Analyze user needs and software requirements to determine design feasibility.
Mehan Observatory / Applied frontier 01
How much of a job
fits inside today's AI horizon?
Not a headline about jobs disappearing. A task-level map of what machines can plausibly assist, what real workplaces make harder, and what remains stubbornly human.
Open the instrument →Task-level explorer
A job is not
one task.
Choose an occupation, then tighten the conditions. The instrument separates work already close to the AI frontier from work still anchored in trust, context, and the physical world.
Software Developers
- U.S. employment
- 1.7M
- Mean wage
- $148,100
Develop software testing or validation procedures, programming, or documentation.
Confer with technical teams to design systems and define constraints and interfaces.
Modify software to correct errors, adapt hardware, or improve performance.
Prepare reports or correspondence on project specifications, activity, or status.
This is an editorial scenario signal—not an automation probability, job-loss forecast, or claim that every workplace uses the same tools.
How to read the instrument
Start with tasks.
O*NET supplies the official task statements. Five representative tasks reveal why exposure inside one occupation can vary sharply.
Stress the setting.
Reliability raises the bar for repeatable success. Workplace friction represents missing context, coordination, approvals, and imperfect inputs.
Keep humans in view.
Physical action, legal responsibility, care, trust, and organizational judgment remain different constraints from producing a correct digital output.
Evidence desk
Sources before
certainty.
Continue the inquiry