My AI Policy

TL;DR: I use LLMs as human team members and despise AI slop

• I use LLMs in the same way I work with humans on my team

• I use LLMs to fix my English

  • After 25+ years, sometimes I still prefer to write in Italian and getting it translated

• I review every single line of code / output from an LLM multiple times in the same way I review PRs from junior team members that are smart, but not trustworthy

• I need to be able to explain everything in my projects

  • Every line of code, every comment, every design decision
  • If I can't do that, I'm using AI improperly

• Results produced with AI should be of higher quality than results without it

  • AI is a tool to enhance my work, not replace my critical thinking

• I watch out for becoming dependent on AI

  • If I can't work without it (e.g., write code, think through problems, debug issues), that's a red flag
  • AI should augment my capabilities, not let me bypass learning or problem-solving

• I use AI for brainstorming, but exercise critical judgment

  • I always verify AI suggestions for accuracy
  • I watch out against hallucinations
  • I challenge outputs and make informed choices
  • I don't let the machine take over my decision-making or give out to AI-lazyiness

• I abhor AI slop and anything that looks, tastes, feels like that

  • AI;DR: If I feel something is AI generated, I lose interest

• My peers will review my work, so I am ready to justify my use of AI and defend every part of what I've created

• I expect people that work with me to follow the same or stricter rules