Author Archive

AI research is progressing in line with MST model of cognitive evolution

Having read Jame5 you will appreciate my excitement over the news that AI startup Vicarious is focusing on training its AGI to be more imaginative:

“Life would be pretty dull without imagination. In fact, maybe the biggest problem for computers is that they don’t have any. That’s the belief motivating the founders of Vicarious, an enigmatic AI company backed by some of the most famous and successful names in Silicon Valley. Vicarious is developing a new way of processing data, inspired by the way information seems to flow through the brain. The company’s leaders say this gives computers something akin to imagination, which they hope will help make the machines a lot smarter.”

Applying the MST model of cognitive evolution we can outline AI development as following:

Control of Position = Movement -> Industrial Robots (1975)
Control of Movement = Simple Reflex -> Roomba (2002)
Control of Simple Reflex = Complex Reflex -> Curiosity Rover (2010)
Control of Complex Reflex = Associative Learning -> Watson (2011) & Deep Mind (2016)
Control of Associative Learning = Imagination -> Vicarious (~201?)
Control of Imagination = Thought -> ???? (~20??)

How much longer before truly creative AGI emerges from this trend?

Comments

This Guy Is 3-D Printing a Classic Aston Martin … That Runs

Not just quite there yet in regards to owning licensed copies of cars but getting closer:

Using Autodesk 3ds Max modeling software and a Solidoodle desktop 3-D printer, Sentch began the painstaking process of printing out individual 4-by-4-inch sections, mounting them on the wooden frame and then gluing each piece into place. So far, he’s produced over 2,500 fiberglass molds and says he’s 72 percent of the way to completion, but that’s only the printing aspect.

Comments

Piratebay offers “Physibles” now

From Jame5:

The rise of self-replicating universal nano assemblers had not eradicated scarcity as expected, at least not in the sense many people had believed it would. It merely shifted the manufacturing cost for goods to the consumer and left it to the assemblers to supply the energy and raw materials to transform design specifications from virtual to physical reality. The cost went down, but large parts of the savings were added right back by the brand owners that henceforth concentrated on research and development, design, market research and advertising. One did not buy a Ferrari, one owned a licensed copy, and those still did not come cheap.

Reality is catching up with another step in the direction when you wont simply buy a Ferrari but download a licensed copy to be printed with your MakerBot 4 ready for the daily commute. Or even better – a pirated copy of physical objects as now on offer in a new category on the PirateBay:

We believe that the next step in copying will be made from digital form into physical form. It will be physical objects. Or as we decided to call them: Physibles. Data objects that are able (and feasible) to become physical. We believe that things like three dimensional printers, scanners and such are just the first step. We believe that in the nearby future you will print your spare sparts for your vehicles. You will download your sneakers within 20 years.

I am looking forward to that day and have been since 2005.

Comments

Biological Simultions Advancing Rapidly

Remember the molecular dynamics simulation of a virus I mentioned in the opening pages of the book? That were 1 Million atoms for 50 ns back in 2005. Now that we are well into the 2nd decade of the 3rd millennium researchers have stepped up their game and are tackling entire bacteria:

Using data supplied by researchers at the Max Planck Institute, University of Illinois postdoctoral researcher Elijah Roberts and chemistry professor Zaida Luthey-Schulten built a computer model of a bacterial cell that accurately simulates the behavior of actual cells.

Amazing! Now – whether you believe Wikipedia or or the Institute for Biomolecular Design, a E-Coli bacterium weights between 10^-12 and 10^-16 grams placing the number of atoms somewhere between half a Billion and 7 Trillion a pop. Taking the conservative approach by assuming half a Billion atoms we arrive at a factor 500 increase over 6 years.

Taking this armchair approach one step further there are about 100 Trillion cells in the human body that each have roughly 10’000 times more atoms than an E-Coli. So, how many years until we can reliably simulate a human being on the cellular level?

10^14*10^4 = 10^18 = ~500^7 ==> 7×6 = 43 years.

My prediction: if you are born post new year’s day 1975 – you will likely be around for the Singularity. Provided of course we will not descent into a post peak oil, fascist-theocratic dystopia before that. Take care of your lawn – you have been warned.

Comments

Jame5 2nd edition – call for testimonials

It has been a good two years now since I first published Jame5 in form of a dead tree and it was a great ride. Not only have I had the pleasure of founding the Beijing Futurist Society and Acceleration Aware Singapore but I also decided to quit my day job in order to focus on my research and entrepreneurial pursuits full time. And boy, oh boy did this rabbit hole turned out to go deep!

By now I managed to give away almost all of the 500 initial copies that I printed and am ready for a 2nd edition that will hopefully iron out the majority of the remaining spelling mistakes. So I thought it would add a nice touch if I got a few more testimonials and put them into the 2nd edition. So if you want to see your praise or constructive criticizm in print on Amazon do make sure to send me a 3-5 sentence blurp including name, URL, organization, title or whatever else you wish to add by the end of October 2009. I would really appreciate it!

For those of you willing to go beyond the call of duty, feel free to read over the final draft of the 2nd edition of Jame5 and let me know any errors my editor and me may have missed for the second time.

Last but not least: 2 years ago I thought self publishing Jame5 would be a good idea since what else could a big publisher do that I could not? Well – turns out that I do not have a marketing machine and while having ~1000 readers is nice, having 20’000 and someone else pay for distribution would be much nicer. So, since I still do not have a professional publisher, anyone with a good idea or connections on how to get Jame5 published properly, please do get in touch.

Comments

« Previous entries Next Page » Next Page »