Ask HN: When will the robot revolution occur?

[Maven911]

The microchip revolution has happened and still going strong, the internet revolution has happened and still going strong.

But when will the robot revolution begin ??

When will the day where instead of having to pay 25$ for a haircut, I can go to an automated barbershop and have it done for much less ? Or we replace our cashiers with an automated money-taking and distribution system ?

From what I have seen, the robotic technology is used in manufacturing plants, researched in industry and a few products here and there like the Roomba.

But I still feel we are far behind, one metric of success will be when the military will deploy robots as ground forces to minimize human casualties, but that still feels far far away. And of course it will still take some time before it becomes affordable and starts replacing “people” out of a job - sorry for the cynical tone but it’s a reality that should happen to stay lean and efficient.

So what are your thoughts on the robot revolution – do you think we will see something successful in our lifetime ?

How do we start preparing, start programming on microcontrollers with servos, motors and actuators ?

Do any of you work in research labs or universities who can provide me some insights ?

[fvryan]

It's already happening, just not in the way sci-fi movies portray it. It's software.

You asked when will robots become barbers and cashiers? They aren't being built for that purpose because they are much more efficient at mental challenges than physical ones.

Is is appropriate to use AI and robotics for subtracting $2.34 from $10.00 and handing over the change at McDonalds? No. It is more appropriately deployed in a payments company to crunch huge amounts of data looking for fraud.

Look at LegalZoom's software and how much it has replaced the need for lawyers when dealing with legal documents.

The advantages to software deployed over the web, to physical robotics are obvious. Software can be everywhere at once and is cheap to replicate. It can be updated and patched easily. The "brain" is not susceptible to the dangers of traversing around and is kept safe and constantly backed up.

I think software is replacing the sci-fi robotics we've seen in movies. But there are still interesting companies working on modern robotics, In fact Anybots is run in the same office as YC and http://www.kivasystems.com/ is another great example.

If you are concerned about robotics replacing "people" in jobs, then rest assured barbers and cashiers are safe and sound. It is scientists, lawyers, doctors, and engineers who should be concerned.

Another interesting thing to watch is whether software can be creative and compete with designers, artists, and writers.

[robotresearcher]

| Or we replace our cashiers with an automated money-taking and distribution system ?

That's an ATM. We did it already for 90%+ of cash-handling interactions.

| But I still feel we are far behind

Far behind what?

It's hard to find things that robots are good for when you factor in the cost. When you optimize for price/performance, you get for example your dishwasher. It's a robot that does 90% of your dishwashing, for cheap. The other 10% is the tricky stuff and you can't afford the robot that can do it. Even if we had the software for cheap, machines that can do accurate manipulation with a strong grip are expensive to make, requiring high-quality motors with strong magnets, for example.

The current killer app for robots is manufacturing. They already do a lot of it, and have done for 20 years. You just don't see them.

The next huge killer app for robots is the self-driving car. This will come eventually, and it'll save tens of thousands of lives every year. It'll come gradually, and at first it will drive you 95% of the time while you handle the tricky bits. This is how Google does it and they have driven tens (maybe hundreds now) of thousands of miles around California.

Otherwise, you'll soon see flying robots doing geophysical surveys, checking power lines, etc. This is current done by people in helicopters and is expensive. UAVs are getting very cheap and we can control them well.

There are opportunities too in labour-intensive business like order fulfilment. Amazon uses lots of people and could replace some of them with robots like Kiva Systems. Amazon is interesting since they will probably do it cheaply and cleverly, now that Kiva has invested and shown that it can be done.

In summary: the robots you see every day are ATMs, dishwashers, etc. The more robot-like robots are hidden away working for you. Down the road you'll see automated cars. You won't think of them as robots either.

[robotresearcher]

| How do we start preparing, start programming on microcontrollers with servos, motors and actuators ?

No, get a CS degree and then go to grad school. Mechatronics is cool and all, but the main advances in robotics in the last 20 years have been in software based on exploiting new state estimation techniques. Stanford's robot car was just a normal car and computer.

On the other hand, you could work on battery chemistry. Advances in batteries will make robots better, along with everything else.

[Maven911 → robotresearcher]

Hi, sorry for the delayed response, thanks for your detailed responses, I don't want to go to school again but is there specific classes you recommend like AI, microcontroller programming etc ?

Do you work in the industry, you sound like you do...are you working on Google's self-driving car per chance : )

[s2r2]

> When will the day where instead of having to pay 25$ for a haircut, I can go to an automated barbershop and have it done for much less ?

It will probably take some time before a robot is cheaper than 4 EUR per hour (hairdresser's collective agreement in Saxony, Germany).