GPT4o + CrewAI: Twice as fast? Half the Cost?
ฝัง
- เผยแพร่เมื่อ 1 ส.ค. 2024
- 🤖 Download the Source Code Here:
brandonhancock.io/gpt4o
Don't forget to Like and Subscribe if you're a fan of free source code 😉
📆 Need help with CrewAI, join our FREE Skool Community:
skool.com/ai-developer-accele...
ChatGPT4o is the latest LLM model released by OpenAI. This new model is supposed to be 2x as fast and half the cost of the previous version. In this video, we put this new model to the test with CrewAI to see if it speeds up our crews and makes them cheaper to run.
📰 Stay updated with my latest projects and insights:
LinkedIn: / brandon-hancock-ai
Twitter: / bhancock_ai
Checkout my other CrewAI Tutorials:
- CrewAI Tutorial: Complete Crash Course for Beginners - • CrewAI Tutorial: Compl...
- CrewAI Tutorial for Beginners: Learn How To Use Latest CrewAI Features - • CrewAI Tutorial for Be...
- How To Connect Llama3 to CrewAI [Groq + Ollama] - • How To Connect Llama3 ...
Timestamps:
00:00 Introduction
00:27 GPT4o Experiment Overview
01:36 Join The Free Skool for AI Developer
02:02 Experiment #1 - LLM Only
04:08 Experiment #1 - Results
06:10 Experiment #2 - LLM + Internet
08:10 Experiment #2 - Results
09:35 Experiment #3 - LLM + Internet + Embedding
11:44 Experiment #3 - Results
12:29 Outro - วิทยาศาสตร์และเทคโนโลยี
Amazing analisys.
I'm working with Crew AI and GPT-4o to classify law firms documents based on the context within it and GPT-4o crushes GPT-4-turbo.
I'm excited to one day maybe see gpt-4o finetuning. It would be game-changing for my company.
Amazing work, Thank you!
Hey Brandon, Excellent tutorial :-)
Another super tutorial, thanks Brandon. I appreciate all the effort in these.
Thank you Rob 😁
Bro, I've seen some intelligent chaps out there talking AI, but you're running some game here dude. Well done. Well evolved.
Thank you! 😅
If you can think of any problems or new videos you want to see, please let me know! I’d love to hear your ideas!
thanks, Brandon!
As always fantastic informative video Brandon 💪👏 see you later!
Amazing video!
Thank you so much! I’ve had a ton of fun playing around with this new version of ChatGPT!
Can’t wait to see what you guys build using this new version!
Super clear, super focused on the topic. Can you share your VS code setting with us? It looks fantastics
Is crewai really something which I should be using for production?
Hi, I'm a big fan of your work! I have a question about streaming intermediate steps in a process. Specifically, I'm trying to see the thought process and task outputs in real-time. I've tried using callback functions, which show the intermediate steps, but they don't actually stream the responses as they happen.
Could you help me refine this approach or suggest a better way to achieve real-time streaming of these intermediate steps? Thanks!
Nice comparison cheers. Id love ro see actual real world usage for these agents... I onlt ever see the examples provided by the tool developers. Ive played around but never get to a completion im after
Hi Brandon, can you make a video on doing data analysis using crewAI?
Nice comparison. New langsmith type service for me.
I want to know do we need to know about machine learning to create ai agent??
Hi, how can I give agents a list or a comments.md file to analyse ? I want to analyse comments but I cannot give agents a list or file could you pls explain me Im really stuck and even gpt-4o couldn’t solve theses problem
Brandon is there any free alternative to this? Something I can run on localhost without API keys?
First of all, I really congrats you bro, your ideas are very useful thanks again. However, I am very curious about your opinion on the following as a custom agent tool structure: “For example, how can you plan a code evaluation agent structure?” For example, .java code will be entered as input, the execution of the code will be evaluated with a certain weight, the proper writing of the code will be added as an evaluation parameter, and the use of commands will also be added as an evaluation parameter, and ultimately a custom agent that returns a score. How do you think we can do it? Wouldn't it be great if you prepared such a content video? I'm curious about your ideas on this subject!
Hey! For more in depth questions like this, I recommend joining our free Skool community:
www.skool.com/ai-developer-accelerator/about
If you could post what you’re trying to do in there with some examples, we can definitely get you pointed in the right direction!
Marvelous guide! Thank you!
i`ve done everything, but i do not have outputs. How i can have separate result output on every my task exexution saved to MD file or smth?
I would make the output file tied to a variable that’s unique to that task! That’s definitely the easiest way!
If you hop over to the free Skool community, there’s a ton of people to help you with questions like this!
www.skool.com/ai-developer-accelerator/about
Hey Brandon, I have had some issues running the Stock Analysis example (Browserless API being a paid API). Did you pay for the API, or how did you solve the problem?
Try scrapingant instead. There is a free tier, and if you search for crewai and scrapingant someone shared a helpful code snippet to sub it in for browserless.
Hi brandon, when do you think ai agents will become good, actually solve hard problems for companies in a lot of industries?
Hey! I think agents are already at the point where they can solve problems for companies.
The biggest is around the tooling for these agents.
Big companies want consistent results and easy to deploy solutions.
I think there is a huge opportunity for those who start AI deployment and work flow companies!
pensé a sortir de chez vous la technologie ne doit pas devenir une addiction 😊😊👍👍🤝🤝🤝🤝🤝🤝🤝🤝🤝🤝🤝🤝👍👍😊😊😊😊😊😊😊😊😊😊😊😊😊🙂😊😊😊😊😊😊😊😊😊😊
Thanks for the great video! I love the whiteboard software you use. Can you share its name?
Thanks man! Love your name by the way!
It’s called excalidraw
@@bhancock_ai ❤Thanks for the nice words! I appreciate it. And thanks for sharing the name of the whiteboard software. It looks really helpful!
They drop the price because we will send them video/audio which take a lot of token each second.
top top top
thought about leaving your home technology should not become an addiction 😊😊👍👍🤝🤝🤝🤝🤝🤝🤝🤝🤝🤝🤝🤝👍👍😊😊😊😊😊😊😊😊 😊😊😊😊😊🙂😊😊😊😊 😊😊😊😊😊😊