Pages

August 14, 2010

Liveblogging the Singularity Summit - Michael Vassar the Darwinian Method

Ad Support : Nano Technology   Netbook    Technology News    Computer Software

Capitalism and Science is what enabled us to go beyond merely rational methods to go beyond sailing ships and pyramids and enabled to take off and go beyond a backwater to dominance from the 17th through the 19th century.

People have always thought and achieved useful understanding of the world around them. Despite this, there is an important sense in which science only began a few hundred years ago. We resolved this apparent paradox by recognizing that humans have developed many different tools to allow reason to guide their collective actions. One of those tools, which we will call ‘Enlightenment Science’, grew out of philosophical skepticism to transform the world quite recently. Many scientific controversies can best be understood as conflicts between those who accept an earlier scientific method, that of ‘Scholarly Science’, and those who believe ‘Enlightenment Science’ to be the only source of authoritative knowledge. In truth, the scholarly and enlightenment scientific methods contain both serious drawbacks and great utility. From Darwin's theory of evolution to the theory of global warming, important science often depends on a proper synthesis of the two methods. The strongest arguments for anticipating a technological Singularity fall squarely into this synthetic paradigm. I will explain why these two scientific methods work, where they come into conflict, and how such conflicts can be resolved.


Archimedes and the ancient rational innovations - math, engineering, rules of logical argument and dialog and weak on naive empiricism.

But not science
No hypothesis testing
No organized literature and publication standards
what are the limitations of this method
* works but requires honest geniuses
* no checks against fooling yourself
* games gives practice
* Other ancient rationality
* nationalistic explanation
* philosophy
* craftsmanship and evolution of arts and technic
* exploratory data collection
* markets and hierarchies (achieve rational aggregate behavior, better than anarchy)
* finally scholarship

Scholarly Scientific Method
* identify people who learn well
* teach them to notice ignorance and want to cure it
* have them read broadly guided by curiosity and seeking surprises
Have them seek and write about diverse life experiences
* look for convergence in what they say
Need to still have competition of ideas and have openness to new ideas

Why does scholarship work
Human children build models naturally
- redundant data sets make pattern detection easier
- averaging together errors with peers eliminate them
Failing error rates reallocate effort from model building to goal oriented action Varied experience maintains high error rates, maintain exploratory behavior with adult minds.
Society develop specialists.
Why scholarship fails
Correlated errors there are ways to mitigate but you can still fool yourself and ever

Radical skepticism

Enlightenment science

do not need to be a genius to do it

seemingly necessary ontological assumptions can be false
- in practice was not a problem until 20th century
* form can be imitated to aid justications
- frequently involves ethical or epistemic assumptions rather than ontological assumptions
- believe humans have free will or your choice is determined anyway
- often smuggles in unchallenged assumptions
- ultimately can end in blatant conflict



How does the Darwinian method achieve long term sustainable incremental improvement and advancement
How Darwin creates the theory of natural selection
Darwin and Wallace had evolution as a scholary hypothesis
But then Darwin developed a follow on scientific hypothesis that was testable and provided more things and ways to follow up and test more things and learn more things
Examples of do not push this red button. May not be able to do an experimental test. Need uncertain logical argument
example of modern environmentalism. Not conclusive with enlightenment science

The Singularity
Independent origination of similar hypothesis even similar terminology
-Von neumann, Vinge
* Logical argument
- Vinge, I J Good
* Inductive data on exponential progress
Kurzweil and other with examples of exponentials in technology
Therefore rational and scientific conclusion

Summary Links to the Entire Summit

The best of the 2010Singularity Summit

Shane legg has developed a formula and program that generates a continuous intelligence measurement for rating algorithms and systems This has the potential to more objectively guide the improvement of AI and algorithms.

Goertzel is using AI techniques to figure out useful things about the Genescient long lived flies.

Lance Becker on using cold and chemical cocktails to enable you to survive having no oxygen to your brain for hours instead of four minutes

The talk was not good but the work has big potential. Anita Goel is harnessing DNA polymerase in devices for rapid sequencing and as actuators.

The interesting inventions of Steve mann.

Details on brain computer interfaces

Combining neuroscience and machine learning as a hybrid approach to AGI

Average Talks

Vasser talk on the going from the scholarly methods to better science methods

David Hanson Builds robots that mimic humans and have character

On using haptics for faster reflexes and remote action

Ray Kurzweil talk on building brains

Reverse engineering brains within reach

Details on regenerating mice. Does not appear to be close to being applied to humans, although other DARPA regeneration work might be

Eliezer - simplified ethics

Ramez Naam the possibility of using advanced biotech to fix global warming and food and other issues

Jose Cordeiro has an overview of the future of energy



Lower end talks

James Randi - no such thing as scientific consensus - getting fooled and fooling yourself. Entertaining but not especially relevant or important.

Gregory Stock now believes the singularity could happen but human values will not survive. I disagree with his argument. It is too general to make a convincing case

A panel on narrow and general AI. Not very illuminating

Tooby talks about evolutionary biology and tries to connect it to AI. Boring and impenetratable.

Irene Pepperberg has smart parrots and chimps and parrots are about equal to 6 year old humans in many cognitive capabilities



what cells can do that robots cannot

If you liked this article, please give it a quick review on Reddit, or StumbleUpon. Thanks

Supporting Advertising

Business Success
   How to Make Money    
Executive Jobs    
Paid Surveys


Thank You
blog comments powered by Disqus