5 ms·
The fact that he was able to do it at all means that this should be a logical next step. There may even be a business idea here. Program good enough Computer Vi
by jxcole 15y ago
The fact that he was able to do it at all means that this should be a logical next step. There may even be a business idea here. Program good enough Computer Vision and planning AI you could loan a set of autonomous robots to someone and pick them up after 6 months when the job is complete. It would probably be much cheaper than the human alternative.
- marshallp 15y agoYou don't even have to program. Set it up to be controlled by people in india or china. This kind of business could be set up in a few days, and which begs the question - how com no-one is doing this. There's companies trying to sell laptops on a stick (anybots) whereas this is a genuinely useful service. Trevor Blackwell should be able to get something like that up in day, he did a segway clone in a day.
- rdouble 15y agohow com no-one is doing this because it takes 7 years to dig out a single basement?
- marshallp 15y agoThat's only because he was doing it as a hobby and because it's one person. A team of workers with larger equipment working 24 hours a day in shifts over the internet could get it done much quicker and cheaper. In addition to digging basements, there's also grass cutting, delivery bots in restaurants, hospitals, tractor driving.
- andylei 15y agohaving people in india control it is probably isn't that much cheaper than hiring someone to actually drive it. it mostly likely does not cover your costs of developing systems to remotely control expensive and destructive construction equipment from thousands of miles away.
- jwn 15y agoThis is where someone on HN developes a remote-system-as-a-service.
- kragen 15y agoShit, I was going to post about someone on HN doing just that, but then I remembered that might still be under NDA.
- Danieru 15y agoIf I understand it right the first rule of NDA is you do not talk about NDA.
- kragen 15y agoIt's been a lot of years, and I think the information revealed by my posting on the topic is sufficiently de minimis that no harm is done. I did hesitate for a while over the "reply" button, though.
- rdl 15y agoThe exception is when you're operating in difficult environments, where the cost of putting people there is huge (life support costs, hazard pay, losses, etc.). Space is the obvious example (manned spaceflight might be a good thing politically, or for inspiring people, but no one contests that robots are far more efficient for almost any mission). A closer to home example is commercial diving. They used to pay people $100-500k/yr to do saturation diving, fairly frequently, at deep sites. Maybe 3-6 weeks of prep, a ship, a transfer chamber, on-deck chamber, a team of support staff, a team of divers, and lots of equipment. Instead, now, they send down a ROV (usually operated from the ship above, but conceivably these could be centrally teleoperated someday too, or autonomous). Commercial divers are still used for smaller tasks, especially where they're not doing saturation diving, but the saturation diving industry in general has been increasingly restricted to shallower jobs or jobs requiring a large amount of work at a single site (involving lots of manual fitting). For the very deepest dives, ROVs and atmospheric diving suits (basically single-person submarines) have basically taken over -- the deep dive record was set in 1992 and no one has tried to match it. There's also the EOD market (iRobot's packbot), hazmat environments, war zones, etc., where teleoperated equipment will get more and more popular.
- rdouble 15y agoThe length of time it took is probably more a factor of the tiny trucks and backhoes. One can dig out a similar sized hole in weeks using regular equipment. Real time tele-operations are very expensive and require specialized equipment. Military drone ops cost billions of dollars. Bomb squad robots and law enforcement drones cost half a mil on up. Even the toy dump truck FTA costs $5000. In addition to the equipment cost, there is no way the latency could be compensated for over the regular internet. None of this would make sense for a residential excavation It's actually really cheap to get a basement dug, in the midwest it's like $1500-$3000 to dig the hole, depending on the size. I've heard as little as $800 for a smaller basement.
- culturestate 15y ago> Real time tele-operations are very expensive and require specialized equipment. Because this is the current state of affairs doesn't mean it has to be this way in 2 years. That's the very definition of disruption.
- marshallp 15y agoReal time teleoperation costs billions? Everyone (almost), even in third countries, has skype running, which probably is more bandwith intensive than teleoperation would be. And with teleoperation, there are parts where automation can make it even less bandwith. Plus, one worker can probably control a few machines at the same time while some parts of operation are automated. irobot has this feature (you tell a robot to go somewhere, it goes there, meanwhile you can do something else).
- rdouble 15y agoYou're baked if you think you can control a bulldozer from India over Skype.
- icebraining 15y agoParent never said "over Skype". And NASA controlled the mars rovers from earth with "nearly five times the speed of home dial-up"; I'm pretty sure you can get more than that from India to the US.
- VMG 15y ago* At least anything to do with humans is probably subject to regulation. * Unions * The technology probably isn't quite there yet
- randomdata 15y agoWhen FarmVille came out, I started wondering how I could leverage the help to man my real farm. People were jumping at the chance to do the work for free, it seemed only logical to capitalize on it. I never did find a solution, but I still believe there is one looming out there.
- icebraining 15y agoI'm pretty sure a real farm doesn't have much to do with FarmVille, except for the names and some concepts. And besides, writing an algorithm to do what people do on FV is easy enough. The hard parts on a farm is data acquisition, crunching it (which requires experts to write software for) and the manual labor.