All podcasts / Lenny's Podcast / Summary

Why the AI's honeymoon is ending (and tech workers are feeling it) | Noam Segal

2026-07-12 - 96 min - source - Read full transcript

Lenny Rachitsky (host)Noam Segal

Key insights

The tech workforce is splitting roughly in half on how AI has changed their professional identity, and this split predicts almost every other wellbeing measure.

In the 2026 survey (~6,000 respondents), about 50% report feeling amplified by AI while the other half splits into redefined (27%), destabilized (14%), and diminished (5%). This identity-stance variable has an effect size on optimism, burnout, and layoff worry roughly 3x larger than previously identified top effects like manager quality or founder status.

workforce-bifurcation

Significant burnout jumped from 44.7% to 54.7% year over year, while career optimism fell from 54.8% to 48.7%.

Noam expected new AI capabilities to reduce burnout by lowering effort required, but the opposite happened: teams are shipping far more (e.g., 30 PRs a day instead of a couple) and the extra output is being absorbed as new baseline expectation rather than relief, so people work just as hard while doing more.

ai-sentiment-and-burnout

No functional group in tech would recommend their own role to someone entering the industry now, not even founders.

On a 0-10 recommend-your-role question converted to NPS logic (-100 to +100, zero = neutral), every group scored at or below neutral. Founders and higher-seniority people (VPs, execs) scored least negative; designers and researchers scored most negative, followed closely by data/analytics workers who are most worried about AI directly replacing their jobs.

role-and-identity-disruption

The dominant fear in tech right now is being expected to do more for the same pay, not job loss to AI.

When asked what they're afraid of, 'losing my job to AI' ranked second to last. The top fears were the expectation to do more for the same compensation and the pace of change (both work velocity and the pace of tooling change) becoming unsustainable, which Noam frames as 'the speed AI unlocked got plowed straight back into expectations.'

ai-sentiment-and-burnout

AI is widely felt to make people faster, not necessarily better, and many report a felt decline in thinking and judgment.

97.2% say AI makes them better at their job and ~50% say much or extremely better, but when probed, that mostly means doing more work faster rather than higher-quality output. Respondents described 'my brain is rotting' and reduced ability to code or write strategy docs unaided, a phenomenon Noam ties to skipping judgment and just accepting AI's first-pass output ('cognitive rot').

productivity-vs-quality

Manager effectiveness is one of the largest levers on burnout and job enjoyment, and most managers are rated poorly.

Only about 25% of respondents rate their manager as highly effective versus 36% rating them ineffective (little changed from last year). People with highly effective managers report roughly 65% higher job enjoyment and dramatically lower burnout, making manager investment, per Noam, 'the biggest lever' companies have to improve retention.

management-and-career-advice

Founders and people at smaller companies remain the happiest across nearly every measure, for the second year running, but even they aren't fully positive.

Founders report 71% optimism, the lowest burnout, and the most AI excitement, and burnout climbs in a near-linear fashion as company size grows from 1-10 person startups to 5,000-10,000 person enterprises. Still, 47% of founders report at least moderate burnout, and founders also do not recommend founding a company to others, undercutting the 'best time ever to be a founder' narrative.

workforce-bifurcation

Designers, researchers, and data/analytics workers report the worst sentiment of any function, both this year and last.

These groups score highest on feeling destabilized or diminished, highest on tired/overwhelmed/anxious emotions, highest on fear of AI-driven job loss (data/analytics specifically), and lowest on recommending their role to others. Noam argues this doesn't reflect AI's actual current capability in these disciplines, only how people in them currently feel.

role-and-identity-disruption

Emotions in tech right now are genuinely mixed rather than polarized into pure hype or pure doom.

The top two reported emotions are positive (curiosity, excitement), but overwhelmed, conflicted, tired, and anxious also rank highly. Noam describes this as 'smiling exhaustion' (a term from Nikhyl Singhal): people feel reborn and energized building with AI while simultaneously having no off switch, and argues most individuals hold both the hype and the doubt within themselves rather than being purely one or the other.

ai-sentiment-and-burnout

Early-career workers report 'AI guilt,' feeling like using AI is a form of cheating, and that guilt declines with seniority.

The guilt is strongest among early-career people and, among functions, shows up most in product marketing and data/analytics. Noam compares it to imposter phenomenon, where competent people attribute their success to something external (in this case, the tool) rather than themselves, and advises leaning into AI rather than feeling guilty about it since 'it's the worst it's ever going to be today.'

role-and-identity-disruption

Concrete advice for employees: go deep on a few AI use cases instead of trying to be a generalist, protect your relationship with your manager, and watch for scope creep without comp changes.

Survey data shows people who feel amplified tended to focus AI on specific tasks/jobs-to-be-done rather than spreading it across everything; over-generalizing correlates with more severe burnout. Other advice: actively flag 'the squeeze' (more scope, same pay) with your manager, invest in the manager relationship since it's one of the strongest predictors of wellbeing, and seek mentorship if early in career since the traditional ladder rungs are eroding.

management-and-career-advice

Concrete advice for leaders: invest heavily in manager training, protect entry-level career paths, and pay attention to design/research team sentiment.

Given only ~25% of managers are rated highly effective, Noam calls manager investment 'probably some of the best money you'll ever spend' for improving retention and burnout, especially given aggressive poaching from AI labs. He also urges leaders not to let 'the bottom rung of the ladder rot' for junior hires (who tend to be the most AI-native) and to watch that negative sentiment among design/research managers doesn't get passed down to their reports.

management-and-career-advice

Media referenced

Tech Sentiment Survey 2026 report (Noam Segal / Lenny's Newsletter) - article - The underlying ~6,000-respondent survey and write-up the whole episode walks through; second annual run after 2025's 'Burnt Out But Optimistic'.
Elena Verna's post on 'AI confidence theater' - article - Referenced at the top as the prompt for framing the survey as a corrective to overhyped 'everything is dead' tech discourse.
Conversation with Scott Wu (Cognition/Devin) on Lenny's Podcast - podcast - Noam borrows Scott Wu's 'ladder' metaphor (high-school CS student to staff engineer) to describe AI climbing skill rungs and pulling them out from under workers.
Conversation with Jeff (Ramp) on Lenny's Podcast - podcast - Cited for the idea that burnout tracks with low velocity, which Noam contrasts with the current burnout driven by high output.
Conversation with Simon Willison on Lenny's Podcast - podcast - Referenced on skill atrophy and the need to consciously resist cognitive rot from over-relying on AI.
Conversation with Nikhyl Singhal on Lenny's Podcast - podcast - Source of the 'smiling exhaustion' term Noam uses to describe feeling reborn by AI work while having no off switch.
The Terminator (film franchise) - movie - Lenny and Noam joke about Skynet as a half-serious analogy for AI capability racing ahead of workers.

Companies

Whatnot - Mentioned only in passing via other episode cross-references, not a focus here.
WorkOS - Season-presenting sponsor; pitched as enterprise-readiness APIs (SSO, SCIM, RBAC, audit logs) for B2B SaaS.
Mercury - Sponsor; banking product for startups, pitched via its new conversational 'Command' feature.
Anthropic - Jony Wend, who leads design for Claude, cited as an example of a design leader publicly emphasizing taste and craft.

Techniques and frameworks

AI identity stance framework (amplified / redefined / destabilized / diminished) - The survey's core segmentation of how AI has shifted a worker's professional identity, used to explain the ~50/50 split in the tech workforce.
Four tech-worker archetypes (energized, conflicted, disoriented, resentful) - Derived from self-reported emotion clusters; used to describe the emotional texture behind the identity-stance numbers.
ARMOR framework - Burnout-fighting framework Noam and Lenny built after the 2025 survey found high burnout; referenced but not re-explained in depth.
Cohen's d (effect size) - Used to argue the AI-identity effect on job satisfaction is about 3x larger than the next-largest known effects (manager quality, founder status).
NPS (Net Promoter Score) applied to career recommendation - Used to measure whether people would recommend their current role to someone entering tech now; every function scored negative or barely neutral.

Summary

Lenny Rachitsky and longtime research partner Noam Segal (former research leader at Airbnb, Twitch, Twitter, Intercom, Zapier, and Figma) unpack the second annual Tech Sentiment Survey, a roughly 6,000-person study Noam ran across product, engineering, design, research, and other tech roles. The headline finding is a stark bifurcation: about half the workforce feels amplified and energized by AI, while the other half splits into feeling redefined, destabilized, or diminished. That single variable, how AI has shifted someone's professional identity, predicts nearly every other measure of wellbeing (optimism, burnout, layoff worry, willingness to recommend the role) with an effect size roughly three times larger than previously known drivers like manager quality or founder status.

The numbers on burnout and optimism worsened year over year: significant burnout rose from 44.7% to 54.7%, while career optimism fell from 54.8% to 48.7%. Counterintuitively, more capable AI hasn't meant less work, it's meant more output expected at the same pace, which both hosts frame as new capability being "plowed straight back into expectations." When asked directly what scares them most, respondents ranked "losing my job to AI" second to last; the top fears were being expected to do more for the same pay and the pace of both work and tool change becoming unsustainable. On productivity, 97% report AI makes them better at their job, but drilling down reveals that mostly means faster output, not higher quality, and many describe a felt erosion of judgment and thinking skill ("my brain is rotting"), which Noam ties to skipping critical engagement with AI's first-pass output.

A recurring, almost NPS-style question, would you recommend your current role to someone entering tech now, produced a striking result: no function scored positive. Designers and researchers scored worst, consistent with last year, followed closely by data/analytics workers who separately report the highest fear of direct AI job replacement. Founders and smaller-company employees remain the happiest cohort for the second year running (71% optimism, lowest burnout), though even founders show meaningful burnout (47% at least moderate) and don't recommend founding to others, complicating the "best time ever to start a company" narrative. Manager effectiveness emerges as one of the single largest levers on wellbeing: only about 25% of managers are rated highly effective, yet those who are drive roughly 65% higher job enjoyment and much lower burnout among their reports, making manager investment what Noam calls the biggest available lever for retention.

The conversation closes with practical advice split for employees and leaders. Employees are encouraged to go deep on a few specific AI use cases rather than trying to be an AI-powered generalist (a pattern correlated with more severe burnout), to actively flag scope creep with managers rather than absorb it silently, to invest in the manager relationship, and, if early-career, to seek out mentorship since traditional ladder rungs are eroding faster than before. Leaders are urged to invest in manager training given the poor baseline, to protect entry-level roles and development rather than letting the bottom of the ladder rot, and to pay attention to design and research teams' sentiment, since manager negativity in those groups risks compounding onto their reports. A closing note names a new pattern, "AI guilt," strongest among early-career workers and in product marketing and data/analytics, where using AI feels like a form of cheating, which Noam likens to imposter phenomenon and argues is unwarranted given how fast the tools are improving.

Notable Quotes

"The honeymoon period with AI is over." - Noam Segal

"Full gas on neutral, right? Like you're pushing the pedal to the metal in your car, but you're in neutral gear. You're not going anywhere." - Noam Segal

"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." - Noam Segal

"We're in the second inning of a shift. No one knows how it will end, but all you can do is keep taking at bats." - Noam Segal (quoting a survey respondent)

"It's the worst it's ever going to be today. It's only getting better. So there's no reason to feel guilty about leveraging it." - Noam Segal