One engineering leader's GitHub data shows AI agents raised the workload

Orestis Ioannou, VP of Technology at Bold.org, pulled their GitHub contribution counts from 2019 to now and found that 2026, with several coding agents running in parallel, is close to their busiest year: 2,102 contributions by 27 October, more than any full year from 2021 to 2025, with August 2026 the busiest month at 427. Their median weekday rose to 7 contributions, from 3 or 4 in each year from 2022 to 2025. They describe bursts of work after each new model or workflow, followed by fatigue, and say the strain comes from switching between agents that each need a review or a decision, not from harder tasks. The author notes contribution counts are only a rough activity signal, and says AI helped draft the post. On Hacker News (100 top-level comments) there was no consensus. The largest group said the premise was wrong: employers turn productivity gains into more output, not shorter hours, so nobody should have expected less work. Many others recognised the strain, describing constant context switching and review load across agents, and a few shared fixes such as a written daily task list, cutting side projects, or keeping agents off the phone. A smaller group of five said they now work less, including a founder and a self-employed commenter. At least seven commenters said the post itself read as AI-written. Sorted take: time saved by AI tends to fill up with new work unless you set limits, so decide what agents may start before they start it.