Hey everyone👋, we are currently looking for AI Agent Developers to join our team. Please apply in our discord in job-posts channel if you are interested. (link in description) Thanks!
Im interested in creating an AI agent for a small project with a finite amount of knowledge database with use for RAG purely for context with fixed API costs. How do I contact you?
Streaming is sick. I made an internal slack bot for people to chat with. First I wasn't using steaming, and whenever people were chatting with it, I needed to wait for the entire request, and everyone was complaining the bot was shit and slow. Then I build in a streaming function, now the bot is creating a slack message almost instantly, and then editting the message as the words are coming in from OpenAI. Now the bot is preferred over using the ChatGPT web-interface, because people don't have to context-switch from slack to the browser
Hi, how could you add a knowledge base to agency swarm? For example, how would you train the AI based on your previous Facebook posts to create new ones with a similar style? Thanks!
Yes, I studied CS, but dropped out of college after my first year. After that I read a ton of books on AI and worked as a freelance ML engineer for 3 years. I am still quite involved in coding, however I am trying to focus more and more on design work, rather than on implementation.
Hello Arsenni, Thanks for the all informative nad great value videos. Could you please create the playlist so it we will easy to follow and watch? Currently, there is no order in video which one to go first.
Hello! Thank you for explanation of creating and using new assistant and for your library dedicated for simplify this process. But i'm lokking for a method of using an exist assistant. Because it must use my database before giving an answer to user. I did it but without streaming and the task is complicated by the fact that assistant returns every paragraph in html blocks, so i need to accumulte chunks until a closing tag. Nonetheless I have a sense how to resolve this task, but existing assistang with streaming still a hard task for me. Can you give me any clue?
Hey nice video! I would really love to know the development process of your saas platform, I want to Launch my own AI SAAS but I have a lot questions and confusions in my head.
hi ,i have app use twilio to get a vioce query from user than take that query to gpt model to get a response than sand this response to twilio to respod to user my probleme is whene user speak its take so long to respond any idea
I don’t think you can use gpt streaming, because if you are sending back a text, you need to have a complete response anyway. The only way to reduce latency in this case is use a smaller model. However, if you are sending back voice on a call, you could probably use streaming when generating voice from TTS model.
The big problem is that assistants API costs are uncontrollable. Until they fix the back-end and billing model, they are useless for anything but a demo.
Yes you are right. They definitely need a way to control amount of tokens that goes into the model. I am sure they will add it soon. In the meantime, I recommend just creating new threads as needed
Hey everyone👋, we are currently looking for AI Agent Developers to join our team. Please apply in our discord in job-posts channel if you are interested. (link in description) Thanks!
Im interested in creating an AI agent for a small project with a finite amount of knowledge database with use for RAG purely for context with fixed API costs. How do I contact you?
Bro you are a gem, other creators didn’t understand the streaming that’s why they didn’t talk about it. Great value 👏
🙌
thanks for the videos, i think you're one of the very few whose ideas have change the way i approach these tools
Streaming is sick. I made an internal slack bot for people to chat with. First I wasn't using steaming, and whenever people were chatting with it, I needed to wait for the entire request, and everyone was complaining the bot was shit and slow. Then I build in a streaming function, now the bot is creating a slack message almost instantly, and then editting the message as the words are coming in from OpenAI.
Now the bot is preferred over using the ChatGPT web-interface, because people don't have to context-switch from slack to the browser
did you tried to use your agency swarm with some Claude 3 model like the cheap Haiku that is beating gpt-4?
Hi, how could you add a knowledge base to agency swarm? For example, how would you train the AI based on your previous Facebook posts to create new ones with a similar style? Thanks!
Amazing Video! Thank you so much!
Hey just curious have u stuided cs or gathered team of devs and u mostly run ideas and business and sales
Yes, I studied CS, but dropped out of college after my first year. After that I read a ton of books on AI and worked as a freelance ML engineer for 3 years. I am still quite involved in coding, however I am trying to focus more and more on design work, rather than on implementation.
@@vrsen Hey do you have a way to account the tokens in my loved framework Agency Swarm.. Appreciate your efforts!
Hello Arsenni, Thanks for the all informative nad great value videos. Could you please create the playlist so it we will easy to follow and watch? Currently, there is no order in video which one to go first.
Right, will do. Thanks
God bless you bro this is the best tutorial ever
Hello! Thank you for explanation of creating and using new assistant and for your library dedicated for simplify this process. But i'm lokking for a method of using an exist assistant. Because it must use my database before giving an answer to user. I did it but without streaming and the task is complicated by the fact that assistant returns every paragraph in html blocks, so i need to accumulte chunks until a closing tag. Nonetheless I have a sense how to resolve this task, but existing assistang with streaming still a hard task for me. Can you give me any clue?
By the way i’d love to learn about your saas development process. It would be incredible.
Love the details.
Hello @vrsen, what are the advantages/disadvantages of Agency Swarm compared to Crew AI?
Hey nice video! I would really love to know the development process of your saas platform, I want to Launch my own AI SAAS but I have a lot questions and confusions in my head.
Great info!
I would like your opinion on Devin.
hi ,i have app use twilio to get a vioce query from user than take that query to gpt model to get a response than sand this response to twilio to respod to user
my probleme is whene user speak its take so long to respond
any idea
I don’t think you can use gpt streaming, because if you are sending back a text, you need to have a complete response anyway. The only way to reduce latency in this case is use a smaller model.
However, if you are sending back voice on a call, you could probably use streaming when generating voice from TTS model.
Thank you!!!
hi VRSEN can you create a video based on perplexity online llms and can I use their API endpoint which offers vast LLMs for creating agents
The website does not work in the creation part of openai
Do you work with smb?
Yes, checkout agents.vrsen.ai/
@@vrsen th-cam.com/video/iHVP7SkpUHE/w-d-xo.html
Definitely would like to hear a video but is that something that someone can call you and consult about?
Yes, of course. You can use a link in the description. I reveal any secrets and insights for my consulting clients
@@vrsen
The website does not work in the creation part of openai
This man is nice. But why is so underrated on youtube. Keep it up man.❣
Go go go!!
The big problem is that assistants API costs are uncontrollable. Until they fix the back-end and billing model, they are useless for anything but a demo.
Yes you are right. They definitely need a way to control amount of tokens that goes into the model. I am sure they will add it soon. In the meantime, I recommend just creating new threads as needed
All talks n no practice step by step funct
The website does not work in the creation part of openai