Conversation

Notices

1. Embed this notice
   jbowles (jbowles@infosec.exchange)'s status on Friday, 28-Mar-2025 00:42:29 JST jbowles

   it's funny there's so much talk around replacing code writers with automation.

   if you've used any of the "enterprise" AI tools to any degree you'll notice one common thing, no matter the domain nor the model:

   they do really well with formulaic data. Things that are within the "common knowledge" sphere of a particular domain... but they struggle with highly specialized data.

   Middle management is all about formula. It seems very much in reach to create a middle manager bot that asks you about timelines, helps plan the next cycle of work, helps estimate work load, reviews performance metrics, and communicates with stakeholders.

   I have some ideas about why the focus is on replacing engineers instead of managers but i'll hold off.

   • GreenSkyOverMe (Monika) repeated this.
   • Embed this notice
     jbowles (jbowles@infosec.exchange)'s status on Friday, 28-Mar-2025 00:42:51 JST jbowles

     in reply to

     • Jeremy Kahn
     • Donald Ball

     @trochee @donaldball don't get me wrong, i'm not advocating for middleManagerBot, just calling out that it seems like all the media attention is on replacing engineers, which seems off when you look at what LLMs can actually do

   • Embed this notice
     Jeremy Kahn (trochee@dair-community.social)'s status on Friday, 28-Mar-2025 00:42:52 JST Jeremy Kahn

     in reply to

     • Donald Ball

     @donaldball

     Preach it

     @jbowles

     GreenSkyOverMe (Monika) repeated this.
   • Embed this notice
     Donald Ball (donaldball@triangletoot.party)'s status on Friday, 28-Mar-2025 00:42:53 JST Donald Ball

     in reply to

     @jbowles While "reviews performance metrics" is very much in the bailiwick of the engineering manager as constructed by big tech nowadays, I strongly question how accurate and also how useful those metrics actually are. Not only are the data exceedingly coarse and readily manipulable, I hold they do an extremely poor job of capturing the business value performed by engineers working at highly variable tasks.

     They measure things that are easy to measure, not necessarily valuable to observe.

   • Embed this notice
     Jeremy Kahn (trochee@dair-community.social)'s status on Friday, 28-Mar-2025 00:43:15 JST Jeremy Kahn

     in reply to

     @jbowles

     TBH senior executive memos, inbound and outbound are also highly formulaic in the case of all but the most idiosyncratic weirdos (everything ever publicly announced or internal memoed from OR TO Jack Walsh, Eric Schmidt, Sundar Pichai, Sam Altman, any of the Facebook execs, the Walton family, etc etc)

     And I think I see a related(?) conclusion — upper mgt has never seen anything that _didn't_ look like AI slop, so *they* see the potential for the LLM to replace middle mgt too