Frank Lauter
Frank is not wrong, and I have been doing similar things with ai (and spending some $$) - I do worry that we will eventually run out of developers who understand what the ai is generating enough to guide it correctly. AI will hit some limits eventually - the term data wall is often used, where it runs out if data to learn from (and there is the worry it will learn from bad data). Eventually human programmers will be extinct - what happens then ![]()
I have found that AI with Delphi works really well with Library code - I have manged to get some really good results.
The results with Application code is much more variable. Perhaps that is because I am reluctant to have it see my entire application codebase.
AI can’t think and never will be able to with LLMs, so can never be creative. 90% of my time is spent trying to work out what the problem is I am trying to solve (and these will be human process issues, mainly), so I spend so little time actually writing code I barely use AI. I also eschew AI mainly at code writing time because the very act of typing the code “burns” the meaning into my brain. This is how humans are “wired” to learn. AI doesn’t change that equation
I am very pessimistic that AI will really help us in the long run because the lack of true understanding of code bases will be so bad as to render the developers unable to fix bugs properly other than “patch over” with AI. AI will mean that technical debt will soar to levels unheard of in the coming years. The world will paint itself into a corner where humans won’t be able to maintain AI codebases. And where are you going to get the software engineers coming through
I just see a disaster happening and fortunately I work with a couple of small clients where I don’t have to be a part of that world
In the age of AI I reckon my cognitive skills have grown much faster than those using AI, because I do things the old-fashioned way, with my brain. If you want to be elite at anything (sport, music, etc), there is no shortcut. You just have to do the time (which in my case is mental exercise)
Cheers, Misha
