My sentiment exactly. The premise of the article comes across a little naive because there are so many fine-tuned libraries for specialized hardware architectures that already do this computation very efficiently.
However, it did make me wonder what this might look like on a gpu-accelerated database engine that is designed to leverage the SIMD parallelism of GPGPU architectures.
Beyond using SQL/NoSQL databases for CRUD apps I am not a "database guy", so I'm not sure about the feasibility, but it would be interesting to see it implemented.
Your question "Has anyone converted stuff like gradient descent to set theory?" doesn't make sense from the perspective that gradient descent uses differential calculus to find min/max points of an objective function and differential calculus requires a lot of additional assumptions on top of set theory.
That being said, current deep learning libraries such as JAX and Pytorch use automatic differentiation to efficiently compute partial derivatives used for optimization algorithms such as gradient descent and it's not clear to me what the level of effort would be to convert that to something that could run efficiently in SQL?
OP's original point still stands. Real estate in America is designed to have many economic advantages for wealth accumulation and preservation. You might even go as far as saying the economic advantages of real estate are a byproduct of regulatory capture
This is an interesting observation. Some of the data scientists I work with have been using Kotlin to define "analytic grammars", and is the first example (in my limited Kotlin awareness) i've seen of Kotlin outside of Android development.
It seems that Roman Elizarov (Kotlin Project Lead) has identified the opportunity for a better language ecosystem to enter the data science space
There's obviously no guarantee in the long run, but Gabriel Weinberg (DDG founder) is arguably one of the most holistic tech founders when it comes to tackling socially impactful problems in tech probably due to his interdisciplinary background. He's detailed his approach in a variety of podcasts and blog posts over the years
There's decades of academic work in the form of "impossibility" results showing that handling byzantine fault tolerance in distributed consensus is alot more computationally expensive than omission/crash fault tolerance. The following (FLP impossibility result) being one of the most famous ones:
One of the contributors wrote a memoir of his time spent working on this project as a PhD student and gives an interesting perspective on the challenges encountered:
Elon is not the first to take on so many ventures at once. Richard Branson has done it with overwhelming positive results (some losses are bound to happen). It seems that after a certain point, an entrepreneur that is so inclined gets more effective at putting the right processes in place so that they're no longer a bottleneck at the management level.
>But now I learned with the minimal dataset I have (<5K samples), linear regression, SVM, or decision tress is the way to go.
Decision Trees are prone to overfitting and especially susceptible for small datasets. Random Forest is
a good substitute that's become standard practice.
https://www.dama.org/cpages/body-of-knowledge