Lenny's PodcastWhy the AI’s honeymoon is ending (and tech workers are feeling it) | Noam Segal
CHAPTERS
- 0:00 – 1:13
AI sentiment survey preview: burnout up, optimism down, and a 50/50 split
Noam Segal and Lenny Rachitsky open with the headline findings: the AI honeymoon is fading, burnout is rising, and the biggest fear isn’t job loss—it’s being expected to do more for the same pay. They set up the episode as a walkthrough of the annual tech worker sentiment survey results and what they imply for careers.
- •Burnout is worsening and optimism is slipping across tech
- •Workforce is increasingly polarized in how AI feels/lands for them
- •Top fear: higher expectations without higher compensation, not direct AI replacement
- •Framing: survey captures emotions and lived experience, not objective job-market “truths”
- 1:13 – 3:01
Who Noam Segal is and why this survey exists (annual benchmark for tech mood)
Lenny introduces Noam’s background across major tech companies and explains why they’re running a yearly sentiment survey to track how tech workers feel over time. The goal is to capture a richer picture than typical internal engagement surveys.
- •Noam’s research leadership background (Airbnb, Intercom, Twitter, Meta, etc.)
- •Annual survey intended as a longitudinal benchmark for tech sentiment
- •Focus areas: AI, burnout, layoffs, career optimism, role outlook
- •Motivation: go deeper than standard manager/leadership pulse surveys
- 3:01 – 5:41
Survey methodology and scope: ~6,000 respondents across tech roles
Noam outlines the study design and scale, noting it’s one of the largest looks at tech-worker feelings about AI and work. They compare this year’s theme to last year’s “Burnt Out But Optimistic” baseline.
- •Second year of the tech sentiment survey; intent to repeat annually
- •~6,000 respondents across product, engineering, design, research, marketing, etc.
- •Last year’s headline: high burnout + meaningful optimism; led to burnout follow-up work
- •This year’s results reveal sharper tension and polarization
- 5:41 – 8:36
Core finding: AI is splitting the workforce—and it dominates every other factor
They describe the main insight: AI’s impact on professional identity is the strongest predictor of how people feel about work, bigger than role, company size, or level. This ‘bifurcation’ explains why some are thriving while others feel destabilized.
- •AI has an outsize impact on well-being compared to other job characteristics
- •Roughly half feel ‘incredible/energized,’ half feel unclear/destabilized/diminished
- •Only ~3% say AI hasn’t shifted their professional identity
- •Effect size framing: AI identity impact is ~3× larger than other big drivers like manager quality
- 8:36 – 13:02
AI identity stance breakdown: amplified, redefined, destabilized, diminished
Noam introduces the specific identity categories and their prevalence. The categories capture not just adoption of tools, but the emotional and identity-level meaning people attach to AI in their work.
- •~50% feel amplified (positive, empowered)
- •~27% feel their role is being redefined (uncertain, mixed clarity)
- •~14% feel destabilized (anxious, shaky ground)
- •~5% feel diminished (AI ‘taking something away,’ expected to worsen as models improve)
- 13:02 – 14:38
How identity maps to outcomes: optimism, burnout, layoff worry, and recommending the role
They show how moving from amplified → diminished produces a clean, linear decline in optimism and rise in burnout and layoff worry. A standout outcome: many would not recommend their role to someone entering tech now.
- •Career optimism drops sharply as identity stance becomes more negative
- •Burnout increases in step with destabilization/diminishment
- •Layoff worry rises as identity shifts negative
- •Recommendation for newcomers declines dramatically, signaling pessimism about the ladder ahead
- 14:38 – 19:35
Four archetypes of tech workers: Energized, Conflicted, Disoriented, Resentful
Using emotion selections, Noam synthesizes four personas that listeners can identify with. They discuss how these archetypes help build empathy across the ‘two halves’ of the workforce.
- •Energized (~41%): playful building, experimentation, “tech amusement park”
- •Conflicted (~35%): simultaneous fun + deep uncertainty about the future
- •Disoriented: shifting roles, unclear path (Industrial Revolution ‘farmer’ metaphor)
- •Resentful (~12%): pressured to use AI, checked out, anger about what’s being lost
- 19:35 – 24:55
Burnout surge vs. sustained enjoyment: shipping faster is fueling exhaustion
They compare year-over-year changes: significant burnout rises while optimism falls. Yet many still report enjoying work—AI enables cross-functional expression and makes previously impossible building feel accessible, even as pace becomes brutal.
- •Significant burnout jumps from ~44.7% (2025) to ~54.7% (2026)
- •Career optimism drops from ~54.8% to ~48.7%
- •Paradox: high enjoyment remains because people can explore beyond swim lanes
- •Hypothesis: shipping velocity and constant learning demands are driving burnout upward
- 24:55 – 29:22
Layoff anxiety and ambivalence: enjoyment mixed with fear of the branch being cut
They explore how widespread layoff worry coexists with excitement and productivity. People can feel energized yet fear that AI adoption and ‘agent-first’ pushes will eventually reduce headcount.
- •~72% worried about layoffs to some extent; ~41% at least moderately worried
- •Emotional clash: energized and enjoying work but pessimistic about long-term security
- •AI adoption feels like ‘cutting the branch you’re sitting on’ for some workers
- •Tech culture is increasingly defined by ambivalence rather than one dominant emotion
- 29:22 – 36:45
Would you recommend your tech role? The shocking ‘career NPS’ collapse
Noam (an NPS skeptic) explains how they adapted the NPS concept to measure role recommendation. Across functions—even founders—scores are negative, suggesting people feel okay now but don’t see a stable future for newcomers.
- •Role recommendation is negative across every function; no group is a ‘promoter’
- •Designers and researchers are least likely to recommend their role
- •Seniority effect: execs more likely than ICs to recommend (ICs feel ‘full gas on neutral’)
- •Interpretation: present experience can be fine while future outlook feels grim
- 36:45 – 45:03
The ladder metaphor: junior rungs disappearing as AI climbs upward
They use a vivid metaphor (from the Devin/Cognition discussion) to explain why early-career workers feel precarious. As AI capabilities progress up the ‘skill ladder,’ it can feel like the lower rungs are being pulled away.
- •AI is perceived as progressing from junior → senior capability over time
- •Early-career roles feel most threatened because the entry rungs weaken first
- •Explains lower recommendation rates among less-senior workers
- •Reinforces need for mentorship and structured growth pathways
- 45:03 – 53:03
AI makes us faster, not better: quality concerns and ‘cognitive rot’
Respondents overwhelmingly say AI makes them better at their job, but deeper analysis reveals ‘better’ often means faster output, not higher quality. A darker theme emerges: reliance on AI may erode thinking, judgment, and self-efficacy.
- •~97% say AI makes them better; ~50% say very/extremely better
- •Deeper read: increased throughput without commensurate quality gains
- •Skill atrophy/cognitive rot: accepting first outputs without applying judgment
- •Self-efficacy decline when problem-solving is offloaded rather than practiced
- 53:03 – 55:56
Top fear isn’t replacement: it’s being squeezed—more output for the same pay
When asked what they fear, job loss to AI ranks near the bottom. The dominant concern is that AI-driven productivity gains reset expectations, creating an unsustainable pace and constant pressure to deliver more without compensation changes.
- •“Do more for the same pay” is the #1 fear
- •Pace becoming unsustainable is the #2 fear (work + tech change/learning)
- •AI speed gains get converted into baseline expectations
- •Creates a downward spiral: more output, less breathing room, rising burnout
- 55:56 – 1:01:24
Emotional landscape of tech: curiosity, excitement, and ‘smiling exhaustion’
They review the most common emotions, showing a mix of positive and negative states living together. The phrase ‘smiling exhaustion’ captures the modern experience: energized by possibility but unable to turn the tempo off.
- •Top emotions: curiosity and excitement, followed by overwhelm and conflict
- •Relief and hope coexist with tiredness, anxiety, and unease
- •“Smiling exhaustion” describes building joy plus relentless pace
- •Rejects black-and-white framing of hype vs. doom; ambivalence is normal
- 1:01:24 – 1:12:19
Who’s struggling most and who’s happiest: design/research vs founders/small companies
Designers and researchers again show the most negative pattern across multiple measures, while founders and employees at smaller companies report higher optimism and lower burnout. They emphasize this reflects feelings more than objective role viability, and argue taste/craft will be increasingly valuable.
- •Design & research: highest destabilized/diminished identity shifts and negative emotions
- •Data/analytics shows especially high worry about losing jobs to AI
- •Founders are happiest across measures (with selection bias acknowledged)
- •Company size effect: larger companies correlate with higher burnout and lower optimism
- 1:12:19 – 1:18:53
Managers as the biggest lever: effectiveness strongly predicts burnout and enjoyment
They revisit a consistent finding from both years: manager effectiveness has a massive relationship with well-being and job enjoyment. In a ‘great flattening’ era with broader spans and fewer layers, manager quality becomes even more critical—and under-supplied.
- •Effective managers correlate with dramatically higher enjoyment and lower burnout
- •Only ~25% rate their manager highly effective; ~36% rate managers ineffective
- •Flattening/founder-mode and larger spans may weaken managerial support
- •Practical implication: improving managers is a top lever for retention and well-being
- 1:18:53 – 1:36:28
Tech feels chaotic—and what to do now (employees, leaders) + AI guilt and closing
A word cloud of open-ended responses paints tech as fast, unstable, exciting, and confusing—an even split of positive and negative sentiment. They close with concrete advice for employees and leaders, plus a note on ‘AI guilt’ among juniors, encouraging people to use AI without shame while protecting their growth and well-being.
- •Industry descriptors cluster around chaos, speed, flux, hype, opportunity, instability
- •Advice for employees: go deep on a few AI use cases; watch scope/comp squeeze; invest in manager relationship; consider smaller companies; seek mentorship
- •Advice for leaders: invest in manager training; manage expectations and pace; protect entry-level growth; support destabilized groups (e.g., design/research)
- •AI guilt: juniors feel ‘cheating’ more; guilt declines with seniority—use AI while maintaining deliberate practice
