I feel acknowledged in my approach working with AI: much the same as I coach new colleagues on their first day on the job after orientation and onboarding. Give a simple task that should be in their range of qualifications. One task at a time, explain or show relevant detail without confusing them, read their face for understanding (never ask “do you understand?”), when comfortable, let them roll with it, and review the task results carefully (see why I never ask?)
Unlike new colleagues who gain competence, and after a while understand “the way we work here”, AI for now remains forever on its first day on the job. It teaches me that I have to grow and gain competence in using AI, practice makes my day. This article helps me see better where and how.
The other day I asked it to give me a SWOT analysis of its own capabilities. I tend to trust AI most when it explains itself, it should have such data, no? But then this may be the domain hardest to verify.
That said, following advice to check Reddit for my use case of local LLMs, I went there for the first time in decades and ran across this post, about 18h old. For me, resonates with this article. I am not qualified to assess how plausible it is.
My AI dev prompt playbook that actually works (saves me 10+ hrs/week)
Thank you, what a bouquet of insights.
TL;DR:
Everyone is a beginner - at something.
#AI is a beginner- at almost everything.
I feel acknowledged in my approach working with AI: much the same as I coach new colleagues on their first day on the job after orientation and onboarding. Give a simple task that should be in their range of qualifications. One task at a time, explain or show relevant detail without confusing them, read their face for understanding (never ask “do you understand?”), when comfortable, let them roll with it, and review the task results carefully (see why I never ask?)
Unlike new colleagues who gain competence, and after a while understand “the way we work here”, AI for now remains forever on its first day on the job. It teaches me that I have to grow and gain competence in using AI, practice makes my day. This article helps me see better where and how.
The other day I asked it to give me a SWOT analysis of its own capabilities. I tend to trust AI most when it explains itself, it should have such data, no? But then this may be the domain hardest to verify.
That said, following advice to check Reddit for my use case of local LLMs, I went there for the first time in decades and ran across this post, about 18h old. For me, resonates with this article. I am not qualified to assess how plausible it is.
My AI dev prompt playbook that actually works (saves me 10+ hrs/week)
https://www.reddit.com/r/LocalLLaMA/comments/1k8hob9/my_ai_dev_prompt_playbook_that_actually_works/?utm_source=share&utm_medium=web3x&utm_name=web3xcss&utm_term=1&utm_content=share_button