There is a lot of genius to Tesla that is out in the open, but this unappreciated--the "experts" often focus on the wrong things. Way back when, one of the many advantages of the Panasonic 18650 was that it had no secrets--Panasonic had sold a gazillion of them and had good operational data on how they behaved, how they degraded etc. This greatly simplified the task of writing a BMS since the operational characteristics of the cells were so well understood. After the first few iterations of the pack, Tesla started to also accumulate mountains of their own data from their own cars which then informed subsequent design and engineering decisions.
No matter how talented your design and engineering team is, a 1.0 product is still a collection of unproven conjecture and that is where everyone else is. Eventually folks will close the gap as they accumulate their own data and expertise, but its going to take a couple of products cycles. Back to the original pack example, LG Chem cells have been around a couple of years, but operational data on them is still going to be limited (especially if manufacturers did not bother to instrument their cars), so writing a BMS to manage an LG pack is still going to involve a fair amount of rolling the dice.