I recently watched Alan Kay's How to Invent the Future (and part 2), and one thing I thought was particularly interesting was his idea of trying to invent the future from the future. That is to say, rather than extrapolating from now to the next 10 years, extrapolate to the next 30 years and then work backwards. That way you're finding an incremental path to a big vision, as opposed to ending up with a small and incrementalist vision.
In that spirit, I thought I'd come up with some 30-year predictions:
Memetics is redeveloped as a practical discipline, becoming the primary driver behind development of entertainment (games, social media) and social influence (politics, PR). Dopamine-farm distraction technology becomes worryingly effective. Deep questions begin to emerge about who is actually making decisions when we're all so easily manipulated on such a large scale. Anti-memetics and non-democratic forms of government become serious areas of consideration.
Software development splits into distinct disciplines, both vertically (architects vs engineers vs builders vs technicians) in its current general-purpose focus and horizontally away from general-purpose towards different special-purpose forms for different domains. This isn't so much DSLs as distinct paradigms like Excel formulas, Labview or MaxMSP. Most people program computers if you consider a wider definition that includes this "soft" programming.
Elimination of all passive media, ie all documents/images/movies etc become executables containing both the data and the code to interpret them. This is partly for DRM and partly to enable new interactive forms. Sandboxing and trust models replace Turing-restriction for software safety. (You can already see this happening on web and mobile). Most media becomes dynamic or interactive in some way, even if the main experience is still non-interactive, eg movies with an embedded chat box.
Human interaction becomes almost entirely machine-mediated, leading to people accustomed to complete control over their social interactions. Huge industry of asymmetric transactional relationships (livestreamers, social media personalities, personal services etc) with defined parameters. This leads to a healthy but small market of countercultural "talk to someone you might not like without being able to immediately get rid of them" services and quaint local-community-based interaction. Most people will be happy enough only interacting with others in controlled circumstances.
Semi-autonomous companies dominate the market as machine learning algorithms demonstrably trounce human judgement in investment, management, and executive-level strategy. People are still involved to do work that is not currently automatable, but that field shrinks from the top as well as the bottom. There is a commensurate increase in value for creativity, which continues to elude AI. Most people are not employed, either spending their time on human-centric creative pursuits like art or sunk deep in a well of machine-optimised distraction.