With lasso I can now feed the entire dataset to Stata and wait for significance asterisk to float and pick whichever I want.
Stata new lasso command saves reg life

ridge reg
Read Mullainathan and Spiess (2017) in Economic Perspectives before carelessly using LASSO for variable selection
why use LASSO for variable selection? there are better methods out there.
What are they?
lmao. so just a different regularization scheme

ridge reg
Read Mullainathan and Spiess (2017) in Economic Perspectives before carelessly using LASSO for variable selection
why use LASSO for variable selection? there are better methods out there.
What are they?
Ridge regression is LASSO lol
only difference is that coefficients can't ==0 and drop out formally, just get really really small 
ridge is not the same as LASSO (LASSO penalizes absolute value of parameters while ridge penalizes sum of squared values) but the concepts behind them are so similar that positing ridge as a fundamental alternative to LASSO is inappropriate
This. Ridge regression tends to excel in cases where there’s *“grouping” of regressors, i.e., a number of the regression coefficients that move in tandem during successive optimization steps. Ridge regression tends to shrink the “group” proportionately whereas Lasso doesn’t since it promotes setting individual regression coefficients to zero to reduce model size. In these cases, in general, purebred Lasso would not perform as well as Ridge regression. Lots of highdimensional problems have regressor grouping so this is not an academic problem (see, e.g., Biostatistics applications with thousands of regressors, for example).
*”grouping” is also known as multicollinearity