Current Infrastructure is Not Enough
We can't accept a partial response to a structural shift
The systems we rely on to connect people to work were built for a different economy.
That infrastsructure assumed that jobs are stable, roles are well-defined, and fit can be inferred from credentials, titles, and past experience. They assume that if we optimize matching speed and volume, good outcomes will follow. For a long time, that approximation was good enough. In the AI economy, it isn’t.
Work is changing faster, roles are more fluid, and the cost of poor matches is higher on both sides. Workers don’t want to churn through jobs that don’t fit. Employers don’t want to absorb the cost (financial, operational, and cultural) of constant turnover. Endurance matters now, not just placement.
Gray collar roles make this especially clear. These jobs often require certifications and relevant education, but they also depend on something harder to capture: judgment, interest, adaptability, and personal ethos. They succeed when the work aligns not just with what someone can do, but with how they prefer to work and what they are willing to be accountable for.
Today’s infrastructure struggles with this on both sides. Workers have little support in developing real insight into themselves and the kinds of careers that would suit them over time. Employers struggle to ask for, interpret, and trust anything beyond surface-level signals. Conveying this kind of insight, honestly and credibly, is difficult for everyone involved.
We will try to retrofit existing platforms and processes. Some improvements will help. But these are partial responses to a structural shift.
The demands of the AI economy require new beginnings, new ways of understanding people, new ways of describing work, and new connective tissue between the two. Without that, we will keep optimizing systems that no longer match the reality they are meant to serve.
Let’s treat this as a signal that something new needs to be built.