Just two prompts. The first asking for advice on addressing the issue (with a photo of the hat, and the second: "interested in the beeswax technique - can ou tell me more (and how it works)" (including my typo)
OK, welll.... I might have been tempted to ask at some point, "So how did you come up with that?;-) Possibly enlightening? I do too easily go down rabbit holes.
I always use Mike Caulfield's fantastic fact-checking prompt which you can find at his newsletter. I have a Fact Checker Project with those instructions built in - you may be interested in the response when I ran your issue through it:
Hey Stephen... I just ran across Mike's Deep Background protocol in a WSJ article. Is that what you are referring to? If your Fact Checker Project is something different, can you share? Thanks!
It's the same one but it runs a little sharper through Claude and the prompt is a little more elaborate. Same basic idea though. I've used it quite a bit and I think does a great job illustrating the process of fact checking as well.
Haha - love the depth of this! Of course, there does need to be something that kicks in that vigilance first ...
One of the things that worries me - and this goes t my work on epistemic vigilance - is that if you are asking questions of AI almost constantly, there's no space for doubt or critical reasoning (case in point would be someone who turns to the app on their phone repeatedly during conversations and meetings to get get AI-enhanced insights, but is not positioned in any way to evaluate or check the responses at the time)
My AI usage has shifted almost entirely from fact based stuff (and I use the fact checker when I do) to grey area analysis where there is no "right" answer and various points of view or the opportunity to converse with a very smart colleague simply isn't available. I guess I can understand these stories of people getting "duped" by AI, but I think that those people who are getting the most out of it instinctively know where to draw the line but I suppose that's a skill in and of itself. I really enjoyed your piece on your AI academic paper!
I guess it'd be a pretty deep hole to go down to figure out how it came up with that?
Or did you actually ask Claude a simple "show me the steps you used and sources that led to your hat advice."
Just two prompts. The first asking for advice on addressing the issue (with a photo of the hat, and the second: "interested in the beeswax technique - can ou tell me more (and how it works)" (including my typo)
OK, welll.... I might have been tempted to ask at some point, "So how did you come up with that?;-) Possibly enlightening? I do too easily go down rabbit holes.
I always use Mike Caulfield's fantastic fact-checking prompt which you can find at his newsletter. I have a Fact Checker Project with those instructions built in - you may be interested in the response when I ran your issue through it:
https://docs.google.com/document/d/1X6WQ1qzV5LrcfBVcX57pPth-RS_kT4asGL2xaOiTm48/edit?usp=sharing
Hey Stephen... I just ran across Mike's Deep Background protocol in a WSJ article. Is that what you are referring to? If your Fact Checker Project is something different, can you share? Thanks!
It's the same one but it runs a little sharper through Claude and the prompt is a little more elaborate. Same basic idea though. I've used it quite a bit and I think does a great job illustrating the process of fact checking as well.
Haha - love the depth of this! Of course, there does need to be something that kicks in that vigilance first ...
One of the things that worries me - and this goes t my work on epistemic vigilance - is that if you are asking questions of AI almost constantly, there's no space for doubt or critical reasoning (case in point would be someone who turns to the app on their phone repeatedly during conversations and meetings to get get AI-enhanced insights, but is not positioned in any way to evaluate or check the responses at the time)
My AI usage has shifted almost entirely from fact based stuff (and I use the fact checker when I do) to grey area analysis where there is no "right" answer and various points of view or the opportunity to converse with a very smart colleague simply isn't available. I guess I can understand these stories of people getting "duped" by AI, but I think that those people who are getting the most out of it instinctively know where to draw the line but I suppose that's a skill in and of itself. I really enjoyed your piece on your AI academic paper!