Really like the concept of posting reviews of recent research and I think the format you use here is promising. Here are a few thoughts:
1) I would urge that you skim recent publications and pick out the ones that seem most promising or useful, instead of choosing what you write about at random. I suspect your RA could be prompted to help with this. 2) I found your "response" more useful than the bullets and wonder if you combined the "Key Insights" with the response you might have a better form for the reviews...a little more prose than chopped up bullets? More "co-intelligence" than artificial intelligence. 3) Similarly, I think your frames at the end of "connections" and "applications" would be a strong way to tie together an analysis of two or three recent research papers. 4) I don't think the numerical rating brings much to the party
Take the suggestions for what they are worth to you, and no more. I intend them to be an endorsement of how you are "leaning" toward your next post. My primary feedback is this is a useful service to your fellow explorers in the wilds of generative AI and education. More please!
Thanks for such a thoughtful and insightful response! They're great and I appreciate you taking the time!
Yes--I agree about less random and more precise. That would be the idea going forward--just wanted to experiment first to do it as proof of concept, knowing it would be iterated along the way.
"More co-intelligence and artificial intelligence"--I like that and see what you mean. I think I would probably reduce the sections as well (e.g. not including this posts the keywords--I want them for whatever I build with this but not needed for the posts).
You're right about the numbers--at least in the high level summary. In the individual summaries there's more details that help to ground them. I will be curious to see if they actually fluctuate at all and provide detail over the long run (e.g. what does a 4 out of 10 look like). I think they have the potential to be useful later on in the larger collection. For instance, if I want to return to an article that I want to consider for reference, recommendation, or use in teaching, they might be insightful as a first glance.
Really like the concept of posting reviews of recent research and I think the format you use here is promising. Here are a few thoughts:
1) I would urge that you skim recent publications and pick out the ones that seem most promising or useful, instead of choosing what you write about at random. I suspect your RA could be prompted to help with this. 2) I found your "response" more useful than the bullets and wonder if you combined the "Key Insights" with the response you might have a better form for the reviews...a little more prose than chopped up bullets? More "co-intelligence" than artificial intelligence. 3) Similarly, I think your frames at the end of "connections" and "applications" would be a strong way to tie together an analysis of two or three recent research papers. 4) I don't think the numerical rating brings much to the party
Take the suggestions for what they are worth to you, and no more. I intend them to be an endorsement of how you are "leaning" toward your next post. My primary feedback is this is a useful service to your fellow explorers in the wilds of generative AI and education. More please!
Hi Rob,
Thanks for such a thoughtful and insightful response! They're great and I appreciate you taking the time!
Yes--I agree about less random and more precise. That would be the idea going forward--just wanted to experiment first to do it as proof of concept, knowing it would be iterated along the way.
"More co-intelligence and artificial intelligence"--I like that and see what you mean. I think I would probably reduce the sections as well (e.g. not including this posts the keywords--I want them for whatever I build with this but not needed for the posts).
You're right about the numbers--at least in the high level summary. In the individual summaries there's more details that help to ground them. I will be curious to see if they actually fluctuate at all and provide detail over the long run (e.g. what does a 4 out of 10 look like). I think they have the potential to be useful later on in the larger collection. For instance, if I want to return to an article that I want to consider for reference, recommendation, or use in teaching, they might be insightful as a first glance.
Thank you again!