This is tech. We make the future. However, we often get a little ahead of ourselves. Oftentimes the promise isn’t fulfilled as soon or as well as we imagined or vapored forth. Here is my list of stuff that may be good, but perhaps isn’t all that it’s cracked up to be as of mid-2017.
It is ironic that I’d call chatbots a bit overhyped given that I work for a search company (full disclosure: I work for Lucidworks, a search technology firm with products in this area). I don’t mean to say that NLP and conversational search and such don’t have a very bright future, but chatbots will only be useful as an interface to a search engine—as the thing that asks follow-up questions to refine your search to find exactly what you’re looking for. All of the other uses, like the one that tries to sell you something or tries to work in customer service, are just fancy dressed up IVR systems.
What is an IVR system? You know how you hated phone trees and thought “this can’t get any worse.” Well, they made it worse by making it voice driven (badly). Instead of just selecting one of the 12 things the IVR system lets you do from a menu, you have to guess. In the end chatbots will be an interface to search; everything else will simply be another IVR system, and you can just go to the website or app and do that stuff instead of waiting on hold to talk to a computer or the live equivalent.
Frankly, I’ve come to regard HDFS as an evolutionary mistake. Distributed file systems are a good idea and SANs weren’t such a great idea given their cost/benefit. However, as I’ve mentioned before, HDFS is a bad design. It warrants replacement. Also, dumping a bunch of files on HDFS in the hope that something “magical” will happen isn’t a data lake—it is a fool’s errand.
AWS is not inexpensive. There, I said it. AWS is actually kinda costly. AWS does not give you magic fault tolerance. AWS is also not magic. Most uses of AWS do involve some maintenance, and some heavy maintenance at that. About the only sure things with AWS are that you won’t buy hardware and you won’t do low-level network configuration. However, some people just read the latter part and started laughing hysterically after spending all day doing a kind of network configuration. When I say you avoid low-level network configuration, I mean you won’t have to shell into a switch and flash its firmware or something.
4. Machine learning
Machine learning is really just math. Although machine learning is a powerful tool that can make the world better and automate a lot of things that couldn’t be automated before, the popular imagination of what it is and what it can and can’t do is mostly wrong. You can’t just take some random unwashed data set, point the “machine learning thingy” at it, and get a real result. Moreover, even if you do find the right algorithm, wash and configure your data appropriately, and get a result, you have to test that result and validate it and tune it over time. I’m not saying machine learning can’t write poetry or jokes, I’m just saying the poetry and jokes will be terrible until the dude that tunes them does all the real work and gives the machine the credit (note: in modern usage “dude” is gender neutral).
It’s a sick accident of history that this turdball scripting language became the basic assembly language of the web. If you’re just munging up HTML and doing some basic database operations, I think it is fine. Otherwise, I think you kids should get off my lawn until you learn a grown-up’s language.
Mainly this is CNN and friends with the latest scare. Don’t install random stuff, apply patches reasonably on time, don’t forget a firewall, keep backups, and don’t run Windows. There is 99 percent of your ransomware problem solved. Now if someone encrypts your files, just format your hard drive and start over. My most important stuff is all on Google Drive and backed up by both the NSA, the Chinese government, and by the President’s employer.
The industry has failed to make PaaS solutions that are sufficient to serve most customers. Meanwhile, most customers still think they are special snowflakes who cannot set up their web server or database the same way anyone else does. Additionally, all of the SaaS vendors are too busy trying to create “stickiness” (read lock-in) for their platforms and pursue a “platform strategy” in order to “completely own the customer relationship” and don’t have time to make their platforms integrate with anything else customers use. Because of this everyone is doing a whole lot of scripting and calling it “devops,” which means shell scripting in less flexible tools.
8. Apple Keynote
This is PowerPoint by Apple. It is actually a lot like PowerPoint except that cut and paste doesn’t work properly, every file is at least 100MB for no obvious reason, and the drawing and layout tools are more primitive. However, it is PowerPoint by Apple (awwww). I look forward to a web-based future free of client-side installs.
9. This year’s MacBook Pros
They were hardly hyped at all—how can I say they are overhyped? Like five years ago I bought a Dell with 16GB expandable to 32GB and it supported a nearly 4K monitor (too early for the standard so it had different geometry). Now you can buy a MacBook Pro with a faster processor but not nearly double as fast with the same number of cores. It still has 16GB and isn’t expandable. Everyone I work with is planning to build their own desktop with multiple GPUs.
10. Virtual/augmented reality
11. Self-driving cars
I do believe this is the future, but the consumer version in the near term will be basically a stopgap for every place we should have built a train or some other form of public transportation. We’re still more than a decade away from buying one for our own personal use or getting into an Uber and giving a friend’s address as a destination.
I used to work with some people who really started drinking that Kool Aid, and I was like “yeah, so what?” It’s an important technology but it won’t change the very fabric of business anytime soon. Most of business is social and tradition. The problem was never merely technical. I expect to see some adoption... albeit dramatically lighter adoption than you might expect. Just like Bitcoin.
That’s my list. What is your hype curve? Are you sporting an 8GB MacBook Pro wondering why I’m not happy with 16? Hit me up.