LLMs can be exhausting

This is a restatement of the old wisdom that to safely use a tool you must be 10% smarter than it is." Or stated differently, you must be "ahead" of the tool (capable of accurately modeling and predicting the outcome), not "behind" (only reacting). TDD is kind of an outgrowth of it. I've lived by the wisdom, but admit that for me there is a lot of fun in the act of verifying hypotheses in the course of development, even in the "test case gap" when you're writing the lines of code that don't make a difference in terms of making a long term test case go from red to green, or doing other exploratory work where the totality of behavior is not well charted. Those times are the best. "Moodily scowling at the computer screen again," has been a status update from chilluns on what I'm doing more times than I like to admit.

[dead]

Classic coding was the process of incrementally saying "Ah, I'm getting it!" -- as your compile your code and it works better each time, you get a little dopamine hit from "solving" the puzzle. This creates states where time can pass with great alacrity as we enter these little dopamine induced trances we call "flow", which we all experience.

AI is not that, it's a casino. Every time you put words into the prompt you're left with a cortisol spike as you hope the LLM lottery gives you a good answer. You get a little dopamine spike when it does, but it's not the same as when you do it yourself because it's punctuated by anxiety, which is addictive but draining. And I personally have never gotten into a state of LLM-induced "flow", but maybe others have and can explain that experience. But to me there's too much anxiety around the LLM from the randomness of what it produces.

Current state of my workflow: (your feedback is welcome, I'm an autodidact and have been self-teaching myself thus far) I also think my workflow addresses some of the pain points the OP mentioned.

If a problem is a continuation of the current or other chat, switch to it. If it is a new problem or sub-problem requiring something more extensive than a tiny refactor, a new chat is started.

From there,

Start in Ask mode. Ask about existing code I'm trying to modify. If I am interfacing with someone else's code, that is put in reach of the project and ask questions about how it produces certain results or what a function does. Ask the foundational 'bottom-up' questions, how does this work? what routine produces x? Call out specific external sources from the web if it contains relevant information, like an API. Iterate until I feel I have a grasp of what I can build with. Not only does this help me comprehend the boundaries in terms of existing capability and/or shortcomings, it seeds the context.

Move to Plan mode. Provide a robust problem statement and incorporate findings from the Ask into the problem statement, and what the desired output is. Throw in some guard rails to narrow the search path used by the LLM as it seeks the solution. Disqualify certain approaches if necessary. If the LLMs plan isn't aligned with my goals, or I remember that thing I skipped, I amend the plan. The plan prompt I typed is saved to a blank file in the text editor.

Implement.

Validate. If it works, great. Read the code and approve each change, usually I speed read this.

If it doesn't work, I tell the LLM the difference between the expected and actual result and instruct it to instrument the code to produce trace output. Then feed the trace output back into it with explanations of where the output doesn't match my expectations (often times revealing weaknesses in my problem statement). Sometimes when it is a corner case that is problematic, several iterations are required then screen for regressions. If I reach the point where I know I screwed up the planning prompt, I trash the changeds, then I revise the copypasta saved earlier and start a new Planning session.

Not sure if it's what you're talking about but I had a coworker trying to break into eSports and he talked a lot about the micro vs macro skills a game requires. Sounds like we all have an aimbot for programming so the competition has shifted hard towards the macro. That could definitely be tiring.