My 10 predictions for 2024

Normally, I try to be very careful with predictions. As Yogi Berra so eloquently said, prediction is hard, especially about the future. When looking at the folks predicting the stock market, the data shows that their accuracy is about as good as random, or even worse. For example, at the end of 2022, many predicted a major recession for 2023 whereas the result was a stock market that went up 20 percent or so. To be honest, I think this limited ability to predict applies to almost all areas of life, including digital technologies.

Despite all that, I thought I’d make a few predictions for 2024 and then evaluate at the end of year how well I did. Of course, going out in public with these kinds of predictions is a bit of a fool’s errand, but one has to take a gamble from time to time. So, here goes ...

1. GenAI code tools are going mainstream in enterprise

In the Software Center companies, I noticed that many were experimenting with GenAI code tools (GPT4, Copilot and the like) during the last year. My prediction for this year is that these tools will be broadly deployed in enterprises as their benefits have proven themselves and they provide a significant productivity improvement for many.

2. Resistance to Agile will reach a tipping point

Last summer, I wrote a post entitled “The end of Agile.” This was in response to the growing resistance against agile methods that I noticed in many corners of software development organizations. With the Agile Manifesto published in 2001 and hence more than 20 years old, it’s time for us to move on to the next paradigm. What this new paradigm could (or should) look like is something I hope to discuss in future posts as I have many ideas. But my prediction is that many will move on to the next paradigm and let go of Agile.

3. Federated learning will become an accepted technology

For years, I’ve been flabbergasted by the lack of interest in federated learning (FL) in the embedded-systems industry. For the millions of devices out in the field, everyone keeps falling back on the AI approaches developed by the SaaS companies, with centralized data storage and training. During 2023, however, I started to see the first signs of increasing awareness and interest in making use of the unique nature of embedded systems. My prediction (and hope) is that we see a broad adoption of FL this year.

4. DevOps in regulated industries will take off

Even if everyone assumes that DevOps is the normal way of operating, it typically isn’t the case in regulated industries. This is because the certification approaches have been unable or unwilling to adopt continuous certification as an approach. With several automotive companies as well as companies in the healthcare space preparing for DevOps, my prediction is that 2024 is the year when we reach the tipping point where several companies adopt DevOps for some or all of their products.

5. Data products will become established

Many of the companies I work with manage data products for internal use. A data product is a managed data set or managed stream of continuous data that has been processed to maximize its relevance for specific use cases. Up to now, companies have been very careful to exploit these data sets externally as the data typically originates from their primary customer base and nobody wants to upset these customers. My prediction is that in 2024, this will shift and we’re going to see commercially available data products to a significantly higher extent and it will become a new source of revenue.

6. Software architecture is going integrated again

For the last decade or so, microservice architectures have been in vogue. The promise was high degrees of decoupling of both components and R&D teams. During recent years, it became clear that the adage that one can do spaghetti programming in any programming language also holds for microservice architectures. Many companies experienced significant inefficiencies due to this architectural style. In 2024, my prediction is that we’re going to see a significant reversal of the trend and a much higher degree of (vertically) integrated architectures.

7. Teams are going truly cross-functional

With the adoption of agile practices, we saw agile teams starting to integrate architects, developers and testers. With the adoption of DevOps in many companies, my prediction is that we need to include more roles and functions in teams, specifically product management and customer support. Product management will work with the rest of the team to figure out what to build and how to experiment with functionality to make sure we only build what customers want. Customer support needs to be integrated as the flow of new functionality is so fast that support needs to sit close to the team to know what’s going on.

8. Continuous business models will be even more prevalent

Especially companies in the embedded-systems industry treated DevOps as a cost center but monetized their products in a transactional fashion. Even if there were several attempts to generate continuous revenue, it often represented only a small percentage of the overall revenue. My prediction is that in 2024, we’ll see a significant acceleration in the use of continuous business models as the primary source of revenue for companies.

9. Year of platforms

All of the companies I work with have used platforms for decades. Traditionally, however, these platforms were predominantly used to capture commodity functionality that every product in their portfolio needed and consequently they were R&D efficiency tools. With the increasing use of DevOps, however, each product in a portfolio needs to be derived from a common ‘superset’ platform as it’s prohibitively expensive to generate a new version of the software for every product every couple of weeks. In addition, this platform can then be used by external parties to build extensions that benefit all products in the portfolio. My prediction is that 2024 is the year that many companies realize this and start executing on their platform strategy.

10. Organizations will “lean out” management layers

The rapid feedback cycles enabled by DevOps and the quantitative data coming back from the field allow for a change in the management of companies. Instead of coordinating among teams using management layers, teams can be given quantitative KPI targets to pursue and then tracked in real-time using the data coming back from the field. This significantly reduces the need for traditional manager-driven coordination and my prediction is that organizations that realize this will start to “lean out” management layers and put more responsibility on teams.

In short, my predictions are concerned with the expanding role of AI, the increasing importance of DevOps, the disruption of Agile as the primary paradigm for development, the commercialization of data, the role of platforms and changes in how companies organize themselves. I’m certain that several or even most of my predictions won’t pan out in the way that I believe they will right now. But failing to plan is planning to fail. As Malcolm X said: “The future belongs to those who prepare for it today.”