Does anyone else here have a fear of missing out because of not working in AI/machine learning? It seems like people working in these fields are the modern-day rockstars.
Does anyone else here have a fear of missing out because of not working in AI/machine learning...
Leo Lewis
Brayden King
no one gives a fuck about machine learning. It's all a giant meme . The general population wouldn't be able to name 1 single person involved in machine learning other than maybe Elon Musk but he doesn't even do that work himself
Oliver Nguyen
>AI/machine learning
missing out? lmao. what are you missing out on exactly? it's a complete fucking farce.
>It seems like people working in these fields are the modern-day rockstars.
they're not. they're just computer programmers with over-inflated egos and self-importance that think their extremely gay python programming skills will build them a totally autonomous machine like the terminator or some other extremely gay fantasy.
Luke Richardson
>fear of missing out on buzzwords
lmao OP
Aiden Sullivan
No, I have a fear of working ON AI/machine learning. Can you imagine? Your whole professional life being based on coping mechanisms - that NP =/= P, that there is no mathematical algorithm to every problem, that you can always do just a bit better, just a bit better. Knowing the asymptotically best ML solution to a problem would be orders of magnitude slower than an 'ideal algorithm' that you just haven't found.
Urghhh
Evan White
I would actually get into it but literally every fucking tensorflow,pytorch,keras,whateverthefuck course is
>10 hours of talking
>0 hours of projects
Does anyone know any project-oriented courses for this shit?
John Thompson
>wanting to be a rockstar
Oliver Walker
the number of underages on this board...
Juan Evans
I work in AI/ML. Honestly unless you're doing world class research, it's pretty easy stuff to implement for any given problem. Hell, a large majority of use cases don't even need deep learning. Linear regression can solve so many problems, and random forests outperform state of the art deep learning methods on tons of real world problem sets.
It's also stupid easy to learn the basics. The real crux of the problem is that ML is just a tool, and in almost every case, requires knowledge of the application domain. Want to do ML for finance? You need to know finance. Want to do image processing and recognition? You need to know computer vision techniques and terminology.
It's not some magic bullet that will solve all problems, but when used sensibly can be used to turn raw data into useful function approximators, that's all.
Alexander Foster
Naw, man. AI is just so much witchcraft. There haven't really been big successes. Everything is still purpose-built and domain-centric.
Grayson Davis
>Does anyone know any project-oriented courses for this shit?
Just look at the top performing kernels on Kaggle and learn from those, it won't get more practical than that. You will also quickly come to the realization that most of the 'solutions' abuse dirty tricks to play the data instead of actually solving the problem in a general sense.
Michael King
>random forests outperform state of the art deep learning methods on tons of real world problem sets
Gradient boosting methods have been dominating nearly every Kaggle competition for a few years, i.e. xgboost, lightgbm etc.
Oliver Martin
so...
tfw no bara bot bf?
Easton Thompson
idk, as data analyst I'd say it's very difficult to find actual problems that need ML to solve.
Most of the time intelligent rule based algorithms are more practical.
Isaiah Smith
I program out of necessity, not for vanity. I make the tools that others and myself want to use.
Liam Sanders
>I program out of necessity, not for vanity. I make the tools that others and myself want to use.
Samuel Brooks
This is true. Gradient boosting machines rule kaggle since 2015.
Nathaniel Reed
>missing out because of not working in AI/machine learning?
im not missing anything because im too brainlet to learn it properly.
also most of Jow Forums do you know what it means to create neural network from scratch?
Connor Ward
No? Stop listening to a bunch of idiot journalists who are so ignorant they put high schoolers to shame.
They aren't rockstars at all, and the only benefit to machine learning is making it easier to copy someone else's visual, linguistic, or auditory style. It's a complete waste of energy for people who create original content. So obviously, journalists who are only concerned about style are very very concerned about machine learning.
Justin Ramirez
Its just a meme user. Kinda like how cloud and IOT were buzzwords a few years back. The hype will die off and those entities whose needs require it will use it. What do you currently do? I'm sure you are fine buddy
This. All the boomer partners and execs are trying to cash in on the latest trends by implementing AI/ML into their companies or selling it to others.
Brayden Ward
Not really.
AI will never be created by logic and reason. Pic related.
Sure they will produce a chess champion, go champion, but they will never create a Rumba that actually cleans your damn floor properly.
If it is created, it will be by accident or a cybernetics abomination. The most interesting theory of intelligence is that it is a pattern seeking imperfect loop.