Earlier this year I had a blog post about how writing software applications in-house is making a comeback. I was reminded of this when I attended the HackNY hackathon at NYU over the weekend. I was there to help Susan with representing Ziggeo, her startup, which provides an API for video recording and playback. The number of APIs and open source projects out there that do amazing things is simply staggering.

Watching some other NY based startups present their APIs and feeling the excitement among the students I felt inspired to write a little hack of my own. Basically the idea was to let anyone record a video and then analyze its contents. Without access to big and powerful services this could be a project for many engineers to work on for many years! Here it wound up taking me a couple of hours.

I used Ziggeo for the recording and video hosting. I then had a bit of Python code to pull down a video from Ziggeo after recording and cut it up into frames using the open source ffmpeg. I then submitted the frames as images to Clarifai for image analysis and reported back the results. You can try it out here. It is fun if you hold up different objects in front of the camera. We have even recorded our dog. There is a lot more one could do to present the results more elegantly but I was pretty pleased for a few hours of work.

Why does all of this matter? Because right now there is a fallacy in Washington and elsewhere that there will be a nearly unlimited future demand for programmer. There is an idea that we just need to retrain as many people as possible and this will solve some of our labor market woes. With software eating software though any extrapolation of current hiring needs is likely to vastly exceed the actual needs.

This is not to say you shouldn’t learn how to program. I have written before about the importance of programming as a way of understanding. But neither individuals nor policy makers should see this as a silver bullet for employment.