AI is useful as today. I, for example, I'm finishing a project where the company I work for wanted to extract some data from bills and documentation. The thing

AI is useful as today. I, for example, I'm finishing a project where the company I work for wanted to extract some data from bills and documentation. The thing is, in my later company I did a project with OCR and deep learning and it was quite useful. But this time, since we don't really need to tabulate the data, just extract some info here and there, what we did is to attach an LLM after the OCR extraction with an API. So right now we are getting a big improvement and flexible queries of information we are required to extract from bills and documentation, with an easy solution and not that heavy or costly. And I'm taking into notes that usually what we are talking about is llms as AI. But in the other hand, my last project for this company was an energy demand and allocation time series with auto-training in case it detects a huge chunk of error in the last statistics stracted from the model compared to the reality. This project was complex but it impacted the company positively with millions of euros in earnings. So, yeah. Statistics and programming works. Usually when we talk we do refere llms as AI, but there are other situations where just "classic" data science is useful and more than enough.