My guess is failing to read between the lines and taking marketing too seriously.
"Removing the need for postfabrication qualification" doesn't mean "perfectly detects errors" it means "detects errors within our business specs and we expect it to be profitable within that margin".
In a typical rocket the payload is the most expensive thing, followed by the engines, and then the rest of the rocket and the fuel. Shaving costs off the engines is well worth it, but not if it sacrifices reliability.
Maybe they have a model in mind where it works. If you use dozens of engines like SpaceX's Starship you can tolerate more engine issues. Of maybe they want to launch really cheap payloads on inexpensive rockets. But in the parameters of traditional rocket design, QA on your engines is one of the last things you want to save money on.
Eh sounds like lying to me. Maybe it sounds like marketing to people who develop software “we can iterate no worries” but I think rocket science has a higher bar for safety.
Reminds me of that failing nuclear startup that tried “fake it until you make it” with the US government.
"Removing the need for postfabrication qualification" doesn't mean "perfectly detects errors" it means "detects errors within our business specs and we expect it to be profitable within that margin".