Feedback on The AI Guidance and Practice section

[sparkslabs]

The AI Guidance and Practice section (netlify preview) goes live during w/c 7/9 - please provide feedback here. Existing comments from the PR which weren't addressed before the PR was merged will also be copied here.

There may be hidden assumptions, things you disagree with. If you have questions - even if you can self-answer them, please raise them - others will have them and they need addressing.

The pages are deliberately tagged as "Beta Guidance" - explicitly to be able to incorporate comments!

[sparkslabs]

Comment received (based on attendees at an event not related to this, but relevant)

"some confusion about skills, loops and harnesses. we really need to clarify that for RSEs"

[sparkslabs]

Comment recieved:

There is no references section here - but there was a recent podcast about this book - https://areamethod.com/the-human-edge-book/ - https://peopleandprojectspodcast.com/ppp-518-why-better-ai-prompts-arent-enough-for-better-decisions-with-author-cheryl-strauss-einhorn/ - it's not about coding per se - but it mirrors some of what the page says - i.e. human agency and thinking first - e.g. "Start from the work, not the tool" & "Using AI well depends on judging what it produces. That judgement does not appear automatically as intensity rises"

[sparkslabs]

Comment received: (paraphrased)

• There needs to be guidance on AI skills - despite (and because) it being a fast moving field.
• This should ideally include: (think PRIMM)
  • What are they? Why would I used one? (etc)
  • How to use others? (considerations, solutions, etc - including costs, portability, testing etc)
  • How to modify others? (similar consideraions)
  • How to make new ones? (similar considerations)