Table of Links
Abstract and 1 Introduction
2. Prior conceptualisations of intelligent assistance for programmers
3. A brief overview of large language models for code generation
4. Commercial programming tools that use large language models
5. Reliability, safety, and security implications of code-generating AI models
6. Usability and design studies of AI-assisted programming
7. Experience reports and 7.1. Writing effective prompts is hard
7.2. The activity of programming shifts towards checking and unfamiliar debugging
7.3. These tools are useful for boilerplate and code reuse
8. The inadequacy of existing metaphors for AI-assisted programming
8.1. AI assistance as search
8.2. AI assistance as compilation
8.3. AI assistance as pair programming
8.4. A distinct way of programming
9. Issues with application to end-user programming
9.1. Issue 1: Intent specification, problem decomposition and computational thinking
9.2. Issue 2: Code correctness, quality and (over)confidence
9.3. Issue 3: Code comprehension and maintenance
9.4. Issue 4: Consequences of automation in end-user programming
9.5. Issue 5: No code, and the dilemma of the direct answer
10. Conclusion
A. Experience report sources
References
9.4. Issue 4: Consequences of automation in end-user programming
In any AI system, we need to consider the consequences of automation. End-user programmers are known to turn to local experts or gardeners (end-user programmers with interest and expertise in programming who serve as gurus in the end-user programming environment) when they are unable to solve a part of the problem (Nardi, 1993; Sarkar & Gordon, 2018). Task-orientation tendencies combined with challenges of completing their tasks easily also leaves end-user programmers with limited attention for testing, or carefully learning what is going on with their programs. Assuming that LLMs and associated user experiences will improve in the coming years, making end-user programming faster with LLMs than without, it is tempting to wonder whether the programmer can be persuaded to invest the saved time and attention to aspects such as learning or testing their programs; if so, what would it take to influence behaviour changes?
Another question is in the role of such experts. We conjecture that LLMs or similar AI capabilities will soon be able to answer a sizeable fraction of questions that end-user programmers will go to local experts for. An open question therefore is how the ecosystem of end-user programmers in organizations will change in their roles, importance and specialities. For example, will gardeners take on the role of educating users on better taking advantage of AI? If so, how can we communicate the working of such AI systems to technophile users and early adopters, so they can enable others in the organization?
