Just refuse to do any AI interview, and refuse early-stage coding exercises. Have the base salary range and the number of in-office days upfront. I add a 10 to 20% premium for each day in office per week.
I got tricked into doing this once, for Mercor. didn't have much else going on, and I'd never actually talked to an AI before. what struck me as very odd is that instead of being obsequious, the AI would ask me an involved question, I would start digging in, and it would very rudely interrupt me to ask something completely different.
after 1 30 minute session I'd had enough, and didn't complete the remaining sessions. the even nastier part is that for the next few weeks I would occasionally receive messages from Mercor thanking me for doing the interview, but saying that the format had been changed and it was necessary for me to go through the process again.
Just refuse to do any AI interview, and refuse early-stage coding exercises. Have the base salary range and the number of in-office days upfront. I add a 10 to 20% premium for each day in office per week.
made me sad reading this as a picture of the state of the world
I got tricked into doing this once, for Mercor. didn't have much else going on, and I'd never actually talked to an AI before. what struck me as very odd is that instead of being obsequious, the AI would ask me an involved question, I would start digging in, and it would very rudely interrupt me to ask something completely different.
after 1 30 minute session I'd had enough, and didn't complete the remaining sessions. the even nastier part is that for the next few weeks I would occasionally receive messages from Mercor thanking me for doing the interview, but saying that the format had been changed and it was necessary for me to go through the process again.
I suggest HN add “is lying to you” to the clickbait title filter.
LLM’s love that one weird trick to get hits.
Honestly, I think this should be a felony.