Going against the current for a couple of things, I guess, but there are two points there that I actually like, but probably not in the way they are usually implemented.

• 4 hours planning is bad if it's 4 hours sprint plannig. Then somebody should be sacrificed for the greater good, but for the love of god, can somebody take their time to plan features and architecture carefully before dumping stories for development? Smaller group and whole long term system overview. Let them spend 4 hours to make something coherent and elimitate tech-debt creation.
• 1 point = 1 hour is fine if you don't think that it's set in stone, but it allows for very easy sprint planning. If your ballpark estimate for the sprint is more than 20 hours/developer for the week, you are overplanning. Normalize creating tasks for thinking how to solve a fucking problem instead of saying it's complex.