A large language model like ChatGPT allows users to write code using plain English.
While feasible to create code with just English input, limitations arise in producing novel, substantial, or iteratable projects.
Writing software with a large language model differs significantly from manual coding, impacting the future of software production.
Metaphors often used to define large language models include 'coach,' 'editor,' 'therapist,' 'collaborator,' and 'genie.'
When utilizing a large language model to write code, the experience can resemble interacting with a mediocre Genie.
Challenges arise when attempting to create new content or when stepping outside the realm of known programming concepts.
The process of producing code becomes more iterative and less contemplative with instant feedback from the Genie.
The Genie transforms the software production process, taking on roles like a Doctor, Mechanic, and Gambler, but diminishing the role of a Detective.
Engineering and artistry differ in their reactions to large language models due to historical practices and economic implications.
Concerns about the ethical aspects of generating code through large language models include potential copyright issues and the need for fair compensation.
The future of software engineering might reward taste and judgement over raw speed or intellect, emphasizing the importance of learning to write code creatively.