Low Code ➡️ No Code ➡️ Vibe Code ➡️ Agentic Code ➡️ No Coders ➡️ Lots of Code ➡️ Oh My God So Much Shitty Code ➡️ Where Are the Coders? ➡️ This Coding Agent costs WHAT now? ➡️ Just Regular Code Again
@maxleibman I know I'm being pedantic all over your humor, but...
IME, the "low code" and "no code" movements were about model-based coding (analysis or design models), not related to the agentic LLM crap. AFAIK, there's been no agentic model creation, as that requires actual intelligence.

@lwriemen Pedantry is welcome, although I think we have a misunderstanding about the points I was making.

I didn’t say low code/no code was about LLMs; it is related because it was the previous wave of products and managerial excitement to claw some of software development away from expensive coders (also, I work in that space, and it IS rancid with AI crap now, too, although that wasn’t my point).

As far as agents go, sure, but that’s not what is meant here by agentic code. Coding “agents” are already running rampant. Phrased that way to be parallel with many of the other trend named here.

@maxleibman Hmmm...I always considered analysts to be more expensive than coders, but point taken. I'm an old Shlaer-Mellor OOA modeler, so that wagon seemed to get hitched to "no code" and "low code" as it tried to stay saleable. I never really grasped what the hard core "no code" and "low code" practitioners were doing, but knew it was tangentially related.
What is "AI" doing in those realms? In the S-M OOA realm some people were looking at it for the model compiler development.
@lwriemen The AI angle comes in a couple different ways—Microsoft (to give the example I’m most familiar with) incorporates Copilot into Power Platform and can be used to start or modify designs for workflows or apps. Doesn’t work well in my experience, but it’s there. They also have an “Apps Agent” in their frontier collection that will build one-off web apps backed by SharePoint data based on user prompts. From the other end, though, there are also platforms that use low-code style interfaces to help non-experts create custom chatbots or workflows/actions chatbots can control.