I'm reaching out to the community to see what the most up-to-date packages are for implementing Lasso, RF, and NN in R.
Lasso
As far as I know, lars has been replaced by glmnet for lasso and ridge regression. I think glmnet is the most cutting-edge package. http://cran.r-project.org/web/packages/glmnet/index.html
Random Forest
Until now I've used the package Breiman and Cutler's randomforest. http://cran.r-project.org/web/packages/randomForest/index.html
I believe there is a new package called party by Hothorn, Hornik, Strobl, and Zeileis, where cforest() implements Breiman's random forests. I haven't used it yet. http://cran.r-project.org/web/packages/party/index.html
Neural Nets
For neural nets, Guenther's neuralnet package was just released on Aug. 29, 2013. Although I haven't used it to speak to whether it's currently considered as the best choice for NN. http://cran.r-project.org/web/packages/neuralnet/neuralnet.pdf