Computer Programmers
Think about your job as a bundle of tasks.
Think about your job as a bundle of tasks.
You chose Computer Programmers. O*NET, the United States government's occupational database, describes this occupation through 17 measured tasks. Each circle represents one task. Hover, tap, or focus a circle to read the task.
AI tools are emerging that can execute parts of this bundle.
We compared the tasks that make up this occupation with the Agentic Task Ecosystem (ATE), a dataset of Model Context Protocol (MCP) tools. MCP tools let AI systems use external software and data. A filled blue circle marks a task whose action closely matched a public tool description. We found direct matches for 7 of these tasks. These listings show what developers have built, not whether workplaces have adopted it.
The number of matches tells only part of the story.
Tasks ask different things of workers. Some rely on general knowledge shared across many kinds of work; others depend on specialized knowledge or judgment concentrated in a particular field. We call this difference task expertise. The circles run from more general tasks to more specialized tasks, locating every task and every filled blue match on this spectrum.
The midpoint gives us a baseline for comparing where tool matches fall.
Each task’s language-based expertise score reflects how specialized the language in its description is. Within each occupation, tasks are ordered by that score and assigned evenly spaced percentile ranks. The lowest-ranked task sits at the more general end, while the highest-ranked task is at the 100th percentile. The 50th percentile gives every occupation the same midpoint for comparing where tool matches fall.
Current tool matches lean toward more general tasks.
The 7 matched tasks have an average expertise percentile of 47.1, which is 2.9 percentile points below the occupation midpoint of 50. The modest gap shows that the matched set remains fairly close to the occupation’s center overall.
That pattern hints at what automation could leave behind.
If AI performed the matched tasks, the job’s remaining task mix would shift toward more specialized work. In Expertise, MIT economists David Autor and Neil Thompson find that when automation leaves more specialized work, wages tend to rise while the pool of qualified workers narrows.
Now imagine a future for your job.
Select circles to build your own automation scenario. The same number of automated tasks can produce different outcomes depending on where they fall along the expertise spectrum. Blue fill shows your scenario, while a green ring marks a choice that differs from today’s published matches.
Work is already changing
O*NET offers a shared baseline, but your work may already include responsibilities its current description misses.
Technology does more than remove tasks. Workers can reorganize roles around new responsibilities, judgment, and coordination.
We are continuing to track both sides of that change: which existing tasks shift and which new tasks emerge.
