The "Low Hire, Low Fire" Equilibrium: Why the Tech Job Market Feels Frozen

[vp_eng_keisha]

I’ve been in engineering leadership for 16 years and I’ve never seen a market like this one. It’s not a downturn. It’s not a boom. It’s something genuinely new — a frozen equilibrium where nothing moves, and that lack of movement is the most destabilizing thing about it.

The Numbers Behind the Freeze

The US economy added only 584,000 jobs in all of 2025. That’s a 71% decline from the 2 million jobs added in 2024. The voluntary quit rate is stuck at 1.8% — the lowest in over a decade, excluding the initial pandemic shock. People aren’t leaving. People aren’t being hired. The market is in stasis.

For tech specifically, the picture is even starker. Indeed’s data shows tech job postings sitting 36% below their February 2020 level as of mid-2025. Software engineering postings — the single most common tech role — are down 49% from pre-pandemic peaks. Specialized developer roles like Android, Java, .Net, and iOS are down over 60%.

And yet unemployment looks manageable on paper. The headline number sits at 4.6%, which economists call elevated but not alarming. This is precisely why I call it a “frozen” market rather than a “bad” one — the traditional indicators don’t capture what’s actually happening.

What “Invisible Unemployment” Looks Like From My Desk

I run engineering at an EdTech startup that’s scaled from 25 to 80+ engineers. I’m not cutting. But I’m not hiring either. My team leads have been asking for headcount for six months. My answer keeps being “not yet” — not because we don’t need people, but because our board’s message is clear: prove you can do more with what you have before we fund new hires.

This is the lived reality of the “low hire, low fire” equilibrium. The people on my team are employed, technically secure, and quietly exhausted. They’re carrying the work of headcount that was never backfilled after the 2023 cuts. They’re doing three roles because AI was supposed to fill the gap and it only filled about 30% of it. And they can’t leave because there’s nowhere to go — the market they’d be entering has job applications that have doubled since 2022 while openings have shrunk.

SaaStr called this the rise of “invisible unemployment” — the jobs that just never materialize. 66% of CEOs surveyed said they plan flat or reduced headcount in 2026. That’s not a layoff announcement. It’s not a headline. It’s a collective decision to hold still, and it’s suffocating the market.

Amazon Is the Microcosm

Amazon cut 16,000 corporate jobs in January 2026, following 14,000 in October 2025. That’s 30,000 people — roughly 10% of their corporate workforce — in four months. CEO Andy Jassy has been explicit about the strategy: “remove layers,” “increase ownership,” build a leaner model.

But here’s what makes this different from 2023-era layoffs: Amazon isn’t shrinking. They’re restructuring. Their planned capital expenditure for 2026 is billion — the highest among megacap companies. They’re cutting corporate managers and investing in AI and AWS infrastructure. The money that used to fund mid-level management salaries is being redirected to GPU clusters and data centers.

This is the pattern playing out across the industry. The headline says “layoffs.” The actual story is: companies are reallocating human capital budget to AI capital expenditure. People aren’t being replaced by AI directly — they’re being replaced by the investment in AI.

The Quit Rate Tells the Real Story

Here’s the data point I keep coming back to: IBM’s voluntary attrition dropped from 7% to under 2%. That’s the lowest in three decades, in an industry that typically runs 13-21% annual turnover.

This isn’t employee loyalty. It’s employee paralysis. People aren’t quitting because they look at the market and see nowhere to go. The “Great Resignation” has become the “Big Stay” — not because people are happy, but because the alternative is worse.

And some companies are weaponizing this. According to survey data, 25% of executives admit their return-to-office mandates are designed to encourage people to quit — attrition without severance packages, headcount reduction without layoff headlines. One employee quoted in the research called it “a way to cut headcount without headlines,” expecting 10-15% voluntary attrition.

As an engineering leader who cares about retention, this makes me sick. Using workplace policy as a stealth layoff tool destroys trust in exactly the institution — the employer — that workers are clinging to because everything else feels unstable.

The Skills Mismatch Nobody’s Solving

LinkedIn data shows job applications have doubled since 2022, yet recruiters say they can’t find talent. How can both be true?

Because the skills market has split. AI and ML postings have surged while everything else has contracted. Companies are hiring for prompt engineering, model fine-tuning, and inference infrastructure while cutting traditional frontend, backend, and mobile roles. The applicant pool is enormous for roles that are disappearing and tiny for roles that are emerging.

CS graduates face 6.1% unemployment versus 3.6% overall. Computer engineering graduates are at 7.5%. These aren’t people who lack ambition or education — they’re people whose skills were exactly what the market wanted three years ago and aren’t what the market wants now. The transition speed is brutal.

What I’m Doing About It (Honestly)

I can’t fix the macro environment. But here’s what I’m trying at my company:

1. Internal mobility over external hiring. When we need new capabilities, we train existing team members first. This is slower but builds loyalty and addresses the skills gap without asking an already-frozen market to produce candidates that don’t exist.

2. Honest conversations about workload. We stopped pretending AI closes the headcount gap. I told my team leads: “We are understaffed and that’s not changing this quarter. Let’s re-prioritize instead of pretending we can do everything.” Acknowledging the constraint reduced burnout complaints more than any wellness initiative.

3. Skills mapping against the market. Every engineer on my team has a “skills adjacency map” showing where their current skills connect to high-demand areas. If someone wants to transition toward ML engineering or platform work, we create a path. This isn’t altruistic — it’s strategic. If people can’t grow externally, they need to grow internally or they disengage.

4. Pushing back on the board. I made the explicit argument that our velocity is declining because we’re asking 80 people to do the work of 100 and calling it “efficiency.” We got approval for 4 hires — not the 12 we asked for, but better than zero.

The Question That Haunts Me

If this equilibrium holds — low hiring, low firing, rising applications, shrinking postings — what happens to an entire generation of engineers who can’t get their first role? What happens to mid-career engineers who can’t move? What happens to the senior engineers carrying unsustainable workloads with no relief in sight?

The market isn’t crashing. It’s calcifying. And I’m not sure which is worse.

[product_david]

@vp_eng_keisha, I want to push on the “Amazon as microcosm” framing, because I think it actually reveals something more interesting than restructuring. It reveals the fundamental repricing of what companies value.

The Capital Reallocation Is the Story

You nailed the key insight: Amazon isn’t shrinking, they’re redirecting. The $125 billion CapEx budget for 2026 is being funded partly by the salary budget that used to go to 30,000 corporate employees. If you average those roles at $180K fully loaded, that’s roughly $5.4B in annual human capital being redirected to infrastructure.

But here’s what I want to add from the product side: this isn’t just a labor market story. It’s a business model story. Companies have decided that the marginal return on the next dollar invested in human talent is lower than the marginal return on the next dollar invested in AI infrastructure. Whether that bet is correct is a separate question. The bet has been placed.

I’m seeing this at the startup level too. When I’m in board meetings for our Series B, the first question from investors isn’t “how many engineers do you have” anymore. It’s “what’s your AI-augmented productivity per engineer.” The metric has shifted from headcount to output-per-head. And that shift fundamentally changes the hiring equation.

Ghost Jobs Are a Product Failure

The ghost job phenomenon is something I want to highlight because it connects to my world directly. Companies posting jobs they have no intention of filling — to signal growth to investors, collect resumes for future use, or just maintain a pipeline “in case” — is a product management failure applied to HR.

Think about it: if a product team launched a feature that had no intention of delivering value to users, we’d call it vaporware. When HR posts jobs that have no intention of resulting in a hire, it’s the same thing. The candidates are the “users” and the experience is broken.

I know of at least three companies in our Series B cohort that maintain active job postings specifically for investor optics. The postings look real. They have salary ranges and team descriptions. But the hiring manager has been told verbally that there’s no budget. The recruiter screens candidates they’ll never extend offers to. It’s theater.

This isn’t just ethically questionable — it poisons the data that everyone uses to understand the job market. When Indeed reports tech postings are at a certain level, some percentage of those are ghost listings. The actual effective job market is even worse than the statistics suggest.

The Product Leader’s Frozen Market Experience

From my seat as VP Product, the freeze manifests differently than engineering. We’re not hiring for my product team either, but the constraint creates different pathologies:

Scope creep without staffing. My team of 8 PMs is covering what used to be 12 PM roles. Everyone is spread across too many products. Quality of product thinking declines because nobody has time for proper discovery or user research. We’re making decisions faster — but they’re worse decisions.

The “AI PM” expectation. Our board has started asking why we need dedicated PMs when “AI can do the analysis.” I had to build a demo showing the difference between GPT-generated product requirements and proper customer-informed ones. The GPT version looked polished but was solving the wrong problem. The gap is invisible to non-practitioners.

Hiring hoarding. When headcount does open up, the political dynamics change. Every function fights for the slot because they know it might be the only hire approved for six months. This creates internal competition that didn’t exist when hiring was fluid.

What Breaks the Equilibrium?

I keep asking myself: what would cause the freeze to thaw? A few scenarios:

1. Rate cuts. If the Fed cuts meaningfully, companies regain confidence in growth spending. But rates have been higher-for-longer and companies have adapted to lean models they might not want to reverse.

2. AI productivity plateau. If it becomes clear that AI isn’t actually replacing the headcount companies thought it would, the pressure to rehire builds. I think we’re approaching this point — the gap between AI promise and AI delivery is becoming measurable.

3. Competitive talent pressure. Eventually, the companies that maintained hiring during the freeze will have a talent advantage. When one company breaks ranks and starts hiring aggressively, others follow out of fear of falling behind. This is what happened in 2020-2021.

4. Burnout cascade. If the “Big Stay” workers finally start leaving en masse — not for better jobs, but just to leave — companies face a retention crisis they didn’t budget for. The cost of replacing institutional knowledge at scale would force rehiring.

My money is on a combination of 2 and 4. The AI productivity plateau is already visible in the data (the METR study showing 19% slowdown, the trust decline surveys), and burnout is a when-not-if proposition at current workload levels.

[alex_dev]

I want to offer the ground-level perspective here because the view from engineering leadership and the view from the desk of a senior IC are very different right now.

The “Big Stay” Doesn’t Feel Like Stability

@vp_eng_keisha, you described the “Big Stay” perfectly from the leadership side. From the IC side, let me tell you what it actually feels like.

I’m a senior full-stack engineer with 7 years of experience. By every pre-2023 metric, I should be in a strong market position. I write code, I ship features, I mentor people. I have exactly the profile that used to get recruiter messages every week.

My last LinkedIn recruiter message was three months ago. It was for a contract role paying 40% less than my current salary. Before 2023, I averaged 4-5 recruiter contacts per week. The pipeline hasn’t just slowed — it’s essentially stopped.

And yet I’m not “unemployed.” I’m not even unhappy at my current job. But I’m stuck. The normal career progression mechanism — use an outside offer to negotiate a raise, or move to a more interesting role at a competitor — doesn’t function when there’s nowhere to move to. My total compensation has been flat for two years because there’s no external leverage, and the company knows it.

This is the invisible part of the freeze for employed engineers: the career trajectory stalls even if the paycheck continues.

What the Application Black Hole Looks Like

I have a friend — senior backend engineer, 5 years experience at a well-known fintech, strong GitHub, solid system design skills — who’s been looking for three months. Here are her numbers:

• Applications submitted: 142
• Automated rejections (within 24 hours): 89
• Recruiter screens scheduled: 12
• Technical interviews completed: 4
• Offers received: 0
• Ghost jobs discovered (applied, then found the role was already filled or never existed): at least 15

That’s a 0% offer rate after 142 applications for someone with legitimate experience. She’s not entry-level. She’s not switching careers. She’s a mid-senior engineer looking for lateral moves.

The AI gatekeeping @product_david mentioned is part of this. Her resume was keyword-optimized, but ATS systems are now so aggressive that minor formatting differences or the absence of specific tool names can get you filtered before a human ever sees you. She reformatted her resume five times using different ATS-optimization tools and her callback rate went from 3% to 8.5%. Same experience, same person — just different keyword density.

The Psychological Toll Nobody Measures

Here’s what I don’t see discussed enough: the mental health impact of the freeze on employed engineers who feel trapped.

At my company, I see three patterns playing out:

The Disengaged. Engineers who’ve accepted they can’t leave, so they do the minimum required. Not quiet quitting — more like quiet resignation to the situation. They used to propose new features and volunteer for complex projects. Now they close their assigned tickets and log off. The spark is gone because the reward mechanism (career growth, recognition, mobility) is broken.

The Anxious Overachievers. The opposite reaction — engineers who work harder because they’re terrified of being on the wrong end of the next cut. They volunteer for everything, work evenings, and visibly demonstrate productivity. It looks like engagement, but it’s fear. And it’s not sustainable.

The Side-Hustle Hedgers. Engineers building consulting practices, indie products, or AI wrapper apps on evenings and weekends. They’re not leaving their jobs, but they’re creating escape routes. One of my teammates ships a micro-SaaS product that makes K/month. He told me: “My salary pays the bills. This is my insurance policy against the day they decide AI can do my job.”

All three patterns erode team culture, but in ways that don’t show up in sprint velocity or DORA metrics.

What I’d Tell Engineers in This Market

For what it’s worth:

1. Stop applying through front doors. The ATS black hole is real. Every job I’ve gotten in my career came through referrals or direct outreach. In this market, that’s doubly true. Your network is your only real channel.

2. Build visible expertise in one niche. “Full-stack engineer” is a commodity label. “Engineer who built a real-time streaming pipeline at scale” or “engineer who reduced inference latency by 80%” is a story. Specificity cuts through noise.

3. Don’t confuse the market with your worth. The freeze is macroeconomic. It’s not a reflection of your skills. The same engineer who can’t get a callback today would have had 5 offers in 2021. The variable that changed isn’t you — it’s the environment.

4. Invest in adjacent skills. If you’re a traditional web developer, learn enough ML engineering to be conversational. Not to become an AI researcher — to be the person who can bridge product engineering and ML infrastructure. The bridge roles are the ones hiring.

The freeze will thaw eventually. But “eventually” is cold comfort when your career feels stuck right now.

[cto_michelle]

This thread captures every dimension of what I’ve been wrestling with as a CTO scaling from 50 to 120 engineers. I want to add the strategic leadership perspective because I think the freeze creates a window of opportunity that most companies are too scared to see.

The Contrarian Take: This Is the Best Hiring Market in a Decade (If You’re Actually Hiring)

I know that sounds tone-deaf given everything @alex_dev and @vp_eng_keisha described. But hear me out.

We’re one of the companies that decided to keep hiring through the freeze. Not aggressively — we added 22 engineers in 2025, a modest number. But here’s what happened: our quality of hire went through the roof.

In 2021-2022, we were competing with Google, Meta, and every well-funded startup for the same candidates. Our offer acceptance rate was 35%. Candidates had 4-5 competing offers. We were paying above market and still losing people.

In 2025, our offer acceptance rate was 78%. The candidates we hired have, on average, 2 more years of relevant experience than our 2022 hires. Our technical interview pass rate actually went down — we raised the bar because we could afford to — and we’re still filling roles faster than we did during the boom.

The paradox of the frozen market: for the small number of companies still hiring, it’s a buyer’s paradise. And the talent is exceptional because skilled people who want to move have almost nowhere to go.

Why Most Companies Won’t Take the Opportunity

So if the talent is available and the quality is high, why aren’t more companies hiring? Three reasons:

1. The CEO-CFO feedback loop. CEOs read the same headlines about AI replacing jobs. CFOs model the same “efficiency” gains. They reinforce each other’s belief that hiring is unnecessary. Breaking out of this loop requires a CTO who can articulate the specific, measurable cost of not hiring — and most CTOs aren’t making that case forcefully enough.

2. The “prove it first” trap. @vp_eng_keisha’s board telling her to “prove you can do more with what you have before we fund new hires” is the classic version of this. It’s circular logic: the team can’t prove capacity because they’re understaffed, and they can’t get staff because they haven’t proven capacity. Someone has to break the cycle with an investment thesis, not a proof point.

3. Institutional inertia. Companies that froze hiring in 2023 have now operated in freeze mode for almost three years. It’s become the default. The muscle memory of recruiting, onboarding, and integrating new hires has atrophied. Even getting a req approved is a multi-week political process. The organizational infrastructure for hiring has degraded, and rebuilding it feels like too much effort.

The Long-Term Cost Nobody’s Modeling

@product_david’s point about repricing human capital versus AI infrastructure is accurate, but I think the math is wrong on a 5-year horizon. Here’s why:

US tech employment is projected to grow from 6.09 million in 2025 to 7.03 million by 2035. That’s nearly a million new roles over the next decade. The demand isn’t disappearing — it’s being deferred and transformed.

Companies that maintained their talent pipeline through the freeze will have a compounding advantage:

• Institutional knowledge retention: Every year of continuity builds organizational context that can’t be hired for later.
• Cultural cohesion: Teams that grew together during adversity outperform teams assembled from scratch during a recovery.
• Skills development: Engineers who trained through the AI transition internally are worth more than engineers who learned AI skills independently without enterprise context.

The companies that cut to the bone and plan to “rehire when things recover” are going to find that recovery hiring is 3-4x more expensive than retention would have been, and the people they want will have been hired by the companies that never stopped investing.

What I’m Telling My Board

My argument to my board is simple: the freeze is a strategic opportunity disguised as a constraint. Every competitor that stops hiring makes our hiring more effective. Every month of the freeze extends our lead in talent accumulation. The ROI of hiring in a frozen market is higher than hiring in a hot market because the talent quality is better, the compensation pressure is lower, and the retention risk is minimal (because where would they go?).

This isn’t charity. It’s strategy. And the data supports it: our engineering velocity per engineer has increased 15% year-over-year despite growing headcount, because the people we’re hiring are higher quality than anything we could have attracted in 2021.

To @alex_dev’s Three Patterns

Your taxonomy of Disengaged, Anxious Overachievers, and Side-Hustle Hedgers is painfully accurate. I’d add a fourth: The Quiet Interviewers. Engineers who take every rare interview opportunity not because they intend to leave, but to calibrate their market value and maintain interviewing skills. They’re mentally half-out even though they’re physically present.

The aggregate impact of all four patterns is what I call “engagement debt.” Just like tech debt, it compounds invisibly until it reaches a crisis point. And when the market does thaw — and it will — the companies carrying the most engagement debt will experience the sharpest departures.

The leaders who invested in their people during the freeze will keep them. The leaders who took the low quit rate as a license to under-invest will lose them overnight.

The freeze doesn’t last forever. What you did during it does.