Why do ML developers all think they're hot shit?
What they do isn't that much more impressive than what regular programmers do. A good full stack developer is harder to find than someone who's done a couple of ML courses on Udemy.
Why do ML developers all think they're hot shit?
Jordan Wilson
Other urls found in this thread:
Caleb Murphy
because the math behind more recent algorithms is outside the grasp of software engineers
Leo Reyes
Bullshit. Reverse engineering is harder
Cooper Powell
'Domain Experts' who import * from sklearn are not ML developers, simply ML users.
Alexander Watson
Reverse engineering is outside the grasp of most software engineers
Elijah Peterson
There are a bunch of people who get into IT for cold hard cash with no real passion or target other than get fat stacks.
These people latch on to certain niches such as ML, cybersecurity, web dev and start to circlejerk with other posers into the same niche. They then form cliques to freeze out people with the real passion who can see what they are up to.
Jace James
I made a gan that produces pictures of cats. Where's my riches? I'm doing dicks next
Angel Brooks
link to that cat repository? im interested
Connor Jones
I'm not linking my private github on Jow Forums this site is banned in Australia
Nicholas Nelson
idk about source code but here's one cat picture generator
Julian Myers
Data 'scientist' here. There's nothing impressive about the programming one does as a data scientist, whatever skill I have is in marrying the business knowledge, statistics and the bare minimum of programming necessary to deliver a product that helps make data-driven decisions. I've seen a ton of retards who fail at least one of these criteria, but there's nothing inherently more difficult about it than software development it's just a slightly different skillset.
Zachary Cruz
OP is correct, the only difference is attitude, and that's because ML developers today made it right before the field got saturated. I'd argue a full stack developer requires more skill and knowledge than a data scientist, but less people have the required skills to be a data scientist, and thus for the next few years, ML will pay really well.
Thomas Myers
I feel like I've opened the pussynomicon.
Nicholas Edwards
Noah Allen
bullets don't sate my appetite, jon
Xavier Mitchell
>process 100,000,000 pictures of cats
>still can't tell apart a cat from a dog
ML devs are total fucking failures, let's face it.
Jordan Wright
They are ML package users not ML developers. The math behind the algorithms is not trivial at all and your run of the mill cs baby that barely passed calc 2 and linear algebra can't understand any of it
Samuel Bennett
As an ecconometrics fag, I absolutely hate using shit like R. Not because it is intelectually challenging, but because it just a precise bunch of syntax bullshit. It like french class, not math class. Why I always cringe when people say coding is math.I wish someone would put the effort into making Eviews or something more complex and customizable. Like an RPG maker for stats.
Caleb Howard
It's not really any worse than any other language in that respect. Plus coding a model in R/Python is like 50 lines including model validation and hyperparameter tuning. Not understanding the hate, desu.
James Bailey
>It's not really any worse than any other language in that respect.
Oh for sure. But I didn't study Econ/stats cause I love programming languages. my point is it is simply an annoying tool of the trade. Not a stimulating part, or where the skill/challenge comes in, it is not why we are paid what we are. Nobody learns R because of an intimate interest in programming. But because they fucking have to. It is like people I know who learned french to help their careers for the Canadian government. It is not because they love linguistics, or that speaking french is magically more difficult, import, interesting or useful than speaking English. It is because they needed to to learn the language for the job.
>Plus coding a model in R/Python is like 50 lines including model validation and hyperparameter tuning.
yeah but shit like Eviews is just a couple clicks. The hate is I have to and do programming when I am not a fucking programmer. And have to memorize a bunch of precise bullshit just do have the software do menial tasks
Cooper Russell
There are no-code solutions if you prefer like Orange, Rapid Miner, KNIME and Weka, though I haven't found one that I prefer to just writing a few lines of Python. I don't like R much either, and don't use it on my own projects.
Sebastian Ward
>other than get fat stacks
genuine question, how old are you?
William Cook
Is there significant demand for people with RE skills? I did some hobby stuff with it a few years ago but I've never put it on my resume since it was very much "learn-as-you-go" and it's technically in a legal gray area cause it was RE'ing copyrighted games (though I wasn't breaking the copy protection)
Dylan Rivera
implementing ML algorithms from scratch is good fun. using prebuilt libraries for them is boring