Does your past habit with editor-based AI tools predict whether you'll keep using a new coding agent, or do your first sessions matter more?

This explores whether a developer's history with IDE-based tools predicts whether they keep using a new AI coding assistant. The short answer is that the corpus has no study that tests this link directly. What it does have points somewhere more interesting: who you're connected to and what you do in your first sessions seem to matter more than your background. The clearest evidence comes from Microsoft's rollout of Claude Code and Copilot CLI Do social networks drive adoption of new coding tools?. There, engineers' social ties, especially to peers one or two levels away on the org chart, predicted first use better than career stage or tenure did. Retention followed a related pattern: it tracked what engineers actually did with the tool, not who they were. Prior tool use is closer to 'who you are' than 'what you do,' so this finding gives reason to doubt that it would predict much on its own.

A second line of work sharpens the point. Researchers mined coding-agent conversations for usage traits and found that these interaction patterns explained coding outcomes beyond prior achievement Can conversation patterns predict coding outcomes better than prior skill?. The catch is that the traits weren't stable or transferable. They didn't behave like a skill someone carries from tool to tool. If even within-tool habits don't carry over reliably, it would be surprising if skill with an older IDE assistant cleanly predicted commitment to a newer, differently shaped one.

Experience can also cut the other way. In a randomized trial, experienced open-source developers working in codebases they knew well were 19% slower with early-2025 AI tools, even though they expected a 24% speedup Do AI coding tools actually speed up experienced developers?. Deep familiarity with your own environment can make an assistant's help feel worth less. Stickiness may also have little to do with real productivity. Fluency illusions and unclear credit for who did the work can make people feel more capable than they are How do AI tools trick users into overestimating their own skills?, so a developer may stick with a tool because it feels productive.

The gap is real. To answer the question properly, you'd need a study that follows the same developers from IDE autocomplete into agentic tools and measures who stays. The corpus doesn't have one. The useful takeaway is that the best available evidence treats adoption as social and behavioral, not as a property of the individual. Your colleagues, and how your early sessions go, are better bets for predicting stickiness than your tool history.