Conversation

Notices

1. Embed this notice
   FeralRobots (feralrobots@mastodon.social)'s status on Tuesday, 26-Nov-2024 21:39:49 JST FeralRobots

   • Dare Obasanjo

   @GrahamDowns @carnage4life
   The problem arises when less-experienced developers (or worse, people with no dev experience at all) are encouraged to use ChatGPT/Copilot for answers. They have no way of assessing quality.
   That's the vision.
   That's why it all needs to be rejected. Not moderated: Rejected.

   • Embed this notice
     FeralRobots (feralrobots@mastodon.social)'s status on Tuesday, 26-Nov-2024 21:39:48 JST FeralRobots

     in reply to

     • Dare Obasanjo
     • Jon PENNYCOOK

     @jonpsp @GrahamDowns @carnage4life
     I see it as all part of the same assumption-set.

     I started out in software training. People invested lots of money in training staff. Now, nobody does. They just assume everyone already knows it.

     What happens is everyone looks it up with a search engine. Up to 2022-ish they could count on finding MOL-valid instructions.

     When all the search engines are LLM-driven chatbots searching an LLM-generated web, they'll no longer find those answers.

   • Embed this notice
     Jon PENNYCOOK (jonpsp@mstdn.social)'s status on Tuesday, 26-Nov-2024 21:39:49 JST Jon PENNYCOOK

     in reply to

     • Dare Obasanjo

     @FeralRobots @GrahamDowns @carnage4life or worse - when management decide they can employ people of lower skill level and not provide training because they think that AI will save them from having to pay for people who understand anything

     Rich Felker repeated this.