For me Artificial Intelligence is both a fascination and a huge disappointment. Frustration with the lack of progress in this field is probably born out of the lofty expectations set back the 60s but even accounting for that decade’s heady optimism, we’re making slow progress.
Bina48 is a strange duck in the AI pond. Not the progeny of MIT or some Government funding Japanese lab, she springs from private funding at the Terasem Movement Foundation in Vermont. The group dedicated to the development and promotion of AI is developing Bina48 as a type of analog of one of its founders.
The Engadet article linked here has an interesting video shot by a reporter. As such the depth of exploration into the technologies in use is lightweight but it does give a glimpse of the amazing progress they are making.
Bina48 is one humanoid robot you’re going to want to hang out with — Engadget.
Just when you thought A.I. was dead, or at least relegated to calculating your credit score, along comes the yeast genome sequencing robot!! Now there’s a long way between designing an experiment to sequence yeast genes and being able to do the housework for you but this is a significant step as the experiment was designed unaided by humans.
Gizmodo – Robot Makes Autonomous Scientific Discovery for First Time – Robotic scientist.
Is it me or does this fembot have really big hands for a, ummmm, fembot..? Apart from the misplaced digit lengthage, it’s pretty amazing tech. Roll over Asimo. You’re competitors are catching up! Japan has lead the world in humanoid robot bodies for decades. Looks like they’re getting close to something that would actually be accepted into society. Now all we need is for Roney Brooks, Marvin Minsky and their buddies at MIT to get the AI sorted out and we’re home and dry. Shame they seem to be nowhere as close to completing their side of the project as the Japanese are.
Read an interesting essay by Minsky over the weekend. It was written a couple of years ago and he says that he reckons by the time computers are capable of 1000 MIPS we’ll have the computing power to support human levels of AI (assuming they can get the software to work!). Today’s fastest desktops deliver about 25 MIPS. With CPU power roughly doubling ever year it will take us 5-6 years to reach that level.
Skynet is upon us.
Robots: The HRP-4C Model Robot Is Programmed to Work…It.