mhoyeFeb 28
What if I could convince you that taking the same time to explain detailed requirements and carefully validate results with a junior colleague instead of a chatbot would not only give you two people who understood the code instead of zero, but if you do it a few times in a row you eventually get a senior colleague out of the deal for free.
Show threadÖlbaumMar 11
@mhoye I think there’s a similarity between the way people ignore the time they waste tweaking their prompts, and the way management doesn’t see the cost of having a two-hour meeting between eight people to discuss whether a $500 purchase is justified.
@maro
Show threadGlyphMar 11
@oscherler @mhoye @maro https://blog.glyph.im/2025/08/futzing-fraction.html is probably the most popular thing I've ever written and pretty much all it does is point out the fact that the null hypothesis for the economic utility of LLMs really *must* be that they obscure the cost of failed prompts but make the success of successful ones highly visible
The Futzing Fraction

At least some of your time with genAI will be spent just kind of… futzing with it.

Show threadGlyphMar 11
@oscherler @mhoye @maro sorry if you've seen me flog this particular post before, nobody hates that it continues to be so goddamn relevant more than me
Show threadÖlbaum
@glyph I had seen it, but forgot about it. It’s probably your post that made me realise how much time my colleagues spend tweaking the prompts for their chat bot to satisfy the teachers because it either gives w vague or too detailed answers. 🤦🏼‍♂️
@mhoye @maro
Mar 14 at 7:38pmWeb