How tech workers are feeling in 2026: a workforce splitting in two with Noam Segal

Why does it matter? Because the same technology is making half the workforce feel superhuman and quietly degrading the thinking of the other half

A survey of 6,000 tech workers — engineers, PMs, designers, researchers, founders — reveals that the most consequential variable shaping career satisfaction in 2026 isn't compensation, company prestige, or manager quality. It's how AI has shifted your professional identity. Noam Segal, who has led research at Intercom, Twitter, Meta, Zapier, and Figma, ran the numbers and found a workforce cleaved down the middle, with the same technology generating superpowers on one side and degrading judgment on the other.

• Your AI identity stance now predicts wellbeing three times more powerfully than manager quality — last year's top predictor. • Burnout jumped 10 percentage points in a single year, caused not by stagnation but by too much velocity. • 97% say AI makes them better; when you press that claim, "better" means faster, not higher quality. • No role in tech — not even founders, the happiest cohort by every measure — would recommend their career to someone entering the industry today.

Your AI identity stance predicts your wellbeing three times more powerfully than your manager does

Where you land on AI's impact on your professional identity now predicts your wellbeing at work more powerfully than who your manager is, what company you work for, or what your title says — by a factor of three.

Segal measured this using Cohen's D, an effect-size metric that captures practical rather than statistical significance. Last year's two strongest effects were manager quality and the "founder happiness effect" — founders, with their autonomy and agency, reported dramatically better wellbeing than typical employees. This year's AI identity finding dwarfs both. "This insight around the impact of AI on your identity and on every single other variable around your job is about three times as large as those other effects," Segal says. Only 3% of the 6,000 surveyed workers say AI hasn't shifted their professional identity at all.

The breakdown: 50% feel amplified — more capable, operating at a scale they couldn't reach before. The rest fractures into three groups: 27% feel their role is being redefined without clarity on what it becomes; 14% feel destabilized, reporting high anxiety and pessimistic outlooks; 5% feel diminished, convinced AI has taken something that isn't coming back. Given models only improve, Segal notes the diminished group will "probably" feel more so over time.

All four groups predict everything downstream — career optimism, burnout, layoff worry, willingness to recommend their career to newcomers — in a clean, linear pattern. If this is now the highest-impact variable on career satisfaction, evaluating your work situation through old lenses (manager, comp, company prestige) without addressing AI stance is optimizing for the second-order variable.

The real fear isn't being replaced by AI — it's being squeezed to do more for the same pay

Losing a job to AI ranks second-to-last on the list of tech worker concerns. What rose to the top: the expectation to do more for the same pay.

Segal surfaced workers' fears directly and watched where things landed. Above layoffs and obsolescence: the quiet grinding reality of scope creep, with every productivity gain immediately absorbed into higher baseline expectations, compensation unchanged. "The speed AI unlocked got plowed straight back into expectations. Every gain becomes the new baseline and the people expected to hit it are running out of room to breathe."

The second fear tracks closely: the pace becoming unsustainable — both the velocity of work output (more PRs, more PRDs, more prototypes, more campaigns, more agents, every day) and the velocity of the technology itself, which demands constant re-learning on top of an already maxed schedule. "It's sort of an evil downward spiral that really prevents you from feeling like you're able to accomplish anything of meaning. You just feel very overworked, very tired."

The retention risk this creates is invisible on most dashboards. Leaders watching for signs of existential AI anxiety will miss the quieter, more pervasive threat: people doing significantly more with nothing changing in return. For ICs, the prescription is to name the scope creep explicitly in the next manager conversation — frame it as a calibration, not a complaint.

97% say AI makes them better at their job — but "better" turns out to mean faster, not higher quality

97.2% of tech workers say AI makes them better at their job. Nearly half say "very much" or "extremely" better. The number is almost certainly measuring the wrong thing.

When Segal pushed below the surface, two findings undercut the headline. First, workers were describing speed, not quality: "I can do more faster, but not better." AI has lowered the floor — far more output is now possible — without raising the ceiling of what that output achieves. Second, and more disturbing: "People essentially told us my brain is rotting. My work feels worse."

Segal connects this to cognitive rot — the pattern of accepting AI's initial output without applying judgment, letting your own analytical capacity quietly atrophy. "Every time you — not the AI — solve a problem and get over a barrier, it raises your baseline of self-efficacy, of self-confidence, of self-belief. And every time you offload that to your favorite AI model, you're lowering that baseline and your thinking and your judgment is wasting. That's a serious problem."

Workers in the survey are aware this is happening. What's absent is deliberate action against it. Segal's own distillation: "The productivity gains are real, but the quality of the work and the sharpness of the person producing it are taking a hit." Treat AI like any atrophy risk — carve out protected time to wrestle with hard problems without assistance before the habit calcifies.

Burnout hit 54.7% — a 10-point surge in a single year, driven by too much momentum, not too little

The old burnout model said stagnation was the cause: effort going in, nothing moving. This data says the opposite — more than half of tech workers are significantly burned out (54.7%, up from 44.7% last year) precisely because too much is moving.

"We're not working any less hard," Segal says. "We're just taking on more stuff, more prototypes, more PRDs, more PRs, more campaigns, more agents." Shipping more than ever, running out of room to breathe. The technology itself compounds this: Segal admits he's been using it "for too many hours a day playing about and building," and the addictive quality makes it nearly impossible to disengage.

Nikhil gave this a name in a previous episode: smiling exhaustion. The old burnout was grim — disengagement, meaninglessness, grinding against inertia. The new version looks different from the outside. "It's about people almost feeling reborn on the one hand. You know, I'm shipping again. I'm building. I'm creating these incredible things with AI. There's never been a more exciting time, but there's no off switch."

Disengagement used to be the early warning signal. You can now be genuinely excited, proud of output, energized by the technology — and burning out simultaneously. Optimism has fallen from 54.8% to 48.7% in that same year. If you're excited and exhausted with no natural stopping point, don't wait for the disengagement that may never arrive.

Not a single role in tech — not even founders — would recommend their career to someone entering the industry now

Founders are 71% optimistic. They have the lowest burnout, lowest layoff worry, and most AI excitement of any group surveyed. They still wouldn't recommend their path to someone considering entering tech.

Segal ran an NPS-style question across every function. Scores run from -100 to +100, with zero representing neutrality — neither promoting nor discouraging. Every group landed below zero. Engineers, PMs, designers, researchers, sales, operations, founders: all negative. Designers and researchers sit furthest into detractor territory; founders are the least-worst. "It's sort of like saying, you know, I'm kind of swimming in this pool. The water's kind of okay, but you shouldn't come into these tech waters. They ain't for you."

This is forward-looking pessimism, not present dissatisfaction. When asked directly about current job enjoyment, most workers said they were doing fine. The gap is between today's experience and what people are extrapolating for one, three, five years out — and what they extrapolate is clearly darker.

The metaphor Segal offers: Devon's capability ladder, moving from high school CS student to college intern to junior engineer to senior. "We are all watching this technology climb the rungs of this ladder and advance and we feel like the technology is pulling those rungs from under us." Lower seniority means greater exposure to that pull — junior ICs are the least likely of all to recommend their path. Anyone advising early-career people should update the conventional script: the emphasis needs to shift toward judgment, taste, and adaptability — the things AI can't replicate.

Effective managers produce 65% higher job enjoyment — and the great flattening is shrinking the supply at exactly the wrong moment

A highly effective manager produces 65% higher job enjoyment and dramatically lower burnout. Only 25% of workers have one — and the industry, in its push toward flatter organizations and founder mode, is actively reducing that supply.

The manager finding is the most consistent result across two years of data: effectiveness is among the largest controllable variables in worker wellbeing. The inverse is equally stark — 36% of workers rate their manager as ineffective, a number barely changed from last year.

What's new is the structural headwind. "We are in the era of the great flattening. People have more direct reports now than ever before. I'm concerned about this." Managers are the primary filter through which the AI-driven productivity squeeze either gets moderated or passes through to individual contributors unchecked. "Who's the person more than anyone else who kind of manages that squeeze? It's your manager. It's your manager who protects you." Removing layers while AI is escalating workload expectations is precisely the wrong trade.

The worst-rated managers cluster in data analytics and design — the same functions under the heaviest AI-related stress — which Segal reads as a transmission effect: managers who are struggling in their own careers tend to pass that strain downstream to their reports.

The conclusion holds even given all the structural friction: investing in manager development is the highest-ROI retention lever available. For anyone evaluating a new role, manager quality should outweigh almost everything else on the list.

35% of tech workers are simultaneously having the most fun and most fear of their careers — that's the statistically normal state right now

Sentiment analysis of thousands of open-ended responses produced a near-perfect even split: 37% positive words, 37% negative, 26% neutral. The dominant emotional state in tech in 2026 isn't enthusiasm or dread — it's both at once.

Segal's four archetypes map the terrain: Energized (41%), who describe tech as a "tech amusement park" with capabilities they've never had; Conflicted (35%), who are having "the most fun they've ever had as builders, as PMs, as designers, as engineers" while simultaneously feeling "the most uncertainty they've ever felt in their careers"; Disoriented, whose roles keep shifting without a visible endpoint ("We're like farmers on the cusp of the industrial revolution and we just don't see a clear path to what's happening"); and Resentful (12%), who feel coerced into using a technology they resent — and are still watching colleagues lose their jobs anyway.

A senior PM surveyed captured the external view: "Tech is manic. Half out of touch, clinging to the bandwagon, probing into the overhype. The other half are exhausted by the first half." On average, respondents selected five emotions simultaneously when describing how they feel; some chose thirteen.

The permission this data grants is worth stating plainly: contradictory emotions aren't a sign of confusion. "We all have both the hype person within us and the doomer to a certain extent," Segal says. "The only difference between us is that we have different amounts of each." Performing certainty in either direction misrepresents what most people are actually living. Stop resolving the tension — start operating inside it.

The split will deepen before it resolves

The AI identity positions Segal maps — amplified, redefined, destabilized, diminished — aren't fixed traits. They're outcomes being shaped right now by choices most organizations aren't explicitly tracking: how they handle the productivity squeeze, whether they invest in managers, how deliberately they protect human judgment from atrophy. These choices compound.

The technology that produced today's 50/50 split will be significantly more capable in six months. For the 5% who already feel diminished, that trajectory is unambiguously worse news.

The second inning looks like this. Nobody knows what the ninth looks like.

---

Topics: AI adoption, tech workforce, burnout, career development, manager effectiveness, product management, engineering culture, workforce sentiment, job satisfaction, AI identity, cognitive atrophy, founders, design and research roles