2018 Tech Predictions
I haven’t done this for a few years, but now is as good a time as any to throw out there some technology predictions for 2018.
- Voice-enabled App Explosion – 2017 was an amazing year for sales of Google Home, Amazon Echo/Dot and many other voice-based devices. The ecosystem for creating apps on these devices, however, is still pretty new. However, you are starting to see some super early adopters figure things out. For example, the KAYAK app for Google Assistant is really good. Expect many more like that in 2018 along with a lot of amazing innovation in the underlying platforms. It won’t be long before getting your business on these voice-based platforms will be as important as native mobile platforms.
- Machine Learning Goes Mainstream – Similar to GraphQL, the interest among developers to learn ML currently far outweighs the actual legit production usage. However, even through this past year I saw I significant change in the quality and quantity of practitioners who attended my Boston AI Meetup. The knowledge base and tooling for ML is increasing and getting better every day. For example, if you haven’t already, you should check out MachineLabs which makes it really easy for newbies to get started and for anyone to share their ML models and code.
- Angular Finds Its (New) Groove – While I don’t think Angular will ever dominate mindshare like it did in 2013, I do think it will pick up steam for people in two specific situations: 1) when you want to have one team that builds one app that runs on multiple platforms 2) when you would rather buy into one end-to-end solution for all aspects of frontend development instead of piecing together your own custom stack of disparate technologies. There is a lot packed into this that I will have to expand on in a future blog post.
- SEO Gets Even Harder to Hack – I have spent the past month focusing a lot of energy on SEO for GetHuman (after taking a couple years off). At first I was really shocked at how bad Google has gotten (compared to 3+ years ago) at picking out obvious spammers (for certain search terms). Then I started reading about how Google is starting to rely more on Google RankBrain (an AI-driven system for processing search results). It typically takes a long time and a lot of data to fine tune large, complex Machine Learning (ML) models. When you see incorrect results with a ML model, it is not always easy to get rid of those results. You have to tweak the model in small ways so that you get rid of the bad results without adversely affecting the good results. Despite this challenge, I know many amazing engineers at Google like John Mu are hard at work on this problem. By the end of 2018, I predict there will be a major breakthrough and spam (i.e. low quality search results that are often trying to make a quick buck off you) will be eliminated from most search results once again.