The Games Still Work. The System Can't Read Them.

I recently dug out my original Nintendo Switch for my five-year-old son.

It had been sitting untouched for a while, but it still worked. We charged it, found some games, and brought it back into service. For a couple of weeks, everything was fine.

Then it stopped reading game cartridges.

At first, the failure was inconsistent. A game might appear after removing and reinserting it. Restarting the console sometimes seemed to help. Cleaning the cartridge felt productive, even when it changed nothing.

Then the failures became more frequent.

The cartridges were still there. They were clean, intact, and full of perfectly good games. But the Switch could no longer reliably recognize them.

From my son's perspective, the games had stopped working.

From the Switch's perspective, they no longer existed.

Blaming the Cartridge

My first instinct was to question the cartridge.

Was it dirty? Damaged? Inserted incorrectly? Did I need to clean the contacts again? Was my son somehow putting it in wrong?

The cartridge was the visible object, so it became the obvious suspect.

But multiple cartridges began producing the same result. Games that had worked days earlier were suddenly unreadable. The common element was not the games.

It was the reader.

Once that became clear, repeatedly cleaning the cartridges stopped looking like troubleshooting. I was polishing the inputs to compensate for a failing system.

The games had not changed.

The interface between the games and the system had failed.

I could not stop thinking about that.

The Other Broken Reader

A qualified person submits a job application containing years of experience, carefully selected accomplishments, a tailored résumé, and perhaps a cover letter.

The system accepts the submission. A confirmation page appears. An automated email arrives. Everything looks as though it worked.

Then the application disappears.

When this happens repeatedly, candidates are encouraged to question the cartridge.

Rewrite the résumé. Change the keywords. Adjust the formatting. Tailor it more aggressively. Rewrite the cover letter. Revise LinkedIn. Add measurable outcomes. Remove older experience. Make it shorter. Make it more detailed. Make it easier for the system to parse.

Sometimes that work matters. Presentation matters. Positioning matters. Clarity matters.

But eventually you have to ask whether the application is broken—or whether the reader is.

Building a Reverse ATS

That question is why I built JobArchivist—what I think of as my reverse ATS.

An applicant-tracking system gives an employer a structured view of every candidate: where they came from, how they moved through the process, who contacted them, what happened in each interview, and why the process ended.

The applicant receives almost none of that structure.

We are expected to navigate dozens of employers, systems, conversations, and decisions using an inbox, a calendar, a few browser tabs, and our memory.

JobArchivist reverses that relationship.

It builds the candidate's view of the market by connecting applications, job descriptions, contacts, interviews, emails, follow-ups, notes, and outcomes into one history.

It does not merely record where I applied. It helps me observe how the hiring system behaves from the other side.

There is some personal symmetry in this.

My earliest support experience came from Bullhorn, whose core product was an applicant-tracking system. I began my career helping support the employer's side of the hiring pipeline. Years later, I found myself building a reverse ATS to give the candidate some of that same structure and visibility.

In a sense, I have now spent time on both sides of the reader.

A job search produces an enormous amount of fragmented information: automated acknowledgements, recruiter messages, scheduling links, interview notes, thank-you emails, follow-ups, rejections, and long stretches of silence.

Looking at each event independently makes the process feel personal and chaotic.

Looking at the entire system reveals patterns.

What the Reader Is Doing

My current dataset contains 21 applications.

The measured response rate is 71 percent, although that includes both captured human replies and documented process progression. It does not mean 71 percent of applications produced a thoughtful response.

The median process contains three documented steps. For applications that reached a conclusion, the median time from application to outcome is 12 days.

Four applications were rejected before there was evidence of a screening call or interview. Nearly one in five never made it beyond the application layer.

Communication tells an even more interesting story.

JobArchivist has captured 33 outbound messages associated with these applications. Seventeen contain explicit follow-up language.

Across the 21 applications:

• Seven received a reply without requiring a documented follow-up.
• Four received a reply only after I followed up.
• Two have a follow-up with no subsequent reply.
• Four progressed, although no corresponding human reply was captured.
• Four produced no measured response at all.

This is not an academic labor-market study. It is a personal operational dataset.

But it is large enough to expose the shape of the experience.

Only one-third of the applications received a captured reply without follow-up. In four cases, the system required another outbound message before producing a response. Six applications currently show either an unanswered follow-up or no measured response.

The candidate is not merely applying for jobs.

The candidate is also operating the communications layer around the hiring system.

The Hidden Second Job

The conventional description of job searching leaves out most of the work.

You are told to find a posting, submit an application, and wait for a response.

The real process looks more like this:

Find the opening. Research the company. Interpret the role. Tailor the résumé. Write the cover letter. Submit through an applicant-tracking system. Preserve the job description before it disappears. Record the application. Identify possible contacts. Watch the inbox. Separate automated receipts from meaningful communication. Prepare for interviews. Send thank-you notes. Follow up. Follow up again. Record the outcome. Try to determine what the outcome means.

Then repeat the entire process while maintaining enough emotional distance to learn from it.

The administrative burden has been pushed onto the applicant, but the applicant has almost none of the system's visibility.

Companies can see their pipeline.

Candidates usually cannot see theirs.

That asymmetry matters.

A Good Game Can Still Go Unread

A failed cartridge read says very little about the quality of the game stored on it.

My five-year-old does not know about contact pins, card readers, hardware failures, or diagnostic procedures. He only knows that a game he enjoyed worked yesterday and does not work today.

The failure is real, but his explanation for it is necessarily incomplete.

Job seekers are put in a similar position.

An application-stage rejection may reflect a genuine mismatch. It may also reflect timing, application volume, location, compensation, an internal candidate, a recruiter's interpretation, an automated filter, a hiring manager's preference, or a role that changed after publication.

The candidate normally receives none of that context.

The system emits a generic rejection—or emits nothing—and the human being is left to invent an explanation.

That is dangerous because people tend to convert missing information into self-judgment.

A system failure becomes a personal failure.

The most useful change I made was to stop treating every rejection as the same event. A company is not simply "rejected." Individual roles have histories.

One application may have ended at submission. Another may have included a recruiter screen, multiple interviews, several follow-ups, and substantive feedback.

Those are completely different processes, even if they end with the same status label.

A red badge reading "Rejected" destroys information.

A timeline preserves it.

Better Questions

Without evidence, the natural question is:

"What is wrong with me?"

With evidence, better questions become possible:

• Which kinds of roles produce human responses?
• Which companies require repeated follow-up?
• Where does the process usually stop?
• How much effort does each opportunity consume?
• Which sources produce interviews rather than acknowledgements?
• How long do different companies take to reach decisions?
• Do personalized messages change outcomes?
• Are certain roles consistently rejected before screening?
• Which parts of my background are repeatedly validated during interviews?
• Where am I investing effort without receiving information in return?

These questions do not remove disappointment. They turn disappointment into something observable.

The goal is to take in what happened, separate signal from noise, act on what the evidence suggests, and learn from the result.

It is not about automating the human parts of a job search. It is about preventing a broken system from consuming human attention without leaving behind useful intelligence.

Inspect the Reader

When a Switch stops reading cartridges, repeatedly polishing the cartridges is not a complete repair strategy.

At some point, you inspect the reader.

The hiring market needs the same change in orientation.

Employers ask candidates for precision, preparation, responsiveness, customization, and enthusiasm. In return, many hiring systems provide generic acknowledgements, opaque decisions, inconsistent timelines, and silence.

That is not merely discourteous.

It is a systems-design failure.

A healthier market would provide clearer status changes, honest timelines, role-specific rejection context, and closure when a process ends. It would distinguish between an automated receipt and a human response. It would treat candidate time as a real cost.

Until that market exists, candidates need their own instrumentation.

That is what JobArchivist is becoming for me: not another place to store applications, but an intelligence system for understanding the labor market from the applicant's side.

The end goal is not a prettier dashboard.

It is a better next decision.

Which role should I pursue? Which company deserves another hour? When should I follow up? When should I stop? What evidence should change how I present myself? Which apparent failure was actually a strong signal? What pattern is emerging beneath the individual outcomes?

My old Nintendo Switch worked when we brought it back into the world. My son played it for weeks. Then the reader failed, and the same cartridges suddenly became invisible.

Nothing about the games had changed.

The job market often does the same thing to people. It accepts their information through a system that appears to work, then behaves as though nothing meaningful was received.

But people are not cartridges.

Applications are not identities.

Silence is not a meaningful evaluation.

Sometimes the game is good.

Sometimes the reader is broken.