Goertzel
* "Arrogance displayed as false modesty"
* Ethics doesn't apply to AI
* Bayesian nets tool

Steve Omohundro 2nd talk [39:59]
* "The Emerging Global Mind, Cooperation, and Compassion"
* Hivemind AI rant!
* So much more untouched on
* Ants and bees, individual cells joining together to create complexity
* Society operating quicker
* Maybe blackholes are the reproductive drive for the Universe

David Dowe [43:45]
* "(Bayesian/Algorithmic) Information theory, one- and two-part compression, and measures of intelligence"
* Universal general measure of intelligence
* Abstraction and abduction
* Measure of intelligence is what is the most complex situation an entity can understand
* Kolmogorov complexity
* Economist article: Intelligence testing - Who are you calling bird-brained?
* Definition of intelligence that is transcendent of humanity
* Steve Wozniak's coffee test

David Chalmers [47:53]
* "The Singularity -- A Philosophical Analysis"
* A philosophical analysis of the Singularity and AI - formulating the assumptions
* AI, AI+, AI++
* It doesn't require consciousness, just intelligent behaviour
* Mind uploading and philosophy of mind
* Merge with your upload
* Simulating virtual avatar self life paths and then picking the best one

Burkard Polster & Alan Hájek [52:01]
* "The problem with probability"
* "A Plea for the Improbable"
* Monty Hall Problem
* Probability and Bayesian networks
* Philosophy of probability

Ben Goertzel 2nd talk [53:25]
* "Genomics, AI and Experimental Evolution"
* Methuselah flies - experimental evolution for breeding flies increase lifespan by 4x
* Comparing SNPs using narrow AI to find differences
* Genescient Stem Cell 100
* Gene therapy for longevity
* The Methuselarity
* Experiment-driven discovery instead of traditional scientific hypothesis
* Genetic Census

Lawrence Krauss [60:31]
* "The future of life in the Universe"
* A Universe From Nothing video
* How long Moore's Law will last for in a finite Universe
* Universal expansion, flat Universe
* Skeptical about other civilizations in the Universe reaching the Singularity
* Sleeping and computing to last as long as possible before the heat-death of the Universe
* The future aint what it used to be

Category:

Science & Technology

Tags:

• singularity summit
• singularity summit 2011
• singularity
• hive45
• nathan waters
• tristan grace
• geoff pidcock
• james newton thomas
• hugo de garis
• stelarc
• ray jarvis
• AI
• steve omohundro
• ben goertzel
• kevin korb
• ann nicholson
• david dowe
• david chalmers
• burkard polster
• alan hajek
• lawrence krauss
• artificial intelligence

[tomjneal]

I don't understand the concept of how to even program or harness within AI the ability to have ambition. It's really interesting to think about, with all this talk about AI trying to wipe out the human race - what is the motive? What would be the point of programming AI with its own initiative, rather than just maintaining it as merely a tool for the human race? Interesting thoughts, guys. Keep it up.

[ThorBarstad]

Nice video.

[timberWolfful]

Nice work dudes.

[logmeinwtf]

@CyberCanadian it's taking some time again lol ^^

[CyberCanadian]

New vid soon ?

[Dbelenit]

best song i heard in a while

[CyberFad]

Amazing work guys! one hour+ .. YAY :D

[Anarkokommunist]

Hey, I'm studying to become an accountant! =O

[CyberCanadian]

Wow you guys do a lot of good work in your videos, you deserve more viewers

[TheRationalFuture]

Ray Jarvis is the head of Robotics dept at Monash Uni. I like the idea of having some sort of Swarm AI technology, perhaps using quadcopters instead of drones to send back all sorts of metadata about the environment, including temperature scans, video, photos, time of flight info (as Vinge suggests) etc.