He highlights a critical point that in a future where transaction fees are high, CPFP may be less optimal than simply replacing an unconfirmed transaction with RBF. This is because CPFP requires an additional transaction and is only possible if the initial transaction includes a change output and meets certain network policy checks.

Folkson references Antoine Poinsot's analysis on Bitcoin StackExchange, which asserts that RBF would be superior to CPFP in most scenarios if mandatory BIP125 signaling wasn't enforced by nodes. Poinsot notes that CPFP is less efficient and can sometimes be impossible, for instance, when the parent transaction's fee rate falls below the minimum accepted by most mempools, making it ineligible for CPFP at that moment.

Furthermore, Folkson poses a query regarding whether CTV-based Lightning Network Symmetry (LN-Symmetry) suffers from the same limitations as CPFP compared to an ANYPREVOUT (APO)-based LN-Symmetry. Theoretically, an APO-based system could adjust fees in every channel update to align with current market rates. In contrast, today's LN without Symmetry relies on justice transactions that may have outdated fee rates if the market fee rate has significantly changed since their creation.

For those interested in delving deeper into Bitcoin's workings, Michael Folkson provides his contact details and suggests a resource, Port of Bitcoin, for learning more about the topic.