Google’s Teachable Machine lets you teach a machine with nothing but your webcam.
By Eric Limer
Everyone from politicians to tech companies themselves are pushing the “learn to code” agenda. On the one hand, it could create a more employable tech workforce. On the other drive, it may serve to drive down the wages for what is today a very high-paying job.
But as machine learning continues to progress, you kind of have to wonder how long it will be before coding, in the traditional sense, is an ancient art form. As Google’s new machine learning demo the “Teachable Machine” illustrates, you can already program a robot brain with nothing more than your webcam.
The Teachable Machine is a limited, but incredible demo. And its premise is simple. Turn on your webcam, teach the robot brain what different motions and activities look like. Then tell it what kind of responses you would like. In a matter of seconds, you can train the computer to react with different sounds when it sees different faces or, as I did, teach it to respond to hand motions with a series of different GIFs.
This is how Teachable Machine works:
Obviously there is plenty of traditional (and complicated) code behind this demo. For the foreseeable future that kind of heavy lifting will be required. After all, the computers can’t just learn on their own. Computers need to be taught how to learn, and built with learning in mind.
Still, it’s easy to see how this sort of basic machine learning could make the gadgets of the future wildly easy to “program” with no sort of coding experience at all. You can imagine a world not too far off where smart home commands like “Always turn the lights on when I walk into the living room” or “Always open the doggy door when Spike sits in front of it” could work seamlessly and effortlessly. It’s an exciting prospect!
While that future might seem positively thrilling in a sense, it does come with downsides. All the microphones and cameras required for such a system would mean the computers are always listening and watching… Who knows who else might be tuning in. It’s a vision of the future that flips from utopian to dystopian depending on the lens you view it through.
One thing is for sure— as the machines continue to learn how to learn, the ways we interact with them will change forever. Hopefully it is for the better, not the worse.