@chrisalbon literally got called a classical ml dude

@hakim_elakhrass Right?!

@chrisalbon When will neoclassical ML come around?

@peterchng Scaled support vector machines!

@chrisalbon My apologies

@rasbt I’m not old I’m wise

@chrisalbon

@Mlondon83 It is so mean!

@chrisalbon We’re boomers now

@HarperSCarroll We're still cool!

@chrisalbon Folks who ignore them are missing out. Also, “classic ML” is a standard in ML breadth interviews: comparisons of training time, inference latency, memory footprint, etc.

@chrisalbon Anything before 2020 is referred to as BC (Before Corona).

@chrisalbon I think I mostly call it "traditional ML" but "classical" is fair too. I also was using sklearn (kmeans) yday & use it sklearn frequently. I think NNs came out and people just forgot traditional/classical ML exists and still works very well and very fast for real deployments.

@chrisalbon I’ve heard this numerous times, makes me feel weird.

@chrisalbon For the past few years Ive been calling it "traditional ML". But scikit learn has lots of goodies besides its models.

@chrisalbon I always call it "classic ML" because it is easy to explain that I am not talking about DL.

@chrisalbon I also use classical machine learning for algorithms pre neural network era

@chrisalbon Feels like still going to LAN parties.

@chrisalbon Better than "legacy"

@chrisalbon If it can reassure you, SVM just got me out of something an LLM would not properly classify. LLMs can do incredible stuff (like identify if a text is sarcastic or not) and yet sometime have a total block on dumb things.