While the current curve is definitely bad and staking ratio is already that big that it should be addressed, decreasing ROI together with issuance (making staking less attractive at all) will hurt solo stakers more than LSTs. Applying alone at the fork it will axe solo-stakers completely which will mean lower security for the protocol.
We need some measures along with this to benefit solo-stakers. It could be:
a. Spread rewards according to the validator ordinal number. If we should have 2.5% reward, it will mean #1 validator receives 4% and #1,000,000 1% (ordinal is not current id, it doesn’t count gaps). This change cannot be applied together with validator indexes reuse. It benefits not only solo-stakers but at least it benefits them more and enough to stay. And it’s deterministic. Until we have a way to change existing validator private key, selling existing validators is not an issue. It is also close to the concept “reward rate is fixed forever at the time of joining”.
b. Every attestation includes voluntary reward decrement with number of validators on this node. Honest nodes count unique validator ids per node and if decrement is lower than the real number of validators, such nodes are banned.
c. anything else benefiting solo-stakers
While I dislike both example measures, at least they directly address solo-stakers’ importance. One curve for all will always benefit LSTs more than solo-stakers whatever it looks like. There was no way found in the real world to tax bigger parties and small subjects with the same rate, while benefitting smaller subjects more.