How tall can you build with 54 Jenga blocks?

The assumption that the tower would fall over at a certain floor tilt seems to rest on the premise that the opposing sides of a Jenga block are perfectly flat and parallel, and orthogonal to one another.

If they aren't then the blocks can be rotated relative to one another to cancel out floor unevenness.

A problem: Jenga blocks have a nominal size, but in reality they vary quite a bit. Virtually any tower from a standard set will have some inherent lean to it. If you were really diligent you could use this to maybe offset lean as you go. If you build naively, you are doomed faster. This is part of the reason why some blocks seem stubbornly stuck in the tower stack and others are loose.

Can’t some of these easily be improved upon?

For example, for the 12.5°, 15° and 20° ones, replace the two blocks in the top layer by the construction used for the top two blocks in the 25° record.

Also, I think “A lone upright block tips over at 11.3° in its thin direction” doesn’t imply “no matter what you do to the base, you can’t stand blocks up vertically at high tilts.”

I would think you could sneak in cage-like structures like these (seen from the top)

   ABB   A
   A D   ABB
   CCD  CCD
          D

with blocks A and D taking most of the north-to-south loads and B and C taking most of the east-to-west loads.

Possibly that will only work near the base, where there is more pressure from above that prevents these blocks from falling over.

Possibly, such a construct with only three (at 120° angles) or two blocks could work, too.

Jenga blocks are intentionally non-uniform thickness.