Is AI a Silver Bullet?
September 3, 2024
Like previous attempts to find a silver bullet such as 4GLs, it does not seem likely that LLMs as a software authoring tool will be succeed. In particular, like 4GLs they suffer from the problem that because software spends most of its life in maintenance, the cost of change makes most improvements to the cost to author irrelevant; in fact, some rapid authoring techniques make maintenance harder and increase the lifetime cost of software.
Source: Is AI a Silver Bullet? — Ian Cooper – Staccato Signals
The term ‘silver bullet’ and in particular Fred Brook’s “The mythical man month” from which it is taken (spoiler alert there are no silver bullets) will be familiar to almost anyone who has studied computer science of software engineering.
The topic of the impact of AI generated code on the profession and practice of software engineering is one clearly occupying quite a few folk’s minds right now.
Steve Yegge recently ruminated on the impact of these technologies on junior developers (with relatively bleak conclusions).
Forrest Brazeal went further, speculating that AI is “about to kill, pretty much every single modifier we want to put in front of the word “developer.””
Ian Cooper is more sanguine–he traces the history of software development tools (a real blast from the past for me, going back to the 1980s)–CASE tools, 4GLs (4th generation languages, higher level, more ‘human like’ languages than 3GLs like C or Java), modelling languages like UML, DSLs (domain specific languages). All held out the promise of more productive programmers, easier to learn programming languages, all largely faded into obscurity (and specific niches where they are valuable).
Generative AI, Cooper argues will be similar–generative AI will have a role to play, but the software engineering landscape may emerge not drastically transformed from it was before.
