I haven't thought about it that way before. "Can we trust it enough to run unsupervised?" Scary, but I guess that's the future.

I like how the comparison table frames this. It's clear at a glance why you can't use the same solution across different NoSQL databases. The DynamoDB row is the one I find most interesting because I think teams coming from PostgreSQL tend to underestimate how different TransactWriteItems really is from a true ACID transaction.

Interesting article, but it has nothing to do with your "child"