- Author
- Alex Martsinovich
- Published
- 29 September 2026
- Kind
- Personal opinion essay on Substack
- Original
- distantprovince.substack.com/p/the-four-horsemen-of-agentic-coding
- Subject
- Harnesses and agent loops
- Themes
- Skill under automation
What it is
Alex Martsinovich writes on Substack. The essay names slop, alienation, deskilling and team fallout. It aims to put into words what the author finds wrong with agentic coding.
What it says
- Martsinovich lists four problems of agentic coding: slop, alienation, deskilling and team fallout.
- The essay says LLM code has a smell that repels human readers.
- The essay says slop looks like a lasting trait of models, not a passing flaw.
- The essay says engineers who skim agent reports feel detached and care less.
- The essay calls deskilling real and doubts that learning AI tools is a skill.
- The essay says team chat has thinned because people ask agents instead of colleagues.
- Martsinovich offers no fix and says the problems have no solution in sight.
What evidence it gives
The essay offers argument and personal observation only. It gives no data, survey or measurement. It uses analogies such as woodworking and IKEA, and unsourced remarks about team chats.
Interests and limits
The essay is a personal view. The author says the effects can be felt in the bones. It names no study. The author also calls agentic coding undeniably useful. Only two replies were visible on the page, and nine more were not loaded.