Personal Robots From Valley Startup

Tjeerd writes ""A Silicon Valley start-up is developing a hardware and software development platform for personal-assistant robots, autonomous boats and unmanned cars. The privately funded company, quietly started almost a year ago by eGroups founder and veteran Google architect Scott Hassan, plans to make its robotics software open source. That way, it hopes to draw a community of developers to build applications in these respective fields.""

Comments

Great start

[Muad'Dave]

Cool. Great start. Call me when you have 'bots that can:

• Load the dishwasher
• Empty the cat litter
• Feed the cats
• Let the dogs in/out
• Feed the dogs
• Do laundry
• Clean gutters
• etc

The vacuuming 'bots are cool, but there's so much more they'll need to do before they're really integrated, Jetson's-style.

Re:Great start

[BadAnalogyGuy]

* Load the dishwasher -> Wife
                * Empty the cat litter -> Wife
                * Feed the cats -> Wife
                * Let the dogs in/out -> Wife
                * Feed the dogs -> Wife
                * Do laundry -> Wife
                * Clean gutters -> Wife
                * etc

The Wife unit also does the hoovering.


COST?

Expert systems or learning?

[BadAnalogyGuy]

While they have no stated goal, one point of the article was the DARPA self-driving competition. This type of goal-oriented competition is really adept in getting people to think of specific problems and devising clever systems to solve these problems. However that's still nothing more than advanced expert systems, and a far cry from a robot that actually "think" for itself. Idiotic contests like the Turing test seem to push AI in the direction of elementary data processing but unfortunately never very far beyond.

I wonder what sort of competition would get people thinking about solving the "thinking" problem. Where robots make informed and appropriate decisions outside the parameters they were originally designed for. Not just to learn, but to take learnt knowledge and apply it in an "intelligent" manner.

Robomaid

[Doc Ruby]

The most promising robot development is Roomba, because it's task oriented on a task humans don't want to do. But Roombas are too flimsy and noisy, and expensive. If companies just worked on that, making Roombas robust and cheap, and expanded them to washing clothes, dishes, and the rest of the home that isn't on the floor, they'd have enough complex behaviors that they could start adding "personality".

Re:Robomaid

[tgd]

The biggest problem with the Roomba is that it just doesn't work. Its a dustbuster, not a vacuum. I know a lot of people who bought them and very few people who kept using them because you realize the first time you bust out a real vacuum how little it really cleans.

And I don't mean to dig at Roomba with this, but any robotics company will have a fundamentally similar problem -- lack of power. AI isn't the only real problem with household robots -- the mechanical efficiency of them and the capability they have to store power are the real limiting factors. It doesn't matter to me if the robot can find my litter box or if it can empty the dishwasher if it doesn't have enough power to do that.

Re:Business Strategy - unproven but hopeful

[eln]

It depends on how long their private investors are willing to continue pouring money into the operation. If all of the investors are passionate about the work and have lots and lots of cash, it might work out. If the company were to go public with that kind of strategy, of course, it would sink like a stone.

Re:There's a reason there are no self-driving cars

[ultrasound]

It's not exactly a mathematical proof, and the logic is flawed. Using the same argument one could 'prove' the non-existence of practically anything that is computer controlled e.g. Automatic Landing Systems, ABS, Lifts, fly-by-wire, food production lines, medical machinery etc. etc.

The corollary is that a litigious society prevents any advancement in technology that may have implications on human life. And if that situation ever comes about it is time to shoot all the lawyers.