Great article! GSDMM sounds like an interesting algorithm that could compete with LDA, but there is yet another algorithm called Top2Vec that was introduced in late 2020.

Top2Vec actually generates embedding vectors for documents and words and then clusters these vectors to identify topics. If you're interested, you should check out this article that I wrote about Top2Vec: https://towardsdatascience.com/how-to-perform-topic-modeling-with-top2vec-1ae9bb4e89dc