Why doesn’t Bitcoin use bech32 encoding for transaction_id and block hashes?

If I understand correctly, transaction IDs and block digests are usually encoded in hexadecimal in the Bitcoin ecosystem, including Bitcoin Core RPC calls, block explorers, etc.

On the other hand, bech32 encoding makes the string smaller and also provides a checksum.

I wonder if there is a technical reason for hexadecimal encoding, or if it’s just a status quo situation.

That is, if Bitcoin Core were written from scratch now, would we likely choose hex, bech32, base58, or something else for hash encoding of blocks and transactions?