I've recently read "You and your research" (link: http://www.cs.virginia.edu/~robins/YouAndYourResearch.html) about great scientists and become interested: what are actual unsolved problems in CS field for which we have a workable attack surface? It seems like everyone keeps talking about AI and blockchain, but it
have become increasingly complicated to distinguish between hype and actual progress. Answers to the same question but in the related fields would suffice too, e.g.: There's some noticeable progress in the protein folding problem, but it's made via neural networks, if I understand correctly.
It’s important to note the “workable attack surface” part of the question, meaning that P = NP is an unsolved problem, but I doubt anyone understands how to approach solving it.