I recently (2 days ago) noticed i enjoy it a lot more if i use a super fast model with low reasoning (gpt 6 luna low effort + fast mode). That way, i can still have hands on the entire time and dont have to wait around for 20minutes for it to finish up and have made a bunch of decisions without me.
Also, LLMs are first and foremost excellent at reading ultra fast. Makes it excellent for summarizing and re-representing modules of your code
That's a great idea, I will try that too. The LLMs are so capable, but the slowness annoys me. I always end up working on a few things at once and always doing rounds of reading its output, giving new orders, and going back to Reddit/HN/YouTube while it's doing its thing for the next 5 minutes.
I’ve also found that fast cheap models are more sustainable because they make it easier to keep technical debt under control. Partially because I’m keeping sustained attention in one place. And partially because small fast models with low reasoning don’t seem to love technical debt either and will (unintentionally) start to give you negative feedback signals when you’re letting things get messy.
I've had bad experiences with gpt 6 luna, but 5.6 luna worked real nicely with e.g React, you can do a back n forth and it'll work like a teammate vs just a "done!" where you then have to clean everything up afterwards
Yeah, I've come to this realization as well. It's like the inverse of that classic interruption comic. The waiting makes me lose the context.
Do you use luna to plan as well, or is it just code? I am interesting in understanding the process you follow.
My workflow is the most common one - use something like Opus to help create a spec document (I decide the specs and then get the LLM to ask me questions to harden it), then generate a plan, and then implement. I use Opus the whole time. Sonnet sometimes if it is a simple task.
humlex · · focus · HN ↗
Also, LLMs are first and foremost excellent at reading ultra fast. Makes it excellent for summarizing and re-representing modules of your code
badestrand · · focus · HN ↗
coffeefirst · · focus · HN ↗
Frenetic multitasking is for suckers.
bunderbunder · · focus · HN ↗
r_lee · · focus · HN ↗
Kiro · · focus · HN ↗
nibbleyou · · focus · HN ↗
My workflow is the most common one - use something like Opus to help create a spec document (I decide the specs and then get the LLM to ask me questions to harden it), then generate a plan, and then implement. I use Opus the whole time. Sonnet sometimes if it is a simple task.