The pitch was that AI would give us our time back, but I have never worked more. I use agents for all my engineering tasks and still read the code, yet this year I notice I am working more than ever and feeling heavier.
My week now runs several agents in parallel—one implementing a change, another investigating a production issue, a third handling a review, and a fourth doing research. Each agent asks for my input at times, forcing frequent context switches that are mentally draining.
GitHub contribution counts from January 2019 to today show the shape of my activity. The busiest year was 2,711 contributions in 2020, followed by 2,102 by October 27, 2026, close to that peak. Months like August 2026 recorded 427 contributions, while my lowest year, 2023, saw 1,189. The median weekday rose to 7 contributions in 2026, up from 3‑4 in earlier years.
The data reveal fewer zero‑activity days—18 through October 27—versus 40‑46 in prior years, indicating heavier, more continuous workload. The shift from fewer serious tasks to many concurrent agents makes the fatigue feel new, not just more intense. Overall, the pattern shows AI has amplified my cognitive load rather than reduced it.
Source: Hacker News · Summarized by HeadlinesBriefing