AI summary by TwinMind
I can't believe this is a real statistic...
Theo - t3․gg
25m 58s
Stanford Research Findings
- Stanford interviewed 50,000 engineers from 100+ companies
- 9.5% of engineers are "ghost engineers" doing less than 10% of median engineer work
- Definition: virtually no work output, often working multiple jobs simultaneously
- Research used AI algorithm + 10 human expert panel to evaluate code commits
- 85% accuracy in identifying ghost engineers
- Published academic paper with ongoing research
Remote vs Office Work Performance Disparities
- 14% of remote engineers are ghost workers
- 9% of hybrid workers are ghost workers
- 6% of office workers are ghost workers
- Remote work enables easier avoidance of accountability
- Best performers (5x engineers) tend to be remote
- Worst performers also concentrated in remote work
- Creates bimodal distribution affecting remote work perception
Commit Activity Analysis
- 58% of ghost engineers make <3 commits per month (not 58% of all engineers)
- Of remaining 42% ghost engineers: make trivial changes (1 character edits)
- Commit counts flawed metric but reveals inactivity patterns
- Some ghost engineers do literally nothing for months
Economic Impact
- $90 billion annually wasted on ghost engineers globally
- Individual company savings potential:
- Cisco, Intel, IBM could save billions
- Combined market cap impact: $465 billion
- Zero performance decrease expected from layoffs
- Conservative global estimate: 6.5% rate = $90B waste
Hiring and Management Implications
- High-skill hiring ranges favor remote workers
- Average-skill hiring ranges favor office workers
- Risk profile: office hiring safer for average talent acquisition
- Elite engineers predominantly work remote, creating bias in remote work advocacy
- Companies hiring top talent see different remote work outcomes
Cultural and Team Dynamics
- Ghost engineers create negative team culture
- Productive engineers get demotivated by ghost colleagues
- "Malaise trap" - good engineers can fall into low-productivity patterns
- Some ghost engineers recoverable with proper motivation/environment
- Ghost engineers actively resist high performers to avoid exposure
Security and Operational Risks
- Ghost engineers create security vulnerabilities
- Higher hack susceptibility due to inattention
- On-call rotation problems when ghost engineers assigned
- Untested, garbage code contributions
- Access control issues with disengaged employees
Industry Examples and Patterns
- Elon Musk's Twitter layoffs (80%) started trend
- Companies realizing massive workforce reductions possible
- Quiet quitting phenomenon evolution
- Overemployment trend (multiple simultaneous jobs)
- FAANG companies particularly affected
Measurement Challenges
- Commit counts poor productivity indicator
- Hours logged unreliable metric
- True impact vs. perceived productivity disconnect
- Need better evaluation methods for engineering work
- Billing by results vs. hours debate
Startup Advantage
- 100% of startup engineers actually work
- Startups better positioned to compete against ghost-heavy corporations
- User understanding correlation with productivity
- Innovation blocked by ghost engineer presence
