1. NASA manager Steve Stich said there's a relatively wide "band of uncertainty" in how risky a Starliner return is. Some (many?) NASA engineers are at the high end of the band and are advocating a return on Dragon instead. Boeing is obviously at the low end of the band and thinks it is a low risk.
The problem is, the data doesn't rule out either side of the band. So they are trying to get more data to narrow the uncertainty (in either or both directions). [Interestingly enough, the data from the White Sands testing made them more worried because it revealed the Teflon seal deformation.]
But my sense is that if they don't narrow the uncertainty (i.e., convince the NASA engineers) then they will very likely choose a Dragon return. That is, it sounds like if nothing changes, the astronauts are coming down on Dragon.
2. Stich said they need to decide by mid-August, in order to have time to prepare the Crew-9 launch for Sept 24th. So we'll know by then.
3. They emphasized that (a) the thruster problems are all fixable (given time), and (b) that even if Starliner returns without a crew, they will have learned enough from the test to potentially certify the capsule for regular service. This is probably the only way they'll be able to keep Boeing as a provider. A redo of this mission would cost Boeing half a billion dollars, easy. And since the contract is fixed-price, this would just add to Boeing's losses. So I expect they will certify Starliner even if it comes down without a crew.
4. In some ways, Starliner is being held to a higher standard than Dragon Crew-2. If Starliner were the only vehicle available, NASA and the astronauts would absolutely take the small risk and come down with a crew. But since Dragon is available, I think NASA is thinking, "why take the risk?"
5. There's a huge difference between how NASA engineers and lay people look at this issue. Many people (particularly on Twitter) have a binary safe/not-safe view of the situation. Either Starliner is safe or it is not. Either the astronauts are stranded or they are not. But the engineering perspective is all about dealing with uncertainty. What is the probability of a bad result? Is the risk worth the reward? Even worse, everything is a trade-off. Sometimes trying to mitigate a risk causes an unintended effect that increases risk (e.g., a bug fix that causes a bug).
I don't envy the engineers, either at NASA or at Boeing.