Juno Maps launches VIBE After School, a free and open framework pairing K-12 students with working professionals and AI tools to build real things. A guide to the shifting job market and what communities can do about it now.
Dustin Rauch, CEO, Juno Maps10 minute read
Let me open with something I did not expect to be writing. I run a technology company. I am also a father. And I am starting to question whether a traditional four-year college is the right path for my own children. If that sentence lands uncomfortably, I understand. It landed the same way on me. It is also, I think, an honest read of where the economy is going. And it pushed me toward a harder question. If the old map no longer fits, what should we be doing instead, starting now.
If you have a child in school right now, or you are watching someone close to you choose a major, start an apprenticeship, or come back to work, the map they are navigating is being redrawn in real time. Week to week the change is easy to miss. In the data, it is not.
The short version. The tools now widely called AI are changing what work looks like faster than schools, employers, and families are adjusting. Some careers are being absorbed into the tools. More are being amplified by them. The careers worth aiming for are the ones where human judgment, care, and craft become sharper when paired with AI, rather than replaced by it.
If that sounds abstract, a chart published by Anthropic a few weeks ago makes it concrete. Two shapes, laid on top of each other, describe where the economy sits today.
What the shapes show.
Theoretical capability
The share of job tasks in each category that current models could perform.
Observed usage
The share of work actually being done with these tools today, based on real traffic on one of the most widely used AI systems.
Figure reproduced from Anthropic’s published research. Used with attribution.
Figure 1. Theoretical capability and observed usage by occupational category. Source: Anthropic, “Labor market impacts” research, 2025. anthropic.com/research/labor-market-impacts
The blue shape reaches into almost every profession. The red is a small, uneven bloom concentrated in a handful of fields. The distance between them is the story.
The gap tells two stories at once. The ceiling of what AI can do is much higher than anyone is using. And the floor of what it is already doing is moving. The useful question is not whether the shift is coming, but where it lands first, and what to do in the meantime.
People who tend to do well through shifts like this share one habit. They choose tools before tools are chosen for them.
One
What the tools do, in plain terms
Think of AI less as a distant science project and more as a new kind of colleague. It has read more than any person alive. It never gets tired. It is confidently wrong often enough that it still needs a human to decide when its answers are good enough to use. Most of what is changing in offices today comes from three capabilities.
It writes.
Emails, memos, reports, code, legal first drafts, marketing copy. Work that used to take a morning can take seconds.
It creates.
Images, video, slide decks, songs, working applications. Often at the level a small business can actually use.
It reasons.
Reads contracts, audits spreadsheets, drafts diagnoses, plans projects. Anywhere a task is a puzzle with a pattern.
The pace is the story.
Figure 2. A rough trajectory of useful AI capability over a decade. The 2030 point is an estimate, included for shape, not precision.
Five years ago, a model could barely complete a sentence. Today it drafts production code. Whether the next five years repeat the curve is genuinely uncertain.
What is less disputed: the tools are already better than most organizations have absorbed. Even if development paused tomorrow, the absorption lag alone implies several more years of change.
Two
The thing we can start doing now
If this is where the world is going, the question is not whether schools will catch up. It is whether we can give kids a better starting point this year. There is a simple, practical way to do that, and most communities already have everything they need to start.
The biggest leverage sits earlier than most people think. Not in universities. In elementary and middle schools, where children are still open to everything and these tools are at their most inspiring. Most districts already run mentor programs, STEM clubs, and robotics teams. A new framework fits right beside them.
A proposal for your community
VIBE After School
An after-school mentor-and-make program that pairs students with working professionals in their own community and gives them the AI tools, time, and encouragement to build real things. The goal is exposure, not another class. A fifteen-year-old who has spent one afternoon with a working engineer, nurse, designer, or tradesperson, and who has vibe-coded her first app, enters high school with a map. Most of her classmates do not.
V
Venture
Real exposure to people doing the work. A mentor in the room, a site visit, a conversation with someone five years ahead.
I
Imagine
Protected time and space for curiosity. The tools reward the student who asks a strange question more than the one who follows a script.
B
Build
Every student ships something small each quarter, with AI as a collaborator. Working apps, circuits, drafts.
E
Express
The work should reflect who they are. These tools are the first medium that lets a ten-year-old release what she imagines.
The ingredients already exist in almost every community. A PTA that invites. A local employer that says yes. A teacher willing to give an hour a week. Reach out at vibe@junomaps.com or vibeafterschool.com.
Three columns of careers: work compounding with AI, work anchored in human presence and physical skill, and work being absorbed into the tools. Compensation ranges reflect current market signals and are not promises.
Column One
Compounding with AI
The tools multiply what a skilled person can do. Judgment is the product. The tools make it sharper.
Column Two
Anchored in humans
Work that depends on presence, trust, or physical skill. AI helps at the edges. The core stays with the person.
Column Three
Being absorbed
Not disappearing, but compressing. Ten roles may become three. The skills transfer. The titles shift.
The left column uses AI heavily, which is the opposite of the instinct to hide from the tools. The middle column is the one most often overlooked: electricians, nurse practitioners, HVAC technicians, welders. Demand here is rising for reasons that have little to do with AI. Demographics, electrification, reshoring, and overdue infrastructure do the work.
The right column is not a list of failures. It contains careers that were safe and reasonable as recently as 2021, held by people who trained seriously and are still needed. The shape of the work is changing, and the skills transfer. This is a transition, not a verdict.
Every absorbed job has a better cousin.
For anyone currently in the third column, or with a child training for one, the most useful pattern in the data is the adjacent role. The skills almost always transfer. The transition is easier the earlier it is made.
Four
What families can do right now
Schools will take time to adapt. Families do not have to wait. A small set of moves holds up whether AI advances faster than expected or slower.
Put the tools in the hands of the kids.
Sit down with them, open a model, and use it for something real. Fluency comes from daily use, not from reading about it.
Encourage vibe coding.
Building software by describing what you want in plain language and letting the AI write the code. A ten-year-old can ship a working website with a chat window and a free afternoon. Give them the room to try it.
Do not over-index on prestige.
The right path fits the actual strengths of the person walking it. The best electrician or HVAC technician in 2035 may well earn more than many lawyers do today.
Take the trades seriously.
Compensation data for skilled trades has been quietly strong for a decade. Electrification, reshoring, and infrastructure are likely to extend the trend.
Focus on skills over credentials.
Employers increasingly hire on demonstrated work rather than degrees alone. Encourage the portfolio early.
Help them start things.
Running a club, building a website, launching a small service. The skill of beginning, recruiting help, and finishing is the rarest capability in the economy.
Arrange a day of shadowing.
Most practitioners say yes when asked respectfully. One afternoon in a real lab, shop, or clinic clarifies a direction that months of research cannot.
Five
A short note for leaders
Policy matters, and four moves stand out for anyone with institutional leverage. None require waiting on Washington.
Make benefits portable. When health insurance follows a worker instead of a job, people retrain faster and the labor market adjusts.
Create lifetime learning accounts. A small payroll contribution and a public match, held in an account that follows the person and pays for retraining.
Build. Housing, transmission, nuclear, rail, water, heat pumps. These projects employ the same trades the AI economy needs.
Modernize the public stack. Permits, licenses, benefits, 311. The state that fixes this first earns real civic and economic dividends.
The only question left.
This is not the end of work. It is the end of a theory of work, the one where a person chose a field at nineteen and expected the choice to hold for forty years. That theory has been fading for a while. AI is accelerating it.
What is emerging is, in some ways, older. A life of small apprenticeships. A portfolio that grows slowly. A reputation that compounds. Tools that sharpen every year. That life used to be reserved for artisans and physicians. It is now available to many more people, with better instruments than have ever existed.
The opportunity in front of us is not to predict the future perfectly. It is to give kids a head start in a world that is already here. The tools exist. The mentors exist. The schools and families willing to try exist. The only question left is whether we put them in the same room.
A note on uncertainty. Everything above is an interpretation of current signals, not a forecast. Compensation ranges reflect present-day market data and are likely to move. The argument is not that any one career is guaranteed, but that the shape of the opportunity is legible enough to make thoughtful choices, and that flexibility will outperform prediction.
Juno Maps builds software for the people who actually run things, part of the RAUCH family of companies. Both are built around the same conviction that animates this essay: the most useful thing AI can do right now is give the people doing the work the judgment and leverage that used to require a much larger team.
This essay reflects the personal views of the author.
Distribution note. Alternate headline: How AI Is Changing the Jobs Your Kids Are Preparing For. Sources and ranges blend Bureau of Labor Statistics projections, the Anthropic Economic Index, and the author’s reading of current signals. Figure 1 from Anthropic research at anthropic.com/research/labor-market-impacts.
