The tech industry promised security and then built the machine that questions everyone's job — including its own.
The Summary
- Tech worker anxiety about AI displacement is "off the charts," according to Menlo Ventures partner Deedy Das, who reports workers are openly questioning whether software engineering and product management remain viable 30-year careers
- NYC entry-level tech job postings dropped 49% from 2022 to 2025 — the steepest decline of any career field — with computer science graduates now outnumbering available positions for the first time since 2010
- The existential questions tech workers are asking: "What does my career and skill mean in the age of AI?" and whether joining or launching a startup offers better odds than traditional employment
- Early-career workers face the hardest impact as AI tools automate code generation, document drafting, data analysis, and routine tasks that once served as entry points to the industry
The Signal
The people who build the future are now afraid of the thing they're building. Deedy Das at Menlo Ventures describes conversations happening across Silicon Valley and beyond: engineers questioning whether their skills will matter in five years, product managers wondering if AI will absorb their workflow, and everyone trying to keep pace with breakthroughs that arrive faster than they can be absorbed. The anxiety isn't theoretical. It's visceral.
The numbers confirm what workers feel in their gut. In New York City, entry-level tech positions collapsed by 49% between 2022 and 2025, the sharpest drop of any profession tracked. 2023 marked a watershed: for the first time since data collection began in 2010, the number of computer science graduates in NYC exceeded entry-level job openings. The conveyor belt from university to tech job broke.
"The deep amount of anxiety people in tech are feeling is off the charts."
This isn't just about fewer jobs. It's about the collapse of a career model that promised durability. Software engineering was supposed to be the modern trade — learn it once, build on it for 30 years, retire comfortable. Product management was the path for people who could bridge technical and business worlds. Both roles now face the same question: if AI can write code, analyze user behavior, and draft product specs, what's the human for?
The tools are good enough to matter but not good enough to replace — which makes them perfect for eliminating the entry-level layer. Companies don't need junior engineers to write boilerplate code when Claude or Cursor handles that in seconds. They don't need as many analysts when AI can parse usage data and suggest feature priorities. What they need are senior people who can prompt, review, and make judgment calls. The ladder still exists, but the bottom rungs are gone.
Key forces accelerating the anxiety:
- Generative AI tools now handle tasks that once served as training grounds for early-career workers
- Interest rate increases and economic pressure push companies toward efficiency over headcount expansion
- The skills gap widens: entry-level workers graduate with traditional training while senior roles require AI fluency
- NYC's high living costs compound the problem — fewer entry jobs in an expensive city means talent migrates elsewhere
Das reports workers are asking fundamental questions about whether they should pivot to startups, retrain entirely, or double down on specialization. The answers vary by experience level. Senior engineers with domain expertise can position themselves as the humans who direct AI labor. Mid-career workers scramble to become AI-native before their skills calcify. Entry-level workers wonder if they picked the wrong decade to graduate.
The Implication
If you're early in your tech career, the path forward isn't coding harder — it's coding differently. Learn to use AI tools fluently, but focus on the judgment calls AI can't make: architectural decisions, trade-offs between speed and maintainability, understanding user intent beneath feature requests. The job isn't writing code anymore. It's knowing what code to write and whether it should be written at all.
For companies and investors, the signal is clear: the next wave of productivity gains won't come from eliminating all technical roles. It'll come from restructuring teams around AI-augmented senior talent and rebuilding entry pathways that teach AI fluency from day one. The anxiety Das describes won't resolve itself. It requires new onramps, new training models, and honest conversations about which roles survive and which don't. The companies that figure this out first will attract talent that everyone else is quietly laying off.