The current market is pretty bad. My position in it might be worse.

It's painful to say, but my faith in the software industry is eroding. After playing the job search lottery for work these past ten years, no previous job search has ever been this dry. I've come to the conclusion I only have myself to blame. Early into my career in web, I looked to diversify out of the desire for additional job security, and I've achieved the opposite. Looking back on it now, maybe it was a mistake.

I don't really know any other way of saying this without being interpreted as arrogant, so try to stick with me here. Someone I've worked with for many years since the very beginning of my commercial experience once called me a unicorn, and I've slowly come to believe them. My background in software is extremely atypical compared to those I've worked with. In 2005 I wrote a cupholder script, and I've been writing software as a hobby ever since. It's been an outlet to create for two decades now, enabled me to put myself through college to obtain a master's in computer science, and allowed me to build a career to a senior level after ten years of doing freelance, research, and commercial positions.

My 2025 application to interview conversion rate is an incredible 0.1% after leveraging my network, manually filling out close to two thousand total cold job board applications, applying through company career sites, and interacting with company recruiters. I'm sure there's always something I could do better. None of this is to convince you I'm not flawed in my approach, but the real reality is how would I know when feedback does not exist? The current status quo is never getting followup, and if it does come, it comes from a no-reply email with a templated rejection.

Overspecialization has taken over

In the past several years, job postings have become more and more specialized. Listings detail deep expertise in a specific stack, tool, or niche, leaving generalists like myself overlooked despite their adaptability. Recruiters seem to conflate specialization with competence, and undervalue the sort of broad skill set and end-to-end vision generalists bring. Additionally, measuring competence between a generalist and specialist doesn't seem possible in the same interview settings.

Tech stack of the month

Industry specialization rewards those who picked the trendy stack and it's slowing shifts in tooling. AI frameworks, frontend, and niche dev ops tools reward candidates who've happened to put all their chips on red 23. This kills generalists and unlucky specialists alike, even if generalists show the most promise in being the most adaptable. Both face rejection for lack of recent experience even if both have used something tangential in the past and could learn the stack in weeks.

Hiring process punishes breadth

Automated resume screening has always favored keyword matches. If we've moved past frequency keyword matching, it's LLMs now. It's always been something. Candidates whom have diverse but non-linear career paths are some of the first two be filtered out. Limiting the candidate pool is no doubt difficult, but it's certainly never given the impression of surgical consideration. Spamming keywords in invisible embedded text used to be heard of, but now it's putting a prompt injection to convince the unfortunate LLM tasked with understanding your career condensed into two pages that you're the greatest talent bar none. If you're lucky enough to get to an interview face, expect a test of hyper specific trivia instead of broad system thinking or architecture design. Not knowing application specific jargon seems to immediately invalidate your abilities as an engineer, even if you demonstrate understanding but not the nomenclature.

AI can't eat generalists yet, but the industry thinks otherwise

Companies lean on LLMs and AI powered scaffolding approaches for doing "easy" generalist tasks, while justifying human hiring for specialized roles. If those in industry were truly hiring specialists for this reason, why do they use AI? Ironically, the flexibility that once gave generalists an edge is now seen as replaceable by automation, and companies are now falling into pitfalls like knowledge and accountability gaps because most of their workforce has homogenized. These forces create environments where knowledge is the only thing keeping a team in place.

Jack of all trades, master of some.

The serotype persists that generalists lack depth, even though many have deep knowledge across multiple domains. Building talent in many domains enables wide architecture thinking, but can be punished by the recruiting process when it is seen as job hopping rather than demonstrating adaptability. Generalists are also placed into environments where they independently lift more than a single contributor is thought capable and this can also be interpreted as overselling one's own contributors in lieu of a team's support.

Career progression bottlenecks

Individual contributor roles where generalists thrive are often reserved towards the top of the talent grading, but because specialist friendly organization recruit and promote often center around mastery of one domain, generalists are often stuck in mid-level positions. In addition to individual contributor positions, leadership roles sometimes demand both technical depth and people-management experience, but ironically, this is the kind of hybrid skill set generalists often have.

Octagon in the triangle hole

Constantly having to reframe experience to fit in a narrow language of job postings often means omitting large quantities of experience, making you evaluated as less experienced than others. The market seems to reward conformity over curiosity, and niche depth over holistic thinking.

Afterword

I'm bound to discuss this topic more in the future, but the above are some of the biggest pain points I thought to mention. Most of my articles usually feel like they miss some valid points when looking back on them in retrospect, but writing them enables me to capture a subset rather than missing all of it to time.

Going back to my point on erosion, it's gotten extremely hard to justify writing cover letters. I've yet to see or feel any tangible benefit from writing them, and with how accessible LLMs are to the average person, I'm sure many recruiters assume AI first. The current culture seems like anything short of a handwritten note is AI, but we're well past the point to emulate even that. It seems to be more of a perspective thing. For example, the research field has seen gen AI be used to fake presidential speeches since the obama administration, but the general public had a barrier to entry to use that software. Now a large number of LLMs are accessible by browser. While times have certainly changed, it's always been more about accessibility and the price of computation than anything else.