What are you bleeding money on today?

I use different AI vendors for purposes involving identifying semantic meaning across varied bodies of text, classification, and most importantly summarization. I guess it’s a good thing that the industry is rapidly developing models that make these tasks more manageable from a financial perspective, but I’ve definitely caught myself losing more than I would’ve liked by using an incorrect model for a particular task. Another place I’ve been caught up that I think also doesn’t get enough attention is not using the MORE expensive options to more quickly conclude some fairly complex tasks. I feel like especially for outfits where there’s not as many resources to spread around there’s always this vision to try to find the cheapest way to do something, not thinking that maybe spending a bit more up front will actually cost less. Last place I’ve noticed I’ve been bleeding money is user feedback. We definitely have a good user feedback pipeline but sometimes that can actually be a bad thing when you balance giving a small subset of users what they want while not alienating the broader user base. Early on we definitely implemented a lot of features that we ultimately had to rollback to stay competitive (and not burn through capital) and it was painful, but I think that’s a story quite a few teams on PH can relate too!