If you’re of a certain age the first personal robot in fiction that made an impression on you may have been Rosie the maid from “The Jetsons” animated TV show. The idea of robotic servants has been around much longer than that, of course, and every year we expect to come closer to finding one available in stores. Well, OK, maybe at Neiman Marcus. But if you were keeping an eye on the recent Consumer Electronics Show in Las Vegas, there really wasn’t much that fits our usual concept. You know: the basically humanoid robot, a similar size to us with roughly analogous limbs and sensors that will do all of the jobs around the house that we don’t feel like doing. A few look cute, but aren’t much use as anything but toys or novelty items. The robots that do useful work are specialized: a small robot that will wash your windows, another that will clean your barbecue grill. Yes, there are scientists all over that are working to develop humanoid robots, but my question is: why bother? Why make a general purpose robot that can “do it all”—like us, only better? It’s a massive challenge, and it isn’t necessary. And from what I see, that’s not the direction things are going.

I have a feeling our cities of fifty years from now will surround us with specialized robots that will each do one thing and do it well. There’s no need for a robot that can clean the house and drive you to an appointment. It’s much more likely we’ll build houses with self cleaning rooms, each with its own Roomba and wipers on the windows (especially with dirt and moisture-repellant surfaces everywhere). And we’re getting closer all the time to self-driving cars. (You can read more about them here.) So-called smart appliances will not only order the groceries but assemble and cook them, too. And brain-computer interfaces will connect us to the internet and, through it, to all of our robotic devices, so we won’t even have to lift a finger to set anything in motion. Why would we need a robotic servant that looks sort of like us?

Child care? Maybe. But a daycare space with one smart computer (or human) in charge and a lot of mechanized baby movers or glorified waldos, for the physical tasks, seems more efficient and more likely. Medical care could use robots, but they wouldn’t have to be mobile. We’d go to them, or remote-controlled gurneys would carry us.

I don’t see a practical need for an all-purpose humanoid robot at all, unless it’s for one of the least practical reasons of all: companionship. (No, I’m not going to get into all the movies and books about glorified sex dolls—you can check them out on your own.) But if it’s companionship we want, don’t make a robot that looks like Rosie. A big, cuddly teddy bear would probably be the way to go.

As for me, I hope we still have dogs.