Ok so, quick question—are u looking to master the heavy math theory or do u wanna practice writing actual code? I ask cuz:

* Brilliant.org: Best for visual neural network logic, but the sub is kinda pricey.
* Enki: Better for interactive Python and ML drills and more cost-effective.

Both are solid for train rides, but it basically depends on if u want formulas or functions!

AI is basically just smart patterns! Late to the party, but do you prefer coding or theory? Sololearn is amazing for free, while Mimo is better for Python drills. Just watch those subscriptions!

Any updates on this?

In my experience, trying to learn neural networks on a tiny screen sounds like a headache, but it’s actually totally doable if u pick the right app. I’m still kinda a beginner myself, but I’ve spent a lot of time looking at the technical specs of these platforms to see what actually works for a busy schedule.

Here is what I recommend for your situation:

* Brilliant.org Premium Subscription: This is highkey the best for the "why" behind AI. It uses interactive visual puzzles to teach things like neural networks and the math behind them without making u drown in dry textbooks. It is super polished and works great for 10-minute bursts.
* DataCamp: Learn Data Science Mobile App: This one is better if u wanna focus on Python and LLMs. They have "Daily Practice" sessions that are literally perfect for a train ride. The technical depth on their machine learning tracks is surprisingly good for a mobile UI.
* Sololearn: Learn to Code Mobile App: A decent option if u want something more casual. It is basically like Duolingo for coding. Good for basics, but maybe not deep enough for complex ML theory imo.

Basically, if u want to understand the math and logic, go with Brilliant. If u want to see how the Python code actually builds the models, DataCamp is the winner tbh.

Quick tip: Stick to one app at a time so u dont get overwhelmed. Consistency during those train rides is literally more important than how many hours u spend on a weekend.

So yeah, it really depends on your needs, but those are the gold standards for mobile right now. gl!

Jumping in here - the market is honestly flooded with AI apps right now, but there's a big gap in quality between the legacy education brands and the newer startup apps. I’m always a bit cautious about where the content actually comes from, you know? Half these apps are just riding the hype, so I’d look at the market from a reliability standpoint first before committing to a sub. 1. Are you leaning more toward a platform with a formal academic pedigree - basically something developed by a university or established tech firm - or are you okay with a gamified startup approach?
2. How much do you care about the long-term data privacy of these platforms versus just having a slick UI for your commute? Tbh, idk if the gamified ones really give you the depth you need for neural networks - it’s easy to make a pretty interface that skips the hard stuff. I’d be careful about what you’re actually paying for, especially with how fast these brands are popping up lately.

Regarding what #3 said about AI being smart patterns... honestly, I've tried many of these platforms over the years and it drives me crazy how predatory the industry has become. You're trying to learn a skill for your future and these companies just see a cash cow. It is such a scam how they lock the actually useful modules behind pro tiers that are nearly impossible to cancel.

• Most of these mobile apps are just shiny wrappers for basic documentation you can find for free, yet they charge a fortune.
• The interactive features are often just glorified multiple-choice quizzes that dont actually teach you how to think or build.
• Quality is going downhill fast because they are all rushing to capitalize on the hype without actually vetting their teaching materials.
• Half the time you're just paying for a UI that looks nice but has zero depth. It is really frustrating for someone just trying to use their commute time effectively. You spend more time navigating dark patterns and subscription traps than actually learning neural networks... honestly its ridiculous.

at Reply 8 - good point! I just caught up on this discussion. Most of the previous suggestions point toward Brilliant for theory and Sololearn or Enki for basic coding. However, you might want to consider hardware compatibility before committing. High-end interactive sims for neural networks can really drain the battery and lag on older devices like an Apple iPhone 11 64GB. If you are focused on the budget, I would suggest looking at the Coursera Mobile App. You can audit many AI courses from top universities for free. You get the high-quality lectures and readings without paying the subscription fee, provided you dont need the certificate. I would also recommend checking out the DataCamp Learn Data Science App for their free daily practice modules. It is a safer way to start than jumping into a recurring payment. TL;DR: Thread recommends Brilliant and Sololearn. My advice is to audit Coursera courses to save money and ensure your phone can handle the app performance before buying.

No way, I literally just dealt with this yesterday. Small world.

Adding my two cents as someone who prefers a DIY approach over paying for a curated course. Honestly, I found that the 'professional' mobile apps were a bit too high-level once I wanted to dive into actual transformer architectures or how gradient descent really functions in a production environment. My current setup is basically a mobile sandbox I built myself: * I use a local inference app to run quantized models directly on my device (great for testing how different context windows affect RAM).
* I keep a library of technical whitepapers in a mobile reader to study the math behind attention mechanisms.
* I run a terminal emulator to execute small Python scripts and test API calls during my commute. It took more effort to set up than just hitting 'subscribe' on a flashy platform, but you'll get a much better feel for industry standards this way. Troubleshooting your own mobile dev environment teaches you way more about how these systems actually tick (at least that's what worked for me). DIY is definitely the move if you wanna go deep tho.

I have been messing around with AI since long before it became such a massive buzzword, and honestly, my biggest breakthrough came when I stopped looking for a specific learning app and started treating my phone like a mini workstation instead. I remember being stuck on a long train ride a few years back, just trying to map out how a simple decision tree worked using a basic drawing tool I had on my device. It clicked way faster for me than any polished course ever did. Nowadays, my setup is pretty bare-bones. I mostly use a cloud-based editor that lets me run scripts directly in my mobile browser. It was a bit of a learning curve to get the interface right on such a small screen, but actually seeing the code fail and fixing it while waiting for my lunch taught me more about the logic than any quiz ever could. Tbh, the best thing I did was stop searching for the perfect app and just started building tiny, broken things during my commute. You learn real fast when youre the one responsible for fixing the bugs yourself.

Finally someone says it. Ive been thinking this for a while but wasnt sure.