At 8:43 I talk about the difference between "bonded covenants" (which bitvm gives us) and "real covenants." The short version is, bonded covenants aren't as good as real covenants because they introduce an additional trust assumption, but they are probably a "close enough" substitute for real covenants in many contexts. IMO the real thing would still be better.

At 10:57 I also talk about how if we had real covenants we could improve bitvm's trust assumptions in cases where you aren't part of the contract. Bitvm follows trust assumptions similar to the lightning network if you have a direct bitvm contract with the prover. But if you don't, you have to trust someone who does. With real covenants, we could improve bitvm so that it works well without the need for a direct contract with the prover. See the video for more details.