A friend shared this Fast Company piece in our group Slack recently on the costs to companies of replacing entry-level workers with GenAI.  Issac, the mid-level Big Tech engineer in Megan Carnegie’s piece, is like a  lot of engineers working on my teams.  He’s seen firsthand what GenAI accelerates in his own work, as well as the downsides of its incorrect outputs and the additional time necessary to fix them.  At my level, I must balance the increased demands for software delivery, a larger people leadership workload, the risks of burnout for my most-experienced engineers, and the downstream consequences of that burnout.

 

Contrary to the assumption that GenAI could make up the difference for fewer junior engineers, much more work rolled up to senior engineers.  Instead of growing junior developers to the point they can do more work autonomously, they’re herding AI agents and spending an extra 4.5 hours a week just fixing their mistakes—in addition to doing their own work.  When you have a dozen or two dozen direct reports, it’s much more difficult to detect and prevent burnout.

 

When a burned-out senior engineer leaves, you don’t just lose their knowledge and expeirence.  The tacit knowledge built through what they worked on with those in their teams and across teams is lost as well.  GenAI appears to be compounding the negative effects we’ve seen of repeated mass layoffs by numerous companies—loss of institutional knowledge at scale.  The vicious cycle of burned-out senior engineers being replaced by less-experienced engineers and smaller workforces overall is already leading to less-resilient systems with more frequent outages taking longer and longer to resolve.

 

As a people leader, replacing entry-level workers with GenAI also eliminates opportunities for me to delegate, mentor, and teach.  The vicious cycle has the paradoxical effective of devaluing me as a person who has grown institutional knowledge over the years and consistently shared it with others to help them grow professionally.  At least one study has shown that this mentoring work increases motivation and psychological well-being.  I’ve worked with dozens of engineers over the past 8 1/2 years in peer learning, people leadership, and mentoring contexts and contributed to at least 10 people getting promoted at least one level in that time.  More than any individual system I’ve led delivery of, helping engineers grow their careers to succeed at even greater levels feels great.

 

These costs of trying to replace entry-level workers with AI are only going to grow as experienced workers who can afford to retire choose to do so, creating a demographic cliff Carnegie’s piece measures at 4.6M more workers retiring than there will be equivalently-qualified younger workers to replace them.  Companies that keep hollowing out their future talent pipelines are placing their continued viability at risk.  Those that use GenAI to augment  their existing workforces instead will have a more sustainable future.