I just think that there needs to be significantly more caution - almost a pathologically high amount of caution.
Think of it from a software engineering perspective. All the unit tests that are done to ensure 100% code coverage, all the testing in identical staging environments, and yet we still have to revert sometimes, once the code is in production.
And this all still happens in environment where the code is something we wrote ourselves, is properly and thoroughly documented, with all dependencies and relationships clearly defined. And still, major problems happen but it's nice that we can just go to Git and try again.
With genetic engineering, the genetic code was not written by us, has zero documentation, the relationships are left for us to figure out, and there's no proper staging environment with a working ecosystem for us to test things in. We basically deploy straight to production after a few unit tests. So, yes, definitely I don't think there is currently enough caution being taken.
Of course, the reality is that genetic engineering won't stop, not until we shoot ourselves in the foot a few times and people realize the seriousness of it.