There’s an awesome tool called Loop Findr. It takes a video and finds the most loop-able sections that could be put into animated gifs. I’m not sure if it allows you to also make custom GIFs (haven’t figured that out yet), but it does find some interesting things. There are some implications to this as well, as outlined by this article.
I wonder what, if equipped with facial recognition, uses this technology could enable. The author in the linked article suggests that maybe you could use it for earthquake detection. Plugging this into radars or surveillance could be interesting.