Ricky
@sixtel
- 19 Followers
- 67 Following
- 120 Posts
Ricky is Human and can make mistakes.
Mostly don't need to read the full post. The quoted (italics) statements are enough for a smart person to interpret on their own.
@donburnside I basically live in the GitHub Copilot CLI interface with it mostly set to Claude Opus 4.6, sometimes Claude Sonnet 4.6. I can choose my models and those are usually my choices (for code). And I'm provided with an all-you-can-eat AI buffet. Use as much as you want, as often as you want.
@donburnside I cheap out for my personal use. It's the Gemini CLI with a lightweight Pro subscription, using the plan mode for most features.
@donburnside Much of my work goes through the Copilot "plan" mode to build and refine requirements, acceptance, etc. before implementation. More complex features or work goes through the process of using GitHub Spec Kit to spec, plan, create tasks, and implement the features.
I basically live in the GitHub Copilot CLI interface with it mostly set to Claude Opus 4.6, sometimes Claude Sonnet 4.6. I can choose my models and those are usually my choices (for code). And I'm provided with an all-you-can-eat AI buffet. Use as much as you want, as often as you want.
@donburnside Thanks! Yes, as someone that uses AI (Copilot specifically) daily for 6+ hours a day doing software development, the changes in the last year have been crazy. I use Gemini for my personal projects so I'm exposed to different LLMs.
Gemini emitted a basic, sterile, scientific graph for my original request (I assumed it would), then my request to decorate the graph for social posting emitted this. So two iterations.
It's baseball season. I asked Gemini to create a graph showing the most complete games pitched by pitcher and decade. Here's what it gave me.
@Dienstag Thanks. He's been an outside friend of the family for quite a few years now. Sort of a garden buddy. And knows when it gets cold that I'll have a warm place waiting for him.
