The debate in AI really comes down to two things:
Whether you think AI will be beneficial to society, dramatically boosting our economy, curing diseases and solving unsolved problems while having more narrow bad side effects/downsides like all things in life, but that the bad is mitigable with engineering and precise, clear targeted laws after demonstrated real harm in the real world.
OR
You believe in the Precautionary Principle and that unrealized future harms that may never come to pass are a suitable way to make policy and that you think AI will be a Marvel movie evil villain, uncontrollable, unknowable, alien and that it will have its own agency, Darwinian drives and that we are helpless and powerless and can't do anything from an engineering perspective or legal perspective and life is hopeless.
In the first, you let the normal flow of history and society and democracy move us forward until we find a level and in the other you slam the breaks, crash the economy and put a lot of self-important self designated "special" people in charge to "steer" us forward (let's stop pretending its a "pause" when there is no real criteria of unpausing other than "we found a way to make it safe by not building it" which is impossible because you can't make an engine safe without building it and iterating).
If you correctly discount or delete the possibility of x-risk then in every other branching future scenario our life is materially worse:
1) Many uncured diseases still killing people
2) A maze of bureaucracy
3) Economic crash that hits you and everyone else and grandma's pension
4) Non-democratic powers taking pole position in the world
If you believe in x-risk then you believe in "by any means necessary" no matter how horrible those actions to prevent an imaginary future disaster are to the world and to today's people.
History is rife with people trying to prevent imaginary disasters and inflicting massive pain on society.
If you're actually serious about probability and real risk assessment and balance, then the only path forward is the one with mostly good futures and the greatest range of good possible outcomes.
And if you're not and you believe you can predict the end of the world (you can't), then you pick the ends justify the means path and we are all worse off for it in almost all the branches of the Monte Carlo simulation of branching futures other than the distantly remote x-risk future.
There's really only one clear, sound, sane choice here.
Choose wisely.