cq about what it's like now. i asked in today's interview how they use AI day-to-day and all the interviewer said is that they all have Claude Code and can use it however they want.

(http://www.autoadmit.com/thread.php?thread_id=5910860&forum_id=2#50178994)

Also sounds like this company is behind the curve in terms of AI adoption. Here's how coding is done at a typical SF startup that's fully bought into AI: you rarely write code yourself; you have one LLM plan out the code you need written, then another model writes the actual code based on that plan, then several other models review and iterate on that code for you, then multiple models review your PR and fix test failures, and then at least one model tries to fix production incidents, etc. You're essentially guiding multiple LLMs to iterate on what starts as AI slop into polished code; the latest LLMs are good enough at coding to do this.

(http://www.autoadmit.com/thread.php?thread_id=5910860&forum_id=2#50179046)