This article is crossposted from IEEE Spectrum’s careers newsletter. Sign up now to get insider tips, expert advice, and practical strategies, written in partnership with tech career development company Taro and delivered to your inbox for free!
I recently had a conversation with a Microsoft engineer who was both excited and concerned that she had tripled her output using an assortment of AI tools.
Wow, I thought, that’s amazing! Even a 50 percent productivity boost is amazing, but a 3x increase is mind-boggling. So, what was her cause for concern?
She felt anxious that she wasn’t actually learning, because she had delegated almost all the implementation work to AI tools like Cursor and GitHub Copilot. The result was a shallow understanding of her work, which led to both anxiety and guilt.
Yes, I agreed, this was a cause for concern. You may be able to achieve more in the short term, but overuse of AI will eventually lead to career stagnation. This led to a broader conversation: How can we ensure our career success as we enter an era where AI is increasingly capable?
An IEEE paper titled “The Daily Life of Software Developers“ found that “developers spend surprisingly little time coding, 9 percent to 61 percent of the workday depending on the study.” In my own experience as a senior engineer at companies like Pinterest and Meta, I probably spent about 40 percent of my time actually writing code. The remaining time was spent collecting information, reading documentation, helping coworkers, or debugging. As AI coding tools become more prominent, more of our time will shift away from simply writing code.
Therein lies the answer to adapting for the future: become amazing at “filling in the gaps” for the work that the AI can do. Utilize AI for rote coding tasks, but maintain your critical thinking skills for the significant part of your job that goes beyond writing the code. Here are examples of premium skills going forward:
- Debugging and reviewing AI-generated code
- Monitoring software and working with other teams (humans) to fix issues
- Decomposing a business objective into smaller milestones that we could feed into an AI
Returning to the Microsoft engineer, I told her that one indication of a healthy AI relationship is to have opinions about the output. You must be able to defend or critique the code generated by the AI tools. If not, you’re at risk of being replaced by AI instead of being amplified by AI. A productivity drop (from a 300 percent efficiency gain back to 50 percent!) is a fine tradeoff to make, as long as you build an ownership mentality.
The ultimate currency in the workplace, both now and in the future, is trust. Are you trusted to be accountable for your work, or are you simply parroting what the AI is spitting out? Whether you’re a new college grad or a seasoned engineer, you must evolve your role in the knowledge economy to become a productive collaborator with AI.
—Rahul.
International Students Are Turning Away From the U.S.
The United States has long been a magnet for international STEM students. But early signals are beginning to indicate that interest in study at U.S. colleges and universities may be flagging as a result of policy changes from the Trump administration. If these signals bear out in enrollment, the trend could have lasting consequences.
Skills useful to learn for robotics engineering
What specific skills do you need to work in robotics engineering? In this edition of the software engineering Substack “The Pragmatic Engineer,” the cofounder of an early-stage startup called Nyro Humanoids shares his perspective on the most useful skills for engineers joining the field.
Profile: The Engineer Who Secured Nigeria’s Democracy
Steve Adeshina helped transform Nigeria’s voting system by integrating technology to improve election integrity and reach remote polling units. But Adeshina’s career has also spanned the private sector and academia, demonstrating the power of keeping an open mind when facing career surprises.
From Your Site Articles
Related Articles Around the Web
