For the record: in mid-2023 I implemented Whisk on Lighthouse and successfully ran it on a small devnet.
I noted a few issues caused by the bootstrapping phase of Whisk. In summary:
- Forced missed slots. Current spec will induce ~2% missed slot rate on the initial epochs after the fork ~ 12,000 cumulative missed slots during the first year post-fork
- Expensive duty discovery for validator clients. Current spec forces validators into doing hundreds of thousands of G1 multiplications per slot and per keypair (attached active validator index)
Note that both issues can be solved if there’s no bootstrapping phase. This proposal uses fancy cryptography to get rid of it The return of Torus Based Cryptography: Whisk and Curdleproof in the target group