Spelling errors are not the most difficult thing to solve these days, we need more corrections based on context, which machine learning is clearly having some advantage. LanguageTool doesn’t participate GSoC since 2018 and I’m sure that with GPT-3 and all the hype around transformers in the last years (just check the number of startups playing with NLP in the last years) made natural language processing much more attractive. It would be very nice to see the community trying to improve it again.