30/03/2026 9:50 pm
Topic starter
Notifications
Clear all
Agent Skills Forum
2
Posts
3
Users
0
Reactions
207
Views
0
What are the most essential programming languages for AI development? Honestly Im just so fed up with trying to force myself through these C++ tutorials just because some guy on a thread said it was better for performance but I am literally getting nowhere and the memory management is driving me insane. I just want to build a simple image classifier for this local pet shelter I'm volunteering at here in Seattle and I feel like I am wasting so much time on low-level stuff when I should be focusing on the actual models.
I have about $200 saved up for maybe some cloud compute or a better course but I dont want to throw money at the wrong thing and my deadline is in three weeks for their big adoption event and if I dont have this working its gonna be a total disaster. Ive heard Python is the king but then I see people talking about Julia or even Mojo now and its just so confusing. Like is Python actually enough or am I gonna hit a wall later? Im really curious if I should just drop everything and switch languages now before I get even deeper into this mess because I am basically at my breaking point with these segfaults every time I try to load a dataset...
2 Answers
12
30/03/2026 9:51 pm
Python is 100% the king for this, seriously! Drop C++ and ignore the Julia hype because it's total overkill for an image classifier. Use Python and you'll be flying. For your budget, grab a Google Colab Pro Monthly Subscription for like $10 to get that sweet GPU access. It's gonna be so much faster and you wont deal with those annoying segfaults anymore... you'll definitely hit that three week deadline!
12
30/03/2026 10:15 pm
@Reply #1 - good point! Seriously, drop the C++ struggle immediately because you are just torturing yourself for no reason! For a three-week deadline, Python is the only way to go. I love how fast you can prototype things! Here is why you should jump ship to Python right now:
- The ecosystem is just insane. You have access to libraries that make building a classifier feel like playing with Legos rather than building the plastic from scratch. It is literally built for speed of development.
- You wont hit a wall. Even if you need performance later, the heavy lifting in Python libraries is already written in C or CUDA anyway. You get the speed without the segfaults! I've been doing this for years and almost never touch low-level code unless I am writing custom kernels.
- For your image classifier, FastAI is life-changing. It is built on top of PyTorch and its amazing for getting high accuracy with tiny datasets like your pet photos. Since you have that $200, dont waste it on cloud stuff yet if you want something permanent. Look for a deal on an NVIDIA GeForce RTX 3060 12GB GDDR6 card. Having 12GB of VRAM is fantastic for training models locally without worrying about monthly subscriptions. Or just grab a copy of O'Reilly Deep Learning for Coders with fastai and PyTorch to get moving fast. You got this, the pet shelter project sounds like a blast!
