I create some repos to jot down thoughts that might grow into something else down the line or just share ideas with others with more bandwidth to execute. This is one of those repos (is this what Github is for? Not realy! But I like to do things a bit differently!)

FYI 2: I have become a fan of typing in plain text sans spellcheck because ... I guess it's what those "focus writers" try to do and I have never been sufficiently interested in any of them to stick with them over just typing markdown in VS Code (how I'm writing this). If I forget to do that or it appears as if a sentence was written because I spilled coooooffffeeeee on my keyboard ... please forgive me.

While I love AI, I use it very sparingly for editing my writing - mostly with very light prompts like "add subheadings," or "I dictated this, add punctuation." Mostly, it's me.

Why (I Aruge) The Job Market Sucks

Just a morsel of context:

I've been exploring all manner of permutations around traditional employment since .. about 10 years ago. I've been in-house, fully remote, contracting. I think that there are merits to each of these relationships and that what works at one stage of life mightn't in another. People also change.

At various junctures, I've dove into the conventional "job market" and gone through the usual processes (repettive interviewing, offers, jobs, moving on - the process).

Naturally - as careers evolve - exposure to the job market becomes more interspersed by longer term stability. So I've seen snippets of it periodically and through chatting with friends. I've hoped that - much as our conception of remote work has changed - the job market will evolve into something that's more empowering and pleasant to navigate. Sadly that hasn't happened.

Another thing that happens as one goes through their career: depending on the situation, you'll find yourself sitting at the other side of the table too - that of the hirer. Unless you work in HR you're probably not the actual person handling the paperwork. But you get to experience the process of what it's like to identify a need for personell and then fill it.

Context part 2: I'm aware that the tech world constitutes only a small part of the economy and that there are many jobs that simply cannot be done from home. As I jot down thoughts about how AI could help us craft a better job marketplace, I'm thinking, primarily, about the types of jobs marketplaces I'm familiar with and likely to encounter during the rest of my career.

---

Jobs Marketplaces Are Based Arounds Needs

Jobs marketplaces, as we experience them in 2025, work kind of against the grain of everything we're told about good career planning. They're not typically places where we try to understand and clarify our objectives and potential. They're focused primarily around keyword searches and discovery.

Jobseekers, it is understood, come to them because they need a job - and they need a job, by extension, because they need money and the things that money buys, like healthcare and shelter.

Employers come for much the same reasons. On the hiring side, job marketplaces are also not built to support long term or even medium term thinking: how do we imagine our organisation growing and what kind of people would help it flourish? They're, in a sense, reactive: we need a bum on a seat by this date. Who's in the talent pool?

For both parties this is disempowering.

Remote Creates Is Own Problems

The broken model of a needs-based marketplace might just about hold itself together when the geographical context is constrained.

Businesses need people to do jobs - and it's taken as a given that they will come to the office every day. So they need to be within a certain geographical distance to make that happen.

Likewise, employees need to find an employer within reach of their home.

When the talent pool and hiring pool both become remote, however, the constraint disappears - and the needs-based marketplace system finds itself simply unequipped to facilitate the kind of matching it may have done before.

Marketplaces find themselves fundamentally unequipped to deal with the basic realities of a remote friendly working world:

When the candidate pool is infinitesimally large, hirers will find it impossible to sort through incoming applications efficiently, ultimately, often, giving up in frustration and returning to place-based hiring.

Employees face a similar struggle: when the only matching criteria for a potential job is keyword-based (skills, job titles), they may find that it is still not sufficiently narrow to conjure up a reasonably sized pool of potential workplaces. On both sides, the system is overwhelmed by scale.

Blunt AI Can Blunt The Job Market Too

When AI is applied as a blunt force instrument to the jobs market, it is likely to only exacerbate these challenges: candidates decide that they can deploy automation and autonomous execution to easily "scale up" and "automate" job apps. But if they're chasing the same number of positions as in the world before AI, all that's happened is that hirers now face an even more impossible job in sorting through a messier-than-ever slushpile.

The innovation here may be self-defeating: the more employees try to "game" the job market to create a more flexible hiring paradigm, the more that hirers may reactively flex back to the kind of rigid 9 to 5 models that many are tired of.

AI Model: Based Around Desires & Vision

I'm sharing these thoughts as a note also because I'm not the right person to execute upon this concept.

Someone immersed in the world of HR and hiring and passionate about AI would be a much better lead for this project than I ever could be.

But the vision I have for the implementation would be something like this:

To overcome the problems mentioned above, I envision a traditional two-sided marketplace with candidates and employers.

But rather than starting on either side with a search process, the system or marketplace would leverage AI in order to generate a specification of sorts.

Rather than thinking of the CV/resume as simply a file to be sent out in batch to applications, AI tools could look to this as only a jumpstarter to a much wider ranging pool of context data which they need to assemble about the candidate in order to understand them as both a human and as a professional with skills to offer.

A model that I've worked with in prototypes is that of agentic interviewing for context development in which AI agents "interview" the user in order to proactively generate context by asking probing questions about specific subjects.

This, I have found, is an efficient method of creating an accelerated context pool and could be perfect in this application.

Of course, users may also seek input from traditional human career counselors and headhunters. Their notes and observations could also be ingested as context data, as well as parametric information such as minimum expected salary, preferred working arrangements, etc.

Employers, for their part, could be onboarded with a similar process which aims to map out the type of world they envision for future teammates at their workplaces.

This would present an interesting means of gathering information about the prevailing culture in a company as well as other soft determinants of employee retention that, at the moment, are primarily extracted through mining data from previous hires and other methods.

The overarching idea in this reconceived system is of creating a narrower talent and opportunity pipeline - not through filtering or search, but by an ever-evolving and more rich definition of self needs and desires for both parties in the system.

AI is layered in at two important levels:

• "Interviewer": clarifying goals and objectives from both sides of the marketplace
• Matchmaker: identifying matches between the sides so create potential hiring opportunities with good fit potential