I have really been enjoying the Real Time Conference these last two days. I gave a keynote this morning and then participated in a really interesting panel and an on-line discussion. In a little while I am going to participate in an evening panel.
Being a presenter in a conference that is held entirely on-line is a fascinating experience. Of course there have been technical glitches, because all of this is new. But on the whole, I very much appreciate how people from around the world can share in the experience, without the great expense, inconvenience and difficulty of physical travel (not to mention the high ecological costs of physical conferences).
I am now watching the session just before the start of my own evening panel. A lot of really smart people are discussing “deep fakes”, algorithms which create very convincing visual simulcra of reality through the use of machine learning.
But my favorite moment came when Andrew Glassner started explaining how many of these algorithms need massive amounts of data, so they must rely on huge corpuses. But I was only half listening, so I misheard.
I admit that part of me was disappointed when I thought back on what I had just heard. Only then did I realize that Andrew was not talking about huge porpoises.