Recently, DevObsessed founder Mark Ruch and I sat down to discuss the trends we're seeing in technology, and understand why now feels like such an unprecedented time. Things used to seem obvious, we mused. Are we just getting old?

Four Technology Tsunamis

Reviewing the projects our consultants have worked on (and, indeed, are still working on), we noticed that there is a significant subset of our work which isn't centered on  innovation, per se. Instead, this work focuses on leveraging technologies (sometimes emerging) which make access to data easier and more ubiquitous.

Broadly speaking, we identified four major shifts over the last twenty-ish years.

1. The move from mainframe to GUIs
2. The move from web apps to service-oriented architectures (SOA)
3. The move off the intranet and into the cloud, culminating in mobile
4. The move of data into enterprise data lakehouses

Editor's Note: Please forgive me for using the term enterprise data lakehouse.

All four of these waves share two common themes: each one serves to make data progressively more accessible and more usable, and each one seemed obvious at the time.

The Outliers

In our discussion, we observed two contradictory things:

1. We technologists viewed all four technology tsunamis as waves that enterprises must ride in order to be successful - they had an air of inevitability about them
2. Many of our biggest, most successful clients chose to wait until long after the hype train had passed (choo choo!) before even considering adopting these technologies

Across the latter, we found that those companies are more focused on market position and relentless customer service than they are on being perceived as innovators.

So What's The Next Tsunami?

It's something about AI, right?

Well, maybe. We're certainly being told that it is. But it takes some imagination to get there. When we use a green screen mainframe application, then someone presents us with a graphical desktop, we see that the GUI is the next step (though my friend Paul Cramer would note, you can often F10 your way to success faster than you can click or tap). Work-life balance be damned, we don't need to be told how useful mobile access is.

From there, things are harder to visualize. If the enterprise data lakehouse (ugggh) lives up to the hype, then maybe we've reached peak data access. Sure, there are marginal improvements to be made in the usability, performance, security, and cost of these systems. But they definitely not vaporware - they're being used with increasing regularity to expose real data and provide real insight.

Indeed that's what we've been building toward for decades - ubiquitous access to data.

My hot take for the horizon is that we will see tsunami-like improvements in how we use all of this data at our fingertips. Instead of code and queries, we'll have systems that help us answer questions like:

• What's the break-even point on our "buy n, get 1 free" promotion?
• How many of our customers come just for the hot dogs?
• How much are we spending on our customer loyalty program?
• How many defects do we think are in the system?
• Can I sell casualty insurance on a 4-bay auto body shop if the nearest fire department is 7 miles away?

The current bleeding edge of machine learning systems is able to answer these questions convincingly, but often in fascinatingly incorrect ways. Despite the clamor, we're just not there yet.

Why Now Feels Different

And that, I think, might be why we're seeing uncertainty. When we told people the web was the future, we were ready to build the web. When we evangelized SOA, we knew we could build those services. When we realized we needed mobile apps, we were using existing widgets and APIs. When we proclaimed the future was in enterprise data lakehouses, we had Snowflake standing at the ready.

But even if you believe me when I say that the next big wave involves how we use the data we've spent the last four evolutions carefully collecting and exposing, the fact is the technology isn't standing by ready to be used. The hype curve is way ahead of the technology curve, and it's leading to FOMO.

Addendum: How DevObsessed Uses AI Today

I'm going to abuse the term "AI" here, but here's some things we are using next-generation data technologies for.

• Copilot/CodeWhisperer/JetBrains AI - especially when coding in unfamiliar languages or building prototypes
• Our internal technology platform - we use a homegrown app and BERT model from huggingface with a vector database to analyze leads and match them with consultants on our roster
• ChatGPT - to answer questions about domains we are unfamiliar with and to make fun of Darrel.