Why AI Generates Different Code Every Run: Convention Drift and the Borrowed Architecture

Why does AI generate different code each time I run the same prompt? Because LLM inference is nondeterministic at the systems level, and the agent re-infers your conventions every session instead of loading them. The pattern has a name: convention drift — the agent borrows a convention from its training data instead of yours; same prompt, different session, a different convention each time.

The research most people haven’t seen

Researchers ran the same coding prompt repeatedly and measured functionally-identical output. On a competitive-programming benchmark, the same prompt produced zero matching outputs 75.76% of the time (Ouyang et al., ACM TOSEM 2024, arXiv:2308.02828).

And it’s not the temperature setting. Sampling a model 1,000 times at temperature 0 produced 80 distinct outputs (Thinking Machines, 2025). Nondeterminism is a property of batched GPU inference and floating-point non-associativity — not a dial you forgot to turn.

Why “prompt it better” doesn’t close it

Because the variance isn’t lexical. Prompts differing only in trivial formatting — separators, capitalization — produced swings of up to 76 accuracy points on the same task (Sclar et al., ICLR 2024, arXiv:2310.11324). The surface form, not the meaning, moves the output. You can’t word your way out of a different architecture per run.

Convention drift is one of three forms

The Borrowed Architecture has three forms: convention drift (above), assumption propagation (a reasonable-in-isolation assumption that passes tests and fails in prod against an older decision), and gate erosion (your review threshold migrating down without a decision). All three are decisions in your code that you never made. The diagnosis behind all of them is the Fluency Trap.

What closes convention drift

Persistent context — a tiered, curated core that loads your conventions every session, so the agent stops re-inferring them. Not a 200-line dump (that’s a monolith); a curated Tier-1 core you build up from tested rules. (See: context tiering.)

See how persistent context is built: curiochat.ai/software-engineer