Elevator Pitch

• The author agrees with most critiques of LLMs (slop, ethics, trust erosion) but still uses them heavily as a thinking-and-quality amplifier—provided the human remains accountable.

Key Takeaways

• LLMs flood ecosystems with low-trust output (e.g., OSS PR spam), making “effort” and “authorship” hard to judge from the outside.
• The most durable value comes when humans keep ownership of thinking and credibility, using LLMs to refine, critique, and accelerate—not to replace judgment.
• Practical workflows (“/grill-me”, tight problem statements, adversarial subagents, local models) help counter LLM agreeableness and slop, but require the user to be able to evaluate quality.

Most Memorable Quotes

• “I almost agree with all of the LLM critics, yet I still use LLMs a lot.”
• “I think LLMs might have serious potential to kill OSS, if we don’t find ways to restore that trust.”
• “What distinguishes ‘AI slop’ from ‘good writing’ is whether a human has put thoughts behind it. And you cannot outsource thinking.”