Looking Back On My 2018 Predictions

For T-SQL Tuesday #100, Adam Machanic’s challenge for SQL bloggers was to predict what databases would be like 100 months from then.

Well, guess what: that’s coming up this July!

I wonder how many of the original bloggers are still even active in the community, and I’m curious to see how everyone’s predictions went. I’ll jump in and go first, since I have no shame, hahaha. Here were my original predictions, and let’s talk about how I did.

2018 Prediction: In 2026, DBAs will still be a thing. Back then, I came under ridicule because some people thought there was no way database admins would still be necessary or around. Well, whaddya know: we’re still here! Score: 1 point out of a possible 1. However, making a prediction like that for 2028-2030 becomes a little tricky: all kinds of tech jobs are in wild flux right now with this AI thing.

2018 Prediction: The data safety business will look like the car safety business. I said that countries would have their own different data protection and residency regulations, and that our work would be harder for that lack of standard. Boy, did this one come true: the US government forced TikTok’s sale to keep the US data & algorithm inside the US, a whole slew of new data laws went into effect on Jan 1, and data localization laws by country are a bewildering mess. I truly pity anybody who has to make a living dealing with this bureaucracy. Score: 1 of 1.

2018 Prediction: <= 5% of SQL Servers will run on Linux. Back then, the Windows Core crowd was super-excited about SQL Server on Linux because it’d get by with even less surface area and less memory requirements. For companies, this prediction is clearly true: the adoption rate is wildly low. If Microsoft ran Azure SQL DB atop Linux, you could argue that >= 5% of databases run on Linux, but that’s not what the prediction was about. Score: 1 of 1.

2018 Prediction: Your developers will have several projects built with serverless architectures. I honestly don’t know what the overall population’s results are for this, but at least for the majority of the clients I work with, this is true. For example, if you’re using Microsoft Fabric, you’re using serverless. Still, though, I don’t really give this one high marks because I said several projects, and I’m not sure that’s the case for most companies. I’m gonna give myself a low score here. Score: 0 out of 1.

2018 Prediction: Your default new database will be in a cloud PaaS. I didn’t make this one as measurable as I should have, given that I wanted to be able to go back and measure my accuracy rate. In the 2026 Data Professional Salary Survey, the majority of responders are using the cloud in one way or another, but the primary database for most responders is definitely not Azure SQL DB. Score: 0 of 1.

2018 Prediction: 2 big clouds will offer an MSSQL-compatible serverless database. This one already came true several years ago with the advent of Azure SQL DB Serverless and AWS Aurora Babelfish. Score: 1 of 1.

2018 Prediction: Microsoft will fix the “String or binary data would be truncated” error. Fixed, amazingly. Score: 1 of 1.

Overall score: 5 of 7. I’m happy with that, but hoooweee, I would not wanna try to predict what the database business is going to be like another 100 months from now! Next, check out Adam’s 2018 roundup of blogger predictions, and have a few laughs about where our minds were at back in 2018. It’s amusing to note that some of the predictions could have been written last week – some things always seem like they’re going to happen real soon now™.