Agentic Engineering for the Amateur Coder: An Extraordinary Skill for Ordinary People?

Alex Reppel; Mark Lycett; Patrick Kirchhoff

Agentic Engineering for the Amateur Coder: An Extraordinary Skill for Ordinary People?

Alex Reppel
, Mark Lycett
, Patrick Kirchhoff

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Advances in AI coding promise to enable people without programming skills to create working software. Popularized as vibe coding, this involves developing software through conversation with AI, without needing to write—or even read—the resulting code. In agentic engineering, users move beyond conversations and towards supervising multiple AI agents. While anecdotal accounts demonstrate its potential for non-programmers, a systematic exploration is lacking from the perspective of those experienced in realizing small software projects (often for ideation and/or personal use) yet without formal training. We call them amateur coders: People able to produce plausible software but lacking the expertise to verify what they are building. Drawing on a practice-led exploration and informed by three industry workshops, we articulate a framework for the amateur coder; responding to the thesis resulting from our research that while understanding code is no
longer necessary, judging whether it is good does not (yet) automate.

Original language	English
Title of host publication	Proceedings of the 60th Hawaii International Conference on System Sciences
Publication status	Accepted/In press - Jan 2027

Keywords

ai-assisted software development
vibe coding
agentic engineering
amateur coder

Cite this

@inproceedings{b55ce9a83316426a9686d2e225aa312e,

title = "Agentic Engineering for the Amateur Coder: An Extraordinary Skill for Ordinary People?",

abstract = "Advances in AI coding promise to enable people without programming skills to create working software. Popularized as vibe coding, this involves developing software through conversation with AI, without needing to write{\textemdash}or even read{\textemdash}the resulting code. In agentic engineering, users move beyond conversations and towards supervising multiple AI agents. While anecdotal accounts demonstrate its potential for non-programmers, a systematic exploration is lacking from the perspective of those experienced in realizing small software projects (often for ideation and/or personal use) yet without formal training. We call them amateur coders: People able to produce plausible software but lacking the expertise to verify what they are building. Drawing on a practice-led exploration and informed by three industry workshops, we articulate a framework for the amateur coder; responding to the thesis resulting from our research that while understanding code is no longer necessary, judging whether it is good does not (yet) automate.",

keywords = "ai-assisted software development, vibe coding, agentic engineering, amateur coder",

author = "Alex Reppel and Mark Lycett and Patrick Kirchhoff",

year = "2027",

month = jan,

language = "English",

booktitle = "Proceedings of the 60th Hawaii International Conference on System Sciences",

}

TY - GEN

T1 - Agentic Engineering for the Amateur Coder

T2 - An Extraordinary Skill for Ordinary People?

AU - Reppel, Alex

AU - Lycett, Mark

AU - Kirchhoff, Patrick

PY - 2027/1

Y1 - 2027/1

N2 - Advances in AI coding promise to enable people without programming skills to create working software. Popularized as vibe coding, this involves developing software through conversation with AI, without needing to write—or even read—the resulting code. In agentic engineering, users move beyond conversations and towards supervising multiple AI agents. While anecdotal accounts demonstrate its potential for non-programmers, a systematic exploration is lacking from the perspective of those experienced in realizing small software projects (often for ideation and/or personal use) yet without formal training. We call them amateur coders: People able to produce plausible software but lacking the expertise to verify what they are building. Drawing on a practice-led exploration and informed by three industry workshops, we articulate a framework for the amateur coder; responding to the thesis resulting from our research that while understanding code is no longer necessary, judging whether it is good does not (yet) automate.

AB - Advances in AI coding promise to enable people without programming skills to create working software. Popularized as vibe coding, this involves developing software through conversation with AI, without needing to write—or even read—the resulting code. In agentic engineering, users move beyond conversations and towards supervising multiple AI agents. While anecdotal accounts demonstrate its potential for non-programmers, a systematic exploration is lacking from the perspective of those experienced in realizing small software projects (often for ideation and/or personal use) yet without formal training. We call them amateur coders: People able to produce plausible software but lacking the expertise to verify what they are building. Drawing on a practice-led exploration and informed by three industry workshops, we articulate a framework for the amateur coder; responding to the thesis resulting from our research that while understanding code is no longer necessary, judging whether it is good does not (yet) automate.

KW - ai-assisted software development

KW - vibe coding

KW - agentic engineering

KW - amateur coder

M3 - Conference contribution

BT - Proceedings of the 60th Hawaii International Conference on System Sciences

ER -