I have stumbled across these two questions and accepted answers:
(1) Do we need gradient descent to find the coefficients of a linear regression model?
(2) Why use gradient descent for linear regression, when a closed-form math solution is available?
The answers to both questions are contrary.
In the answer to question one it is stated that gradient descent is the worst algorithm for solving linear regression and should never be used.
On the other hand, in the answer to question two it is argued that gradient descent would be much more efficient for huge systems than using QR, e.g.
How do these questions relate to each other? Is one of the answers incorrect / outdated? Or did I miss the difference between the two questions?