80% of developers say AI coding feels like addiction, not advantage
AI coding tools are productive. They are also, according to a growing body of survey data, turning software development into an always-on loop that developers can’t step away from.
The addiction data
Coddy Tech surveyed 305 developers and found that 80% said their AI use felt more like dependence than an advantage. Specific behaviors tell the story: 43% kept coding with AI after hours even when they intended to stop, 32% put off sleep to keep going, and 39% said AI tools made it harder to switch off from work entirely.
Quentin Rousseau, CTO and co-founder of incident report company Rootly, described the pull in a LinkedIn post: it was 2:47 a.m., no deadline, no outage, just watching Claude Code refactor a module and unable to stop. His explanation: agentic coding delivers a dopamine hit when the agent succeeds and an adrenaline rush when it fails. Rousseau said he couldn’t sleep and had to seek medical help.
Trust is eroding while adoption climbs
The 2025 Stack Overflow Developer Survey adds a second pressure point. AI tool adoption reached 80% of developers, but trust in AI accuracy has dropped from 40% in prior years to just 29% this year. Positive favorability toward AI tools fell from 72% to 60% year over year. The trigger: 45% of respondents were frustrated by AI answers that were almost right but not quite, creating difficult debugging work rather than eliminating it.
The verification debt problem
The Coddy survey also found that 74% of developers believed heavy AI use made them more likely to earn a raise or promotion. But 51% said they were also more likely to burn out. That tension has a structural cause: AI output arrives fast, but developers still carry the full burden of understanding requirements, catching conflicts with existing architecture, testing edge cases, and managing security risks. The code ships quicker; the judgment work does not disappear.
When organizations treat AI as a capacity multiplier, the workload pressure compounds. More features expected, more tickets to close, more generated changes to review in the same number of hours. Time saved on individual tasks gets absorbed by larger pull requests, more dependencies to validate, and more operational risk to manage.
The operator takeaway
If you are a solo developer or running a small team, the risk here is not that AI tools are bad. It is that the feedback loop they create has no natural stopping point. The developers who report the healthiest relationship with these tools are using them to remove specific routine toil, not to accelerate every stage of the pipeline simultaneously.
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.
