Absolutely, this ^^^^. Agile just strongly promotes the things which are important to any delivery using any delivery methodology: ability to react to change, fast feedback, small measurable iterations with predicable milestones, open communications and collaboration, outcomes versus tasks and complete products (ie. product + docs + support tools + training etc.) You absolutely cannot impose and off-the-shelf agile framework onto a waterfall organisation and hope for a positive outcome. Using agile thinking (note - not an agile framework) has always allowed me to address the kinds of inefficiencies which result from an organisation which has evolved into distinct silos who no longer engage effectively with each other. Plus the docs are actually of better quality if you get it right.

This is the kind of result you get when clever people are presented with a problem and allowed to attempt to solve it.

Totally agree with this: meetings kill innovation. I've worked over video links with suppliers and customers for years and it does take a little while to get used to. I don't miss awful conference call quality and struggling to listen to a call in a noisy open-plan office. I also find remote working makes it easier for those less dominant to get their point across. Video calls seem to neutralise alpha males (another innovation killer). It's tough for new starters, unless mentoring is more formal - ie. the new starter is invited to everything by their mentor to get going as it's easier to do intros as part of other conversations. Agree that ideas need to be fleshed-out too. Business proposals with clarity and demo material work wonders.

My wife's old Fiesta used to rev uncontrollably occasionally. More often in wet weather. We kind of learnt to drive it that way though until it was scrapped. No "runaway vehicle" type incidents occurred, however, probably due to a thing called the "clutch". Bit embarrassing at traffic lights though.

NCR Unix had (in the early '90s, possibly still has) a "maximum logged in users" limit which was configured via a kernel parameter. NCR customers could purchase upgrade licences which altered this parameter (via a patch + kernel rebuild) to increase the logged-in-user limit. The patch contained no other changes so the OS was capable of supporting more users than the licence allowed, but artificially limited (by software) depending on the licence purchased. Does this qualify as prior art ?