This referendum proposes carrying out the first step proposed by Daniel Olano (@olanod) in Adapting Kusama to demand and its JAM future: reducing the Kusama validator set size from 1000 to 700.

In addition to the long-term goal to have Kusama operate a lightweight version of JAM, as expressed in WFC #573 and approved by the Kusama community, this reduction will also provide relief to Kusama validators given the increased difficulty in operating a Kusama validator economically - especially operators of singleton validators.

Adapting Kusama to demand and its JAM future proposes a multistep reduction timeline to 120 validators over a 6 month timeframe.
This referendum is not a commitment to the full timeline, but only the first step.
The reasons for staging only the first step for now are:

1. It removes the current over-provisioning: reducing the validator set to 700 aligns it with the paravalidator population of 700 (for 140 cores) that Kusama actually makes available today. Any further reductions would require reducing the available cores.
2. While there may be broad agreement that a reduction to 700 is sensible, there could be diverging opinions on the final validator count the network should land on, and how many cores the network should provide. In particular, neither Polkadot nor Kusama have ever reduced their validator set sizes to date, so it's also prudent to gather relevant data from this first reduction's impact before proceeding with any further reductions.

If this referendum is approved, after a few weeks, we can assess the impact of reducing the validator set to 700 and decide whether to proceed with further reductions as outlined in Daniel's plan, or alter the path based on data and community feedback.

For the preimage and integration tests, see https://hackmd.io/@Lederstrumpf/referendum-kusama-validator-set-reduction-to-700-runtime-2_3_0