No understanding needed?
"The pairing enables quantum researchers and educators to focus on application development uninhibited by low-level code [...]"
One of the greatest dangers of reliance on tools-based training as opposed to understanding of principles is that outputs from the said tools can not be readily validated. This happened long ago in fields where use of statistics packages have in general superseded informed application of statistical techniques, and the dangers are exacerbated in the AI arena as the transfer functions of the tools are essentially non-transparent. In the absence of a groundwork of understanding about the ways the tools arrive at their results, there are going to be lots of questionable conclusions based on their outputs.