Insights on the New Meta DS/Product Analytics Interview Process

Mainly want to understand two parts:

1. Analytical Execution
What is the recent format like?
Is it more about A/B testing, experimental design, causal inference, metrics, product cases, or will they give a scenario and keep following up?

Specifically curious about:

Will they ask about sample size / power / significance
Randomization / selection bias / confounding — how deep will they go?
Is it more framework-based, or do they have you calculate problems on the spot
How far do they usually dig into one question?
Any noticeable changes lately

2. AI-Native Technical Skills
This is the part I'm most uncertain about right now.

Questions for those who recently went through it:

[/i] Is it still mainly SQL + Python?
Is the difficulty roughly comparable to the old Meta DS technical round?
How exactly is the AI assistant used in the interview?
Do they have you prompt the AI to write SQL / Python, then you check and debug yourself?
Does the interviewer focus more on the final answer, or on how you interact with the AI and verify the results?

• Will there be data analysis / debugging / code review type questions?
  

If anyone has gone through the Meta DS IC4 / IC5 interviews in the past few months, please share the general format, difficulty, and preparation focus. No need to reveal exact questions.

Especially requesting recent experiences on the new Analytical Execution + AI-Native Technical Skills sections. Thanks so much!