Except that for someone outside CS theory it will be hard to tell what is routine and what is ground-breaking. The proceedings for the recent editions of top theory conferences are here:

- FOCS (Foundations of Computer Science) https://www.computer.org/csdl/proceedings/focs/2022/1JtvJWOs...

- STOC (Symposium on the Theory of Computer Science) https://dl.acm.org/doi/proceedings/10.1145/3519935

- SODA (Symposium on Discrete Algorithms) https://epubs.siam.org/doi/book/10.1137/1.9781611977554

They're mostly paywalled but conference or full versions can usually be found on arxiv.org, and there are successful efforts to host paywalled papers illegally for the public good.

Personally, I am impressed by the progress on approximate counting and sampling. Developments include localization schemes (FOCS), average-case entropic independence (FOCS), the sampling/counting Lovász local lemma in (FOCS and SODA), and approximate counting with global constraints (STOC).