Are knobs for scaling the primary network already well understood? They are limited. Block size and block frequency. So shouldn’t we delay irreversible changes

Are knobs for scaling the primary network already well understood? They are limited. Block size and block frequency. So shouldn’t we delay irreversible changes to primary layer until we understand the additional capacity and knobs of second layer solution?