This week I saw a post by a professor saying there’s no point in calling yourself an expert in (Gen)AI if you don’t know about optimization algorithms. To me, that’s the same as saying someone can’t be a spreadsheet expert without knowing linear algebra.
This kind of “academic arrogance” shows the distance that often exists between academia and practice. Of course, knowing more doesn’t hurt, so much so that I’m (re)studying all the concepts and the algorithms. Anything useful so far for my life as a product person? Very marginal.
Understanding how a vector search is done and what it takes to build some tools helps, and a lot! BUUUUUUT what helps the most is closely tied to the social sciences and less to optimization/statistics/probability.
I really liked Anthropic’s article on context engineering (https://lnkd.in/e-PxeWSd), and notice that it’s more important to understand how to split the work and pass the right instructions at the right time than the math behind it.
It’s the old 80/20 rule. The 80 will be solved with context engineering; the other 20 — I think it’s more like the other 1% — is solved with the math.
My opinion — for now.
Links
Originally written in Portuguese and published on LinkedIn on October 10, 2025. I translated it into English with the help of AI (Claude): it let me bring my LinkedIn posts to people who don’t read Portuguese without rewriting each one from scratch. I reviewed the translation to keep my original meaning and tone.