Came across a good Twitter thread covering a range of Google course modules in aspects of AI from the fundamentals of LLMs through how Transformers work (GPT stands for Generative Pre-trained Transformer), The Attention Mechanism, and more.
It’s no light and casual skim-read, but most of the modules seem to be less than an hour, and this really is an investment in understanding how things actually work.
Definitely worth taking a look and powering through if you want a better understanding of the space.
Thanks for sharing!
I took and passed the first one. I found it helpful to get the context of Machine Learning and LLMs. I need to go through the rest of the courses. The first module is great background information, but nothing immediately practical.
I think, to a very real extent, the practical comes in afterwards, with our own projects.
This course is more about giving us a solid foundational understanding, upon which, we can build our own practical experience and experimentation. Even to the extent of using cloud services to train our own LLMs or Diffusion based AIs if we have the time and money for that endeavour. After all, that’s pretty much what OpenAI themselves did, though with millions of dollars to burn on their AI training runs.
Everything is going to move pretty fast, unless or until various authorities start to get involved and set laws and limitations. GPT5 may well be ready before the end of this year… Or it may be delayed, of course, and I’ll be amazed if Google and DeepMind aren’t aware of that and planning a big release of their own before that. They can’t allow OpenAI to keep building their lead (in public perception, which also affects share prices).