I recently wrote a blog post by first using Simon Buckingham Shum’s QReframer Chatbot prompt. While I had some ideas for a blog post, they weren’t fully developed and there were multiple possible angles. I didn’t want to start writing until I’d first really honed a single idea. I wanted a punchy, pithy and practical post rather than a rambling post resembling a Frankenstein’s monster of half-formed ideas. Of course, the editorial team would do a great job in providing me with appropriate feedback if I did write a mash-up using a mixed bag of malaphors but as no one says – why ‘burn that bridge when we get to it’ if we could just not?
The specs
- I used Microsoft Copilot in Web mode (not Work mode) so that Copilot could draw from anything on the internet that could help. I lazily couldn’t be bothered doing a Google search myself.
- I spent fewer than 10 minutes but am sure I saved myself at least an hour or more of time. Certainly, I progressed my thinking efficiently rather than sporadically spending a few minutes here and there over an unnecessarily drawn-out period causing a feeling akin to the princess and the pea – small but noticeable annoyance of an unfinished to-do list item.
- Once I’d crystallised and honed the single angle for which I had the most enthusiasm and energy, I wrote the blog post myself.
Recent chats
While in my original blog post I was warning against the dangers leveraging GenAI’s efficiency at the sake of deep learning, I’m writing this blog post to summarise conversations I’ve been having with colleagues on judgement making skills and GenAI. Unrelatedly, we’ve also been discussing how best to peel a mango and when the rain will stop 😊.
We feel that as more experienced professionals we’re already able to write effectively in a variety of genres for different audiences and when using GenAI we can quickly spot what needs editing, or at times completely delete the writing long with any evidence of the embarrassing prompt we tried. For us, the use of GenAI is not coming at the expense of learning how to write or accomplish any other day to day task.
And while I asserted that for students, some learning should still be completed manually before/while using GenAI, I’m now wondering how often I should try to organise my thoughts manually lest that skill go the way of navigating using a paper map, while the car driver yells at you that they’ve missed yet another turn in a city where hook turns baffle many. Moving on from that memory…how often do you think certain tasks should be completed slowly, manually and mindfully to maintain the muscle? Weekly? Monthly? Quarterly?
You can read my Copilot chat transcript here (PDF download).