Here's a recent, modern application of time-lock puzzles to a private voting scheme. Developed by a16z crypto:

https://a16zcrypto.com/posts/article/building-cicada-private...

If you're interested in time-lock puzzles, drand [1] achieves this but through randomness instead of computational effort.

[1] https://drand.love/

Pretty cool though the Rivest one got broken a few years ago (ahead of time and quickly on new hardware!)

I think the impact of advancements in computing could be mitigated by giving certain guarantees about the best-case and worst-case scenarios for computing advancements. For example, rather than saying "this will take 35 years to crack" you could say "this will decrypt in 20-45 years, depending on how much computing efficiency improves".