Everybody has heard of Einstein lensing: a galaxy in between you and some bright object in back of it will bend the light from that object, sometimes into a ring shape. Note that the light you see didn’t come “straight” from the object—the path is quite a bit longer than it would have been without the intervening galaxy in the way. In the image shown at that second link the light probably took a thousand years longer than it would if the near galaxy weren’t there.
That’s one key to the story at the first link. The other key is that just as a galaxy can bend light a lot, a large star or black hole in the galaxy can bend it a little. Maybe just a day’s worth.
That’s what the authors looked for in studying blazars. There were only a couple of candidates, but as one of them flashes in different wavelengths (gamma rays, radio waves, etc) they found that there was an echo for the gamma rays, of about 10 days. This is presumably from “microlensing” by something almost in the line of sight between us and the blazar. That’s cool by itself, but from the estimated size of the “microlenser” you can estimate how big the region was that the gamma rays came from. This is impossible with normal telescopes and utterly impossible with gamma ray telescopes—they just don’t have the angular resolution. But working back from the “microlenser” size they estimate that the gamma ray creation region is of order 10^14 cm or less. If it were igger you wouldn’t get a nice “echo”.
It is a bit like the way stars seem to twinkle in the night sky but planets don’t—the planets subtend a larger solid angle than the almost point-like stars do, and so our atmospheric variations don’t change the apparent direction much. In this case the gamma rays get lensed into an echo/flicker, but the radio waves, coming from a bigger area, don’t.
This isn’t exactly a surprise, as most models predict that gamma rays would be produced near the base of the jet from a black hole, but it is a clever analysis.