1. I assume this training is for at least semi-technical people, showing how to install python and that to create a file you need to click icon is stating the obvious 2. Enable line wrap when showing code 4. Imho agents without LLMs are not like team that cannot communicate. Imho LLMs are not for communication, please update "analogies". I think it would be close to team without brains ;-) 5. Rethink how you show the code. We're adding lines without description what they do, then a video later we say oh, now we need to add search tool, but ou, we have it, and we have that too and we have that. So maybe it's better to add search tool only on that later video when we talk about it? 6. I can bet that not everyone has OpenAI and Serper accounts, you may want to make a remark about signing in OR creating 7. Have the content reviewed by someone next time, please 8. Another weird thing - first we add process = Process.sequential, in the code and only in the next video we talk about different process types. Sequence seems to be super random in this training. 9. We built code and now we're back to theory without running it? After creating code I would like to run it vs listening to another piece of theory. We describe tools AFTER we created one. 10. It's not easy to understand the speaker sometimes, the voice is going up and down, please use text to speech engine for the voice over 11. I think the sample assumes some OpenAI account with $$, I get error "You exceeded your current quota, please check your plan and billing details" - that's not a good experience. 12. Clearly state what will be the $ cost of running those examples 13. Output from the sample is providing information from 2023 as cutting edge.. that's a bit crappy example imho 14. Why do we repeat content between tasks? we talk about top 3 AI articles both to researcher and writer, shouldnt writer write blog based on the content that researcher provides? I think we're making writer too specific to what we know that researcher provides, while we should make it generic enough to handle other researchers... 15. Link Agent Swarms, Implementing Open Source LLMs with Crew AI points to generic https://docs.crewai.com/#connect-crewai-to-llms (I think the anchor is gone)

course doesn't load properly