Conversation

Notices

1. Embed this notice
   Reg Braithwaite 🍓 (raganwald@social.bau-ha.us)'s status on Wednesday, 04-Jun-2025 21:43:55 JST Reg Braithwaite 🍓

   Thoughts from Eric Forman's basement:

   1. We'll know that generative AI is really here when the entire stack from LLM vendors all the way down is written by LLMs, and everyone who remembers the code that was originally written by humans is gone.

   👇🏽

   • Embed this notice
     Reg Braithwaite 🍓 (raganwald@social.bau-ha.us)'s status on Wednesday, 04-Jun-2025 21:43:51 JST Reg Braithwaite 🍓

     in reply to

     Eric Forman’s basement is a place where even the most banal observations provoke unstoppable giggles:

     —————

     "When nobody remembers how to read code, who will review the code written by AI?"

     _More AI, of course._

     "Quis custodiet ipsos custodes?"

     _You can't fool me, young man. It’s AI all the way down._

   • Embed this notice
     Reg Braithwaite 🍓 (raganwald@social.bau-ha.us)'s status on Wednesday, 04-Jun-2025 21:43:53 JST Reg Braithwaite 🍓

     in reply to

     👆🏽

     And finally:

     3. When people speak of “Nineteen Eighty-Four” as prophetic, they’ll be speaking about Ken Thompson’s work, not George Orwell’s.

     “To what extent should one trust a statement that a program is free of Trojan horses? Perhaps it is more important to trust the people who wrote the software.”

     —Ken Thompson, Reflections on Trusting Trust, Turing Award lecture, 1984

     https://www.cs.cmu.edu/~rdriley/487/papers/Thompson_1984_ReflectionsonTrustingTrust.pdf

     HistoPol (#HP) 🏴 🇺🇸 🏴 repeated this.
   • Embed this notice
     Reg Braithwaite 🍓 (raganwald@social.bau-ha.us)'s status on Wednesday, 04-Jun-2025 21:43:54 JST Reg Braithwaite 🍓

     in reply to

     👆🏽

     2. “The AI Stack” will not be an architectural chain, it will be a recursive pile of AI, where humans prompt AIs, which in turn do the programming by prompting other AIs, which in turn do the programming by prompting other AIs, which in turn...

     Because prompting AI will itself become so large and complex that AI is needed to prompt AI. And no, we won't recognize any but the human-most “prompts.”

     It’ll be AI all the way down.

     👇🏽

     HistoPol (#HP) 🏴 🇺🇸 🏴 repeated this.