I disagree – for the foreseeable future, hard forks are needed to implement scaling improvements. What needs to be set in stone is which changes we'll approve when fork time occurs.