Google Colab

There's open weights models that you can use.

In our class, we use Google Colab because they have access to that and can get GPUs through that and things like that.

So it makes the experience much better.

But there's tons of, you know, available resources out there to kind of get started.

Now, this is something you can just do with a single prompt today, but back then you had to go and take 10 or 20 different photos of yourself at different angles, different lighting, usually just dug through your camera roll, got them.

Then you had to feed that into like a Python script that was either hosted on Colab.

And then the Colab thing was weird because it was a luck of the draw what GPU you got.

So you would like sign up and create an instance.

And then sometimes it would just be like, you got a 1080 TI.

And there's some evidence that they can be.

You know, Google has Colab, Claude has ClaudeScience.

So they're trying to build tools that do these things.

I have a hobby project called OpenScience.ai that's trying to do these things.

I recommend using a Python notebook.

And the easiest way to get started is Google Colab.

It's really easy to get set up.

Or Jupyter Notebooks, that's more difficult to get set up.

So once you've got your subscription on ChatGPT, you need to turn on Code Interpreter.

That's under beta features, as you can see on the screen, hopefully.

Get Google Colab.

If you're already on Google Workspace, it's free, so great.

Yeah, yeah.

Yeah, I was just gonna say that it's funny you brought that up, too, because, um, when I was in school, um, I think, I think it was like my second to last semester, my professor w- like I took an artificial intelligence class, and that was like, uh, 2021, and, and it, this was like he, he was showcasing OpenAI, and OpenAI was, was basically, um, led, giving you access to like Atari games, and then you, you could just, um, basically put it in your, uh, a Google, uh, Colab notebook or a-

Right

... or an Anaconda notebook, and you would just play like Pong or something like that in, like, in a, in a, in a, in a web interface.

And this is how logistics companies, you know, pack their trucks and they have supercomputers running all these knapsack problems, essentially.

So I began to do that and it taught me how to use Google Collab at first.

and then it taught me how to use Python, and it was giving me code, and I would test and run the code, and sometimes it would work, and sometimes it would come back with an error, and I'd put the error in JetGPT, and it would correct it, and went back and forth, and what felt like a know a collaborative experience as a regular person it felt like i was working towards an end goal and i just want to say that like this is how this is what the tool is designed for like there are developers who do use gpt i mean this isn't um you know i wasn't using it for some strange reason like this is how entire teams are being replaced by by large language models right because they can do the code they can correct the code and all this so i thought that i was just You know, there's a really strong there's really strong messaging coming from the corporations building these products.

Right.

It's all Python based.

You should also be familiar with, I would start with Google Colab.

You can use it for free.

And you're going to have to do like Jupiter Pi notebooks for a lot of your AIML coding.

The jobs tell you the exact areas that you need to learn, right? And then a lot of this stuff is open source.

I think all of Hugging Face models are open source, if I'm not mistaken.

So you can literally go and create your whole workspace on Google Colab for free.

You get to test on different GPUs.

And then I literally did like a multi-layered approach with Python and the Gemini API.

I run a lot of Python on Google Colab.

I feel like it's an extremely underrated tool.

I currently, like I'm speaking to, I have three Google Colab tabs open right next to our tab.

So I run break this down.

like break these massive documents down into facts.