Yeah, ChatGPT, you’ve pretty much got the patter down. Thanks for regurgitating what I assume is an amalgam of assorted templates you found online.
I’m warming to the current generation of LLM AIs being a positively useful tool for setting a “what is the obvious, standard bullshit” baseline for a given topic. I’m writing a letter of recommendation; what can I do that an automated BS generator can’t? That clarifies my purpose.
I asked my students to consider that question (https://hachyderm.io/@inthehands/109479808455388578), and now I’m considering it myself and finding it useful.
In the case of my letter of recommendation, what can I provide that the algorithm can’t? Actual information about the student. Not just judgements and platitudes, but specifics.
The hollow familiarity of GPT’s output above might prompt some reflection around the role of letters of recommendation. Might, or at least ought to.