- Offers a quick path to MVP and immediate usefulness to me.
- modular, foundational (can be abstracted and repurposed)
Thoughts/Ideas
- After implementing this as a GPT, I’m thinking:
- it would be ideal if it shared the context of prior evaluations so that it could hone the scoring over time
- it could also get better with the questioning over time as it notices trends in the kinds of things I’m evaluating
- Updates
- The impact. Is it impact for me or potential for society? probably need to focus it on the latter. Also, I need more probing help with this question.
- It shouldn’t jump straight to confidence. It needs work through the technical details, entertain some solution options and then narrow down.
- it got lost when I told it I wanted to focus on the technical solution.
- and it then just provided the rankings and score when I tried to get it back on track, instead of talking me through it.
- while estimating impact, it conflated this with ease.
- can the GPTs learn as they go? or can I only update them in edit mode?
- it gave a summary instead of a transcript and it couldn’t generate a file
- move the scores to the top
Ease, key (i’m not sure this factors correctly though given the times…)
- 5 is ≤ 1 day
- 4 is 2-3 days
- 3 is 4 days to a week
- 2 is 2 weeks to a month
- 1 is possibly multiple months