I haven't heard this fork discussion but it sounds like more gaslighting. This is still based on the false premise that relay filters don't work and only consensus validity matters.
Soft forking for spam mitigation would be a mistake IMO - There's no way to play the "cat and mouse" game adequately if you have to have 3-5 years to design and activate a soft fork every time. It should only be done if it's a fork that probably removes the spam permanently and probably will never need to be updated. Otherwise we're just adding tech debt.
