Join my SKOOL community here: www.skool.com/the-ai-agent-5174/about It's a community for AI Agents and I can better answer questions. If you would like: 👉 developing these agents 👉 understanding workflows with them 👉 using CrewAI as the framework to accomplish this 👉 be a part of a community of like-minded individuals
A possible video for you. Compare cost of the same task with different agent frameworks. Use AgentOps to compare AutoGen, CrewAI, and something else to see which is most cost efficient and quickest.
This is a great idea! I have a few videos similar to this coming out! I enjoyed giving my honest thoughts in this video about it, whether good or bad. I have an overview of my thoughts on crewai vs. autogen coming soon
@@TylerReedAI My framework uses Directed Acyclic Graphs and JSON-defined agents that act as both agents and tools. A developer agent that can create new agents on the fly, if needed. example: what's the weather like in Detroit? The intent is an action that there is no agent for, so the system will ask, do you want to create this agent. It will create the agent, (prompting for any needed info) and then execute the agent and give you the weather. As a UI-based system, it communicates through natural language and executes tasks seamlessly. Designed to enhance the lives of persons with disabilities, a blind person can perform tasks via voice commands. Currently a diabetic patient manages her condition by integrating with her Dexcom glucose meter, tracking food intake, getting nutrition info, and calculating insulin doses. A few elderly people use it to manage household tasks like meal planning, chores, and shopping (Kroger API). It's highly modular and adaptable for almost any application. I may Open-Source it shortly. @ScottLahteine got a peak at the orchestrator code and how it flawlessly handles complex requests. This one exactly: User_input = "Retrieve the customer data for Alice Johnson using her first and last name, then obtain her contact information. Next, find another customer by her phone number '555-9876' and get her contact details. Check the availability of both Alice Johnson and the second customer, along with team members Bob Smith and Carol White, for a joint client onboarding meeting on December 5th at 11:00 AM. If all participants are available, schedule the meeting, add all four participants, and send personalized confirmation emails to Alice and the second customer with the event details, including the meeting agenda and a follow-up survey link."
😆😆 i don’t want to say it’s bad because it supposed to be educational, I’m just not sure why they put time and energy into it either. I don’t know, just my opinion
I helped fix the vector memory stream in LangChain about a year ago. The team was fantastic, and it’s an awesome framework. I’m just cautious about black-box solutions, as I prefer having transparency and control over the internal workings.
Join my SKOOL community here: www.skool.com/the-ai-agent-5174/about
It's a community for AI Agents and I can better answer questions.
If you would like:
👉 developing these agents
👉 understanding workflows with them
👉 using CrewAI as the framework to accomplish this
👉 be a part of a community of like-minded individuals
A possible video for you. Compare cost of the same task with different agent frameworks. Use AgentOps to compare AutoGen, CrewAI, and something else to see which is most cost efficient and quickest.
I second this suggestion! I also would really want to see the best framework, as well as which one you think is the best and why.
This is a great idea! I have a few videos similar to this coming out! I enjoyed giving my honest thoughts in this video about it, whether good or bad. I have an overview of my thoughts on crewai vs. autogen coming soon
@@TylerReedAI Looking forward to that
How did you get personal-shopper to work? It is missing a package and an issue said it does not exist.
It’s more like a barebones setup compared to other workflows, isn’t it?
Yeah because it’s supposed to be educational and that’s about it
Do you have a video or stated preferred AI stack you use to build Agentic apps?
That will be coming soon!
Have you used autogen 0.4? They say it was rewritten from the ground up.
I haven’t yet! I’m waiting just a bit to really dive into the re work
Hyped too much. There are like 500 lines of actual code. They committed logs from April 2024... which interns wrote this?
good question lol. I was kind of curious how much money or time they spent creating it
Time to release my framework open source. It’s way better than swarm. Great ending to your video, way to keep it real.
An agent that authors new agents as-needed and adds them to the router agent…. How quickly does this bootstrap a huge amount of functionality?
@ wait! How did you know this was a capability? Kris?
Thank you! What's your framework?
Oh Crap! the video was set to public. It wasn't meant to be. - REMOVED! 😲
@@TylerReedAI My framework uses Directed Acyclic Graphs and JSON-defined agents that act as both agents and tools. A developer agent that can create new agents on the fly, if needed. example: what's the weather like in Detroit? The intent is an action that there is no agent for, so the system will ask, do you want to create this agent. It will create the agent, (prompting for any needed info) and then execute the agent and give you the weather. As a UI-based system, it communicates through natural language and executes tasks seamlessly. Designed to enhance the lives of persons with disabilities, a blind person can perform tasks via voice commands. Currently a diabetic patient manages her condition by integrating with her Dexcom glucose meter, tracking food intake, getting nutrition info, and calculating insulin doses. A few elderly people use it to manage household tasks like meal planning, chores, and shopping (Kroger API). It's highly modular and adaptable for almost any application. I may Open-Source it shortly. @ScottLahteine got a peak at the orchestrator code and how it flawlessly handles complex requests. This one exactly: User_input = "Retrieve the customer data for Alice Johnson using her first and last name, then obtain her contact information. Next, find another customer by her phone number '555-9876' and get her contact details. Check the availability of both Alice Johnson and the second customer, along with team members Bob Smith and Carol White, for a joint client onboarding meeting on December 5th at 11:00 AM. If all participants are available, schedule the meeting, add all four participants, and send personalized confirmation emails to Alice and the second customer with the event details, including the meeting agenda and a follow-up survey link."
Oh, it creates "agent" instances, and these "agent" instances can delegate and pass messages to each other? Just sounds like OOP.
😆😆 i don’t want to say it’s bad because it supposed to be educational, I’m just not sure why they put time and energy into it either. I don’t know, just my opinion
Tyler will be very confused with that email lol
😅😅 🤷♂️🤷♂️
Langgraph wins
LangGraph Studio!
you like it that much? I only played with it once or twice
I helped fix the vector memory stream in LangChain about a year ago. The team was fantastic, and it’s an awesome framework. I’m just cautious about black-box solutions, as I prefer having transparency and control over the internal workings.
They blatantly stole it from a framework called Swarms, which is better and more fleshed out than the OpenAI version.
I haven’t tried swarms yet but I have heard of it and was curious about that
Garbage…that’s just not true