We’re getting constant pressure from the management to use LLMs which, in our work, have so far only shown negative ROI. And the time and effort spent on them has been displacing some ML projects we were working on that actually looked like they’d give us some concrete, positive results.