Advancing your Tech Career in the AI Era

In the past year, I've seen an increasing number of people on social media fretting about the future of tech jobs in an AI world and whether it's even worth it to go into the field. The sensationalist, click-bait pieces from media and influencers haven't helped matters, so I figure it's time to take a few minutes and help people take a deep breath, calm down, and have a rational discussion about what is happening today and my prediction for what is coming and how to adapt.

Factoring in the Current Job Market

To start, we have to look at the current job market. 2023 was a brutal year for technology professionals, though one could argue that the correction was due. The combination of over-hiring from the pandemic, salary bands that far outpaced inflation (in some regions), and the end of cheap debt financing all combined to form a significant correction. Until late 2022, demand was way outside the norms, reminding me of the time before the Dot Com Bubble (and subsequent crash). For my younger readers, if you were in tech in the late ’90s’, there was a short window where if a person could spell HTML, they were getting jobs at well above market rates even though their skills were demonstrably shallow. I see a lot of parallels to that time and the “learn React in 90 days and get a 6 figure job” bubble that we’re coming out of.

Taking a step back, though, we need to consider a few things about the job market going forward:

1. Even with 2023, overall employment numbers are still net positive, so the field has continued to grow.
2. The unemployment rate for software developers is still far below that of other fields, including healthcare, where the same media frequently chronicles worker shortages.

Yes, the market is still challenging, and many individuals have been and will continue to be impacted, but signs of recovery are starting. I’m sticking with my Q2/Q3 prediction of stabilization.

Is AI Taking Developer Jobs?

I’ve been using AI tools in coding tasks for nearly a year, and I can assure you that it is nowhere near taking developer jobs. Again, we can look back at history. SQL and COBOL were supposed to be so easy that business people could write their code, Visual Basic and Windows Forms were supposed to make drag-and-drop applications so easy that you wouldn’t need huge teams of programmers to build applications, and recently, low-code and no-code solutions make claims of substantially reducing the need for developer talent.

I am orders of magnitude more productive than I used to be due to the advances in tools, and yet, despite these advances, the world keeps demanding more software. We have no evidence that if AI tools make developers 30% more productive, it will hurt the field’s growth or overall employment. Some studies are already seeing a drop in the quality of code authored as AI tools allow low-skilled developers to copy-paste-flail their code into repositories, and then someone has to go in and clean up the mess.

But What if AI Gets Good?

As a thought exercise, let's assume that AI tools continue to improve, bringing them closer to mid and senior-level developer output. What does that mean for the field? Does it mean the "death of the developer career"? Here's why it doesn't, and to help, I will use a normal curve. Let's assume this curve represents developer skills, with the left being lower-valued "junior level" skills and the right being higher-valued "solutions architect" skills.

Here are the traits I would use to separate professional developers into parts of the curve:

Left (lower-valued)

Everyone starts their career on the left side of the curve, and we typically see:

• They perform low-complexity tasks, including modifying UI components, adding additional columns to reports, and scripting small batch jobs, and basic troubleshooting.
• Their tasks, such as building UI components based on database queries or API results, basic CRUD (create-retrieve-update-delete), and other boilerplate/repetitive code types, are easily templated.
• Their primary workflow resembles “copy-paste-flail” because they lack solid foundations and struggle with creating and comprehending anything but low-complexity code.
• They have poor debugging and defensive coding skills.
• There is a lack of general knowledge about adjacent technology topics such as networking, deployment, security, etc.
• Those who are poor communicators and collaborators.
• Those who require detailed specifications and code reviews to be effective.

Middle (average-valued)

The traits of these developers are:

• Strong problem-solving skills. They can integrate new features into existing codebases cleanly, anticipate and address issues, and recognize opportunities to make processes more efficient and effective.
• They can work with people and processes to gather and translate business requirements into technical specifications (and AI prompts).
• The capability to design medium-complexity solutions with reasonable autonomy.
• The technical skills to fact-check AI-generated code for accuracy, security, and performance.
• An understanding of maintainable and professional grade code and the ability to use AI tools effectively to generate it.

Right (high-valued)

These developers will likely always have employment opportunities because:

• They can architect complex systems, considering security, performance, reliability, and accuracy concerns.
• They’re capable of complex debugging and analysis.
• They have deep expertise in areas where AI models don’t have enough data or fall short.
• They can engage in new research and invent new processes or technology.

Today, all of the developers in this normal curve can find and maintain employment, but developers on the left side of the curve should be concerned. If AI code generation capabilities continue to improve, this group will be the most redundant. Unfortunately, many of these traits overlap with entry-level developers, especially those from front-end boot camps and low-quality degree programs.

However, everyone has to start somewhere, and the developers in the middle and right part of the curve will not be keen to spend their days using AI to do these lower-valued, repetitive tasks, so there will still be a place for these developers as the technology improves. However, I anticipate we will need fewer of these developers, and if you are a developer in that part of the curve, your primary objective should be moving to the right. If you are unwilling or unable to do so, your earning power and job security will not be great.

Non-technical stakeholders and users cannot today and likely will not be able to ask the right questions, generate the right prompts, or validate the approach of AI tools in the future. So, people in the middle and right side of the curve will likely be fine. I predict that people considering tech careers for the money without genuine passion or interest won’t have the drive or discipline to move to the right on the curve and, over time, will be pushed out of the field.

But What About AGI?

This isn’t something I worry about at all. Suppose the technology gets to the point where it can replace the left and middle parts of the curve. That means the tools have also gotten so good that most white-collar jobs like accountants, marketers, financial professionals, middle managers, call center employees, and paralegals are also obsolete, to name a few.

The social contract must change if this happens, and all the bets are off. We could evolve into a society with Universal Basic Income, reduced work hours, and pursuit of meaning. We could have a modern French Revolution, smash the machines, and become Luddites. Or we might enter a dystopian future where the gap between the haves and have-nots is similar to the movie Elysium.

No one can predict that, but it’s not a reason to give up on a tech career today.

Moving to the Right Side of the Curve

“Ok, I buy this. How do I move to the right side of the curve?”. Here are some suggestions:

1. Deepen Technical Expertise: Make the effort to understand how processes and systems work at least one level down from where you are working. Consider specializing in security, performance, architecture, or other skills in high demand.
2. Enhance your Soft Skills: Improve your written and spoken word abilities, especially becoming proficient at translating technical concepts to non-technical stakeholders and working effectively with teams and leadership if you are advancing in that direction.
3. Continuous Learning: This was true before and is even more important in an AI future. You must have a plan for continuous improvement.
4. Gain Domain Knowledge: Understanding technology is far more powerful when you understand how it can be applied to specific domains (finance, healthcare, etc.).

In Conclusion

The best strategy for preparing and thriving in a potential AI future is to build a solid foundation of technical skills that allow you to work effectively with and verify the output of these tools. Just "making it work" will not produce high-quality, professional code, and, in the long run, is a losing strategy. The ability to solve problems and design solutions appropriate to budget, time, performance, and accuracy constraints is not something AI, or a non-technical human using AI, is likely to be able to do anytime soon.

So calm down and focus on a career path that moves you to the right side of the curve.