Uber released Ludwig, a deep learning toolbox where hardly any coding skill

uber.github.io

23 points·udayrddy··9 comments

I mentor a lot of young data scientists, and the advice I give them, over and over, is, "your ability to code as a data scientist is not what makes you employable." I use projects like this to make the point.

Not taking away the great help to the open source community. Apart from high tech and advanced AI, did uber team missed the basics ? I mean, half of the 1.3MB webpage load is being consumed by 3 images. That "screens-2800.png" which consumes 586KB, and it conveys nothing.

Is anyone else a little bothered that Uber uses machine learning for “identification of points of interest during conversations between driver-partners and riders”

Title gore, but super interesting project.

> Doesn't require coding

> First step is to run pip

How many actual non-coders are going to go beyond step 1?

How many actual non-coders are going to go beyond step 1?

Hard to say. Are we defining “non-coder” as a person who can neither use a command prompt nor code?

Your project managers, your functional QAs. Your manager's manager.

Outside SV, those people have never, and will never open a terminal.

Horse shit.

Plenty of my relatives don’t have hot SV titles, and yet have successfully navigated the console / prompts / command line.

So in general: Citation Needed.

Somehow your anecdote seems to be better than mine. I've only worked in two other countries, aside from the US, and 4 distinct industries, and, again, except in SV, no PM or functional QA wanted to even get close to a terminal. Closest thing would be Excel, and always going through the formula UI.

Reply on news.ycombinator.com