80% of developers say AI coding feels like dependence, not advantage
If you’re building with AI coding tools every day, the productivity gains are real. So is something else: a growing body of survey data suggesting those tools are blurring the line between productive and compulsive.
The Numbers
Coddy Tech, a programming training company, surveyed 305 developers and found that 80% say their AI use has felt more like dependence than an advantage. The specific behaviors that surfaced:
- 43% keep coding with AI after hours even when they meant to stop
- 32% have put off sleep to keep going
- 39% say AI tools have made it harder to switch off from work
On the career side, 74% said heavy AI use made it more likely they’d earn a raise or promotion. But 51% also said they were more likely to burn out.
The Trust Problem
The 2025 Stack Overflow Developer Survey adds another layer. AI tool adoption hit 80% of developers this year, but trust in AI accuracy dropped from 40% to just 29% year over year. Positive favorability toward AI tools fell from 72% to 60% over the same period.
The core frustration: 45% of respondents said AI answers were almost right, but not quite. The output arrives fast. Validating it still takes real work.
The Hidden Cost: Verification Debt
Quentin Rousseau, CTO and co-founder of incident report company Rootly, described the pull in a LinkedIn post: coding at 2:47 a.m. with no deadline, just watching Claude Code refactor a module and unable to stop. He described agentic coding as addictive in a specific way: when the agent gets things right, you get a dopamine hit; when it fails, you get an adrenaline rush. Rousseau said the experience affected his sleep and required medical attention.
The structural issue is what the article calls verification debt. AI output arrives quickly, but developers still need to confirm it’s correct, secure, maintainable, and appropriate for the specific codebase. If employers treat AI as a capacity multiplier rather than a toil reducer, the time saved on individual tasks gets absorbed by larger pull requests, more generated changes to inspect, and more operational risk to manage.
The Operator Takeaway
If you manage developers or are one, the question worth asking is not whether AI can write code. It can. The question is whether your team’s relationship with these tools is structured around reducing specific, bounded toil or around shipping more of everything, faster. The first has real upside. The second just moves the pressure somewhere harder to see.
I’m founder of BizStack at Cagri Sarigoz LLC and a passionate advocate for entrepreneurs.
With over 14 years in tech, marketing, and AI, including my role as Head of SEO at CitizenShipper and co-founder of TaleBot at Intale AI, I’m dedicated to sharing genuine, useful product insights and tips.
At BizStack, I aim to cut through the digital noise to provide clear, actionable advice.
And more than all else, I’m a father to a (always) little girl and a husband.
Contact me at [email protected] for assistance.
