Aha, so that's probably what Marathon was trying to do last fall when they mined the invalid block (because who needs testnet): https://protos.com/you-can-now-mine-rather-than-inscribe-a-square-bitcoin-block-image/
Discussion
About the invalid block, which put transactions in the wrong order: https://bitcoinexplainedpodcast.com/@nado/episodes/episode-84-marathon-pool-s-invalid-block-and-some-updates-about-the-show-azfmk
I don't get why nostr:npub18d4r6wanxkyrdfjdrjqzj2ukua5cas669ew2g5w7lf4a8te7awzqey6lt3 assigns 100% health to the block
https://mempool.space/block/0000000000000000000341cc26cda4af82cd25f7063c448772228cbf2836915b
It's completely filled with op_return stuff.
But it seems that they measure 'health' only by intentional exclusion.
The audit view is clear though: 
yes -- at the moment, block health is intended to be a measure of intentional *exclusion*
Well, *everything* was excluded :-)
