Posts by kbmcgowan@lingo.lol
 (DIR) Post #APR6qejH69xUm2CLGC by kbmcgowan@lingo.lol
       2022-11-09T18:57:53Z
       
       1 likes, 0 repeats
       
       I admire the way duckduckgo keeps me safe from the prying eyes of trackers and marketers. But I wish it was less effective at keeping the web pages I want to read safe from the prying eyes of me.
       
 (DIR) Post #ASZg5WGnlT12rce6pE by kbmcgowan@lingo.lol
       2023-02-10T11:56:58Z
       
       0 likes, 1 repeats
       
       subscription service idea: box of the month for cats! there is nothing in the boxes, they've just been used to ship various interesting-smelling things.
       
 (DIR) Post #AV0usclv6IUeAw3o3s by kbmcgowan@lingo.lol
       2023-04-25T14:52:11Z
       
       0 likes, 0 repeats
       
       Inspired by @grammargirl; today is the day I'm going to finish writing this overview of theories of speech perception and what they have to say about social variation.
       
 (DIR) Post #AVSgmDlDDlfVvyexxg by kbmcgowan@lingo.lol
       2023-05-08T17:08:17Z
       
       0 likes, 0 repeats
       
       welp. I will never assign another online essay question.Fuck you, LLM researchers, and fuck you, lazy piece of shit cheating students.
       
 (DIR) Post #AVTp2RpY9fMHYvBcLQ by kbmcgowan@lingo.lol
       2023-05-09T12:38:16Z
       
       0 likes, 0 repeats
       
       @writerethink @grammargirl when you have 90 students replying to the same 3 prompts it becomes pretty obvious that you're seeing the same adjectives, low frequency nouns, and entire NPs over and over.  So I wrote a little thing to find long substring matches between two files and started comparing particularly-vague, formulaic answers to it and found entire sentences chatgpt had generated for me and for the student.
       
 (DIR) Post #AVTp2Svy3I5Oz7CFzE by kbmcgowan@lingo.lol
       2023-05-09T12:43:06Z
       
       0 likes, 0 repeats
       
       @writerethink @grammargirl just grading the answers on their merits, though?  without having to accuse anyone of cheating, many of the essays were just awful.  these LLMs might be able to fool someone who is trying to cheat on an exam, but I literally wrote the textbook for this particular class. When the AI bluffs that "the authors argue X in chapter 9", it's pretty cringe.