The Solution to Automation

The job market sucks ass right now. I'm not going to bother with the links and exposition to support my argument, because we all feel it in our bones. Every layoff, a possible death sentence; every AI project, a reminder of your dwindling utility; every new contractor, your potential replacement. Such is the popular narrative, consensus, gestalt of civilization at present.

Which is why I had a radical idea today that made me damn-near spit water all over my displays: automation has become a problem, and the solution is the shitty career advice we got from our Boomer parents that failed to support us in the modern world.

Applying in-person.

Hear me out! I promise this isn't a waste of your valuable time.

A Brief History of Hiring

In the before-times, job applications were done on physical pieces of paper - sometimes as a proper form to complete, other times as a resume to be submitted, and others still as a readily-available CV. You would - and I know this is going to blow the minds of younger readers - physically hand these documents to a living person, who - again, bear with me here - would have to read the documents to judge whether or not the applicant was a good fit for the role, and whether or not an interview should be scheduled.

This could be stressful, especially with large stacks of physical applications. Hiring managers would also have bias, either towards their own social networks or their own social classes, which hindered the ability to find qualified candidates who weren't already "in those circles". Hiring lacked objectivity, KPIs, and measures, yet this was enough for the biggest conglomerates of our history to thrive for decades, even centuries.

"There has to be a way to lower costs," thought executives, and thus the ATS was born: a computer system for scanning resumes for keywords chosen by the employer and surfacing "suitable" applicants for the job based on what the employer specified. Thus the arms race of modern hiring began in earnest, and everything went to shit. ATS systems meant reams of specialized recruiters and hiring managers could be shitcanned, since a system operating on buzzwords was all that was needed to weed out the bad seeds, thus saving money...until applicants understood how these systems worked (typically via lectures, seminars, and recruiting firms made up of recently laid-off people managers whose jobs were taken by these computers), at which point they were fed garbage resumes filled with suitable buzzwords to game the system.

This was a problem, and so ATS systems worked on improving algorithms. Claims of improved accuracy were met by criticisms about bias and discrimination, which merited newer versions that employers must pay for to get the fixes and not get in legal trouble. Except applicants once again learned to game the systems in question, using specific fonts, colors, and formatting to get higher scores and thus a stronger likelihood of being hired. Companies responded in turn with more automation - OCRs for paper, new algorithms, centralized applicant databases, personality tests, you name it - which applicants met with automation of their own.

Today, AI has made the arms race a lose-lose battle for both sides: employers are trying to use AI to scan the thousands of applications they get, and applicants are using AI to juice their applications and send out thousands of them a day with bots. AI then rejects applicants at 2am, three minutes after receiving the application, likely because it lacked enough buzzwords or vocabulary to reach an arbitrary scoring threshold for review - but also potentially because of the applicant being too old, or having a non-white name, or some other protected class that's legally defensible when the algorithm is a black box with no documentation.

This is a hot mess that's completely untenable any further.

The Answer is Humans

Humans are messy creatures, filled with inherent biases and often acting in our own self-interest at the expense of others. No system is perfect enough to eliminate these problems, or the many others humans innately have.

For hiring, though? The answer is absolutely getting more humans involved somehow, and axing automation where possible. Accepting resumes in a physical office is a good start, with mandates for specific formatting acting as both a barrier to entry (gotta demonstrate that word processor knowledge) and an obstacle to AI (until someone invents apply-by-drone). It gives applicants a chance to show their best selves, and hiring managers an opportunity to knock out a quick first-round interview if time allows, culling the stacks of applications further, faster.

What about remote work? For countries with a functioning postal service, this is an opportune time to accept mailed-in applications. Encourage candidates to decorate their envelopes to stand out in lieu of a cover letter, for instance, and set a cutoff date (e.g., two weeks from posting) to reduce the volume of applications to process. If your company is worried about someone mailing dangerous devices or items, well, a postal inspection service or workflow is probably a heck of a lot cheaper than the latest AI tooling, with the added benefit of providing actual utility compared to yet another automation escalation.

Human interaction isn't just a potential solution to the current hiring crisis, it's also a valuable data point for companies with storefronts or operating in the B2C space. If you can build an enjoyable human interaction for hiring, you can do so for your products as well. As more and more companies shift to data harvesting and AI bots to eliminate roles, there's an opportunity for companies who focus on employing, training, and retaining more humans to be viewed as desirable, even luxurious brands by comparison. Ask yourself: now that everything is expensive, everywhere, would you be willing to pay an extra 1-5% for human interaction throughout the process? Would you pay $50/month for an ISP that directs you to an automated chatbot for support before foisting you to a call center abroad, or $55/month for one that directly connects you to a local human who can fix your issue or might be aware of a local event/outage? Would you rather pay $300 for a full cart of groceries via self-checkout at a tiny stand, or $320 for a manned checkout counter and trained bagger?

Sure, a lot of people will go with the automated solution if all things were equal, but that's increasingly not the case. We don't save time or money doing self-checkout anymore, not with so many cameras and algorithms making sure we're not Danny Ocean performing a heist of groceries. Likewise, employers aren't actually saving time and money with automated applicant and talent systems anymore, but instead exposing themselves to risk, friction, and stress by relying on computer crutches for what's innately a human process.

With everyone falling over themselves to proclaim AI is going to take all the jobs, I'm increasingly bemused that our parents' advice to "pound the pavement, shake hands, make an impression, and hand in your resume in-person" might be the solution to the harms of automation itself.

I'm curious if this hunch of mine will prove correct, in time.