Google Colab

Pros: It is free or cheap and good for beginners. It offers a platform to start AI for those who don't have appropriate hardware in personal use.

Cons: It have been two days that I don't get any GPUs even though I paid for Pro. When I was a regular member I used to get those pretty consistently, maybe not fast and fancy, but at least I had something. Now I have nothing! I will not pay a penny any more and I don't case about quotas etc. You offer a deal, you hold to your promise. There was nothing about quotas in the offer that I went for!

Overall: Do not pay for subscription. They don't do their part of the deal, just collect your money... Use it while you are kicking off your ML coding skills and move on to personal hardware when ready.

Pros: Google Colab is absolutely great. It is so good, that some teachers of my otherwise Google-sceptical computer science department used to recommend it for projects in class as it basically does not need any initial setup and is completely free (the "basic" plan). One of the absolute highlights is the RISE-extension, which enables you to run snippets of code when in presentation mode, without having to switch windows. This really is a great way to present code.

Cons: You can't really make a virtual environment persist longer than for one session. Running the respective cell with all packages to be installed would take a few minutes at most though. The portability and ease of use of this web-based IDE makes up for this disadvantage.

Another disadvantage is the limitation of RAM and computing power but this is hardly reached if you are "just" learning and doing prototyping. If you would want for example to train a somewhat bigger model, it could become very unhandy, as you'd have to do a lot of checkpoints (because you keep getting interrupted by Google every few hours) and memory reallocation, but I guess it would not be impossible.

Another disadvantage is that resources are not guaranteed. Depending on the general load and on how much ressources you already used in the past, you can not be sure that you get computing power for your code for a longer time.

Nevertheless, being able to use it in its basic version for no money is extraordinary and a huge step towards democratization of, for example AI.

Overall: For small to medium-sized projects it is a fantastic Python-based IDE, which can easily be shared with others. For serious projects, you'd probably need a stable environment and guaranteed resources to work with.

Pros: Used to live up to their motto: one account to rule them all something something. Now it's 15 accounts to make things work.

Cons: Removed the ability to mount different drives to one Colab account thereby breaking entire university projects overnight with no warning. Beyond dumb.

Overall: Like sawing my nads off with a spoon.

Pros: It's online, it could be useful do do machine learning with high GPU computer.
It's linked to Google Drive.

Cons: - extremely slow
- lot of crashes
- not intuitive at all
- poorly documented
- bash and shell extremely horrible
- no support for python 2.7

Overall: Not easy to use bash on it, all command are limited, really slow.
It's not even 10% of what a random linux can achieve and that's sad because it could have so much potential with high power GPU

Pros: It is free and looks like a good opportunity, Google Colab.

Cons: The service, also if the paid version, always interrupts the training.

Overall: The service is always interrupted with no sense, it is ok, it is free, so you can't pretend nothing, but the project is based on IA and needs a really long training. I can understand the free service is interrupted after 12 hours, but I pay the service and there is nothing than free. The training is interrupted after 2 or 3 hours, if you don't stay in front of the the laptop, you never grow your model.
The paid version of the service promise more ram and the cut off after 24 hours but it is totally fake.
I am thinking to interrupt the payment, is easy find online solutions to have high memory for free just cheating the platform.
Why pay to have the same service for free?

Pros: Google Colab is a tool which helps data scientists collaborate online. It has a smooth desktop interface and has a lot of features which help in analytics and collaboration.

Cons: It is still in nascent stage and features are still being added. The mobile app is not as robust as well.

Overall: I have had a very short experience with Google Colab, but it seems like a great tool especially for independent researchers and freelancers.

Pros: • Integrates with Google Drive and saves all your work on the cloud, so it saves local space.
• Easy to run python code.
• Sharing and working on the same code with your team is easy. The collaborative tools are pretty efficient.
• Has many built in modules.
• No need for any kind of installation.

Cons: • Since you can't edit python codes directly from the drive, it's sometimes a pain to edit it locally, upload it, and then open it on the Colab notebook.
• Debugging can be confusing, especially since line numbers aren't included in code.

Overall: I like Google Colab mostly for its integration with Google Drive and since it is all in the cloud. But otherwise, it doesn't really compare to its competitors like Jupyter Notebook. Anyway, a great tool for developing code.