Mark Friedenbach explains that addresses are primarily a user-interface issue and don't factor into the Bitcoin protocol itself. He argues that newer bech32 addresses offer better security guarantees or lower fees compared to older address formats like P2PKH or P2SH. The discussion then turns to the cost of SegWit transactions, which are generally more expensive than legacy transactions due to the signature framework used. However, they can still be more cost-effective in terms of blockweight and fees for certain types of transactions. Alberto De Luigi raises concerns about SegWit adoption by exchanges, particularly when it comes to transactions with many outputs. While SegWit can save space and reduce fees in some cases, it may not be beneficial for all types of transactions. He suggests that a coordinated upgrade to bech32 could be a consensual solution to promote SegWit adoption while also increasing block size. The email thread includes links to various resources discussing the benefits and costs of SegWit adoption and different Bitcoin address formats.