Grepper CEO reports tenfold speedup in complex coding tasks
Jeb Naus of Grepper states that AI coding agents have reduced the time for complex production engineering from months to days. He argues that this technology widens the productivity gap between senior and junior engineers.
Jeb Naus, CEO of the 10-person company Grepper, reports that AI coding agents have drastically reduced the time required for complex production engineering tasks. He cites a specific example of implementing 'auto-promotion' or 'shredding' in a data lake, a task involving metadata coordination and column-oriented Parquet and Iceberg files. Naus states that this work took him 10 days using AI agents, a task that would have taken a single engineer six months previously. His workflow utilizes the Windows Subsystem for Linux and Claude Code with the Opus 5.5 model. He uses command-line sub-agents to maintain persistent sessions and improve context efficiency, typically managing three or four concurrent session windows. Naus notes that the disparity in productivity between 'okay' and 'really good' engineers has multiplied by 10x. He argues that senior engineers are better at breaking down complex distributed system problems for AI execution, while weaker engineers see less benefit. Consequently, the team has shifted its code review focus toward tests and golden expectations to validate AI-generated code. Naus contends that AI amplifies the capabilities of strong engineers who can handle high complexity, creating a wider gap in engineering productivity.