Some excellent comments already here. I'll just add a bit of my own thoughts (and seconding some great suggestions from others).
- Scalability, scalability, scalability. Nodes are struggling with the increasing demands of bigger and more expensive hardware as the blockchain grows and it grows a lot (roughly a gigabyte per day?). People are shutting down nodes and relying on a small number of public API endpoints that make the entire ecosystem very fragile. Witnesses are relying on shared service providers and this problem has the potential to get a lot worse. This is unsustainable and dangerous.
- Simple UI change to show or hide resteemed posts in a blog. If I want to look at someone's posts and find one, it is nearly impossible to do if they are a frequent user of resteem.
- Recommended posts (a form of algorithmic feed). This used to exist but was removed as unscalable when hosted within steemd. Now that there is a multi-layer architecture and all requests not hitting steemd itself, please bring it back.
- Communities (perhaps a temporary work-around is to be able to follow a tag; I've noticed some tags are working okay as mini-communities now).
- Traditional currency gateways so non-crypto users don't have to use crypto exchange to buy Steem Power and cash out their rewards. I understand there are regulatory issues. Partner with service providers if necessary.
- Deep revamp of reputation and curation into a better voting/reputation/curation/web-of-trust type system. This is a complex problem and suggest rather than trying to solve it by tinkering and tweaking as has been the case in the past (and present, afaict), to really undertake to study it and conduct the necessary original research. This means hiring the right multidisciplinary researchers (keyword soup, probably incomplete: economics, game theory, mechanism design, auction theory, voting theory, fraud detection) either inside the organization or outside by sponsoring academic research. I doubt it would produce usable results in 2018, so the 2018 task would be to start on a multi-year effort.
I’ve created a word cloud for the comments section of this post that I hope you will find of value and will want to share. While the creation of the actual word cloud is relatively easy and straightforward, the creation of a relevant word list that will produce a meaningful word cloud is quite another matter, as I’m sure you know, and takes considerable time and effort. Unfortunately, as of now, the post has only had 41 views. It needs a little love. Perhaps a resteem might be of help, if you think it is worthwhile, and, of course, it would be much appreciated too. TIA
https://steemit.com/steemit/@cryptographic/word-cloud-for-steemit-roadmap-2018-community-input