Perhaps whoever posted that should be corrected, then. Also noteworthy: there are transactions from before bip16 with outputs that match the bip16 template, like this one: https://mempool.space/tx/9c08a4d78931342b37fd5f72900fb9983087e6f46c4a097d8a1f52c74e28eaf6

And some of them were spent, including that one: https://mempool.space/tx/6a26d2ecb67f27d1fa5524763b49029d7106e91e3cc05743073461a719776192

Per BIP16, I think they should be treated as simple hashlocks, i.e. if they have a redeemscript, it should not be evaluated, as bip16 wasn't active yet. But block explorers like mempoolspace display these txs as if they are P2SH spends, i.e. they DO evaluate their redeemscript. (See screenshot.) I suspect that's a bug.