I think he was talking about angular resolution. EM wave arrives at different antennas at slightly different time. The phase difference allows for very high resolution of the direction the wave is coming from. [VLBI](https://en.m.wikipedia.org/wiki/Very-long-baseline_interfero...) effectively produces an antenna with an aperture the size of the Earth. To make it work one needs extremely accurate time synchronization and spatial location. In the interview he talked about continental drift of several millimeters per year. Those are the things they need to compensate for.

A sibling comment mentions raster scanning: building up an image one pixel at a time. That is largely true for a single telescope (though it is possible to have an array of receivers taking multiple pixels at the same time).

However, VLBI is an array of telescopes. Using interferometry - measuring the interference pattern of the signals of each pair of telescopes - we get simultaneous data on a variety of angular scales and orientations. As the Earth turns, the sky appears to turn, and more points in the UV plane can be measured. Once you have enough data, you can essentially get an image by doing a Fourier transform of the data.

It's a raster scan. Larger images are built one pixel at a time.