4:54 Ackchually it's not a DAG ☝️🤓 The A stands for acyclic but allow to build cyclic graphs is the reason why they created langgraph. If you don't need cycles in your prompt chain I would suggest you to use langchain. You can build cyclic graphs with langchain as well if you have to.
Great video! Thanks for explaining these concepts. I had a quick question: do you think there's a clear distinction between AI Agents and Agentic AI? From my understanding, tools like Langchain, AutoGPT, and CrewAI are primarily designed for creating Agentic AI applications, which focus on natural language understanding and generation, while AI Agents are a more general concept involving autonomous entities that may not necessarily rely on language models. I was also wondering if you see these tools as mainly facilitating the development of individual Agentic AI applications rather than multi-agent systems. I'd love to hear your thoughts on this :) Thanks again for the insightful content!
I think you can use Flowise to create agentic type of behavior using almost no code. Having general understanding on how to code is worth learning though
Interesting review of Langraph, AutoGen, and Crew AI. Each framework seems to have unique strengths for different use cases. Have you explored KaibanJS? It is a JavaScript framework for building and managing multi-agent systems with an intuitive Kanban-style workflow. It might offer an interesting perspective for developers focused on JavaScript or those looking for a highly visual way to manage agentic workflows
I have a simple question: Can AI agents take action on my computer? For instance, can they delete a directory or rename a folder? I worry about the potential consequences if an AI agent mistakenly deletes a crucial directory on my machine.
The Agents usually operate with the "tools" they have been given. So unless you have explicitly created a function for accessing your directories with permissions to delete then It shouldn't happen. Also most frequently they are run in their private environments such as docker so if things go FUBAR then it should be limited to that docker instance. At least that's what I've experienced. Very good point though.
Hi, thanks for clearing this up. I want to build software using AI agents only And I'm not talking about a workflow but actual software. . Do you have experience with this?
3 AI agents , In Depth in 8 minutes. I dont think so. More like a cursory glance Edit: I swear the title said, In Depth... My bad. I'll leave this here and let others revel in my shame.
4:54 Ackchually it's not a DAG ☝️🤓
The A stands for acyclic but allow to build cyclic graphs is the reason why they created langgraph. If you don't need cycles in your prompt chain I would suggest you to use langchain.
You can build cyclic graphs with langchain as well if you have to.
Great video! Thanks for explaining these concepts. I had a quick question: do you think there's a clear distinction between AI Agents and Agentic AI? From my understanding, tools like Langchain, AutoGPT, and CrewAI are primarily designed for creating Agentic AI applications, which focus on natural language understanding and generation, while AI Agents are a more general concept involving autonomous entities that may not necessarily rely on language models.
I was also wondering if you see these tools as mainly facilitating the development of individual Agentic AI applications rather than multi-agent systems. I'd love to hear your thoughts on this :)
Thanks again for the insightful content!
how do you do this without coding?
I think you can use Flowise to create agentic type of behavior using almost no code. Having general understanding on how to code is worth learning though
which open source model do you recomend for ASR?
Interesting review of Langraph, AutoGen, and Crew AI. Each framework seems to have unique strengths for different use cases. Have you explored KaibanJS? It is a JavaScript framework for building and managing multi-agent systems with an intuitive Kanban-style workflow. It might offer an interesting perspective for developers focused on JavaScript or those looking for a highly visual way to manage agentic workflows
How about different is between the MetaGPT and ChatDev? And 3 framework here
I have a simple question: Can AI agents take action on my computer? For instance, can they delete a directory or rename a folder? I worry about the potential consequences if an AI agent mistakenly deletes a crucial directory on my machine.
The Agents usually operate with the "tools" they have been given. So unless you have explicitly created a function for accessing your directories with permissions to delete then It shouldn't happen. Also most frequently they are run in their private environments such as docker so if things go FUBAR then it should be limited to that docker instance. At least that's what I've experienced. Very good point though.
It lamens terms, yes 😅
This video is pretty good made, should have more views
This video is 🔥🔥🔥🔥
Good stuff! Love seeing new quality AI channels!
Thank you 🙏🏻 next video is dropping tomorrow, you might like it even more
Interesting comparison of top AI agent frameworks. For seamless AI orchestration, consider SymthOS. #AI #SymthOS #TechReview #Innovation
I appreciate the AI's ability to learn from my behavior, but it would be great if there was a way to see exactly what data it’s using.
Hi, thanks for clearing this up. I want to build software using AI agents only And I'm not talking about a workflow but actual software. . Do you have experience with this?
What about Devika or OpenDevin?
I would have to test that
Helpful. thx
And flowise?
Subscribing! Very good insights
CrewAI is collecting user data, that makes me uncomfortable
Didn't know about that. Very good point. How about Flowise/Langraph?
It's also open source so you can just host yourself
helpful, i love this
Reading others reviews and converting them into a video
"run out of your openai budget very fast" - then run open local models
Basically a script that executes scripts that executes scripts
@brettmiddleton5013 this is how all of the coding works lol
yes but with new and hip buzzwords
3 AI agents , In Depth in 8 minutes. I dont think so. More like a cursory glance
Edit: I swear the title said, In Depth... My bad. I'll leave this here and let others revel in my shame.
More like a cliff notes and certain insight that I found very valuable when i started working with them.
The words deep dive were used at around 1:29 so please feel vindicated 😊
I got flashbanged by subtitles
“You’re sitting too close to the 📺 lol. I’ll tone them down for next time!”
Great stuff, hope you check out our new multi-agents and AI teams release! :)
music, music, why, why? why the music? it ads no information and distracts and annoys