Did Pinecone Just KILL OpenAI’s Assistant API? | Pinecone’s New Assistant API

Рет қаралды 3,944

Күн бұрын

In this video, I dive into Pinecone's latest feature, the Pinecone Assistant, showcasing how to effectively use its core functions to enhance your AI agents. You'll learn about the simplicity of adding files and building powerful AI applications using the Pinecone API. I also demonstrate how to achieve high-quality results with relevant answers grounded in your proprietary data, ensuring full control and security. Plus, I'll show you how to quickly prototype and ship AI assistants with the new drag-and-drop upload and chat interface. This tutorial is ideal for anyone looking to build and manage AI assistants seamlessly.
🦾 Link to Pinecone Assistant Tutorial Demo: bit.ly/3W038PH
🌐 Visit My Agency Website: bit.ly/4cD9jhG
🚀 Work Together on Fiverr: bit.ly/3XorT7R
📅 Book a Consultation: bit.ly/3Ml5AKW
📰 Join My Newsletter: bit.ly/3WVEHlK
_________________________________________________________
👋 About Me: Hello! I'm Mark, a seasoned Data Science Manager by day and an AI automation agency owner by night, hailing from Canada with a decade in the AI space. At Prompt Advisers, we specialize in cutting-edge AI solutions, helping individuals, businesses, and agencies fully harness applied AI. Having been featured in interviews and recognized for our innovative contributions, we're dedicated to guiding you through the AI landscape.
🚀 My Goal: My mission is to empower you with the knowledge to explore AI technology in your ventures, whether you're an individual, a business, or an agency. I aim to help you leverage applied AI to its fullest potential, providing insights, sharing experiences, and offering partnerships to bring your visions to life.
#PineconeAssistant #AIagents #AIIntegration #TechTutorial #ArtificialIntelligence #AITools #PineconeAPI #GPT4Omni #GPT4o

Пікірлер: 16

@BrockMesarich 2 ай бұрын

thanks for breaking this down so simply Mark!

@GetzAI 2 ай бұрын

Great review. How will this work in production? With more files and my own LLM? Why could this not be pointed to my current indexes?

@Mark_Kashef 2 ай бұрын

thanks so much for the feedback! In terms of production, I would say this is definitely not ready for full-scale deployment in its current state. I think the idea is that in the typical RAG architecture, this could replace or augment the portion of the pipeline designated in pre-processing, formatting, chunking, embedding, and upserting vectors in one-step. Especially since that's Pinecone's forte, they likely have tons of knowledge on how to dynamically chunk and rerank based on document contents versus using static or systematic methods. Remains to be seen how they're going to prepare. this for concurrent usage and large scale applications..

@kk-nw2rx 2 ай бұрын

Really interesting. Please give us more Bland content too!

@Mark_Kashef 2 ай бұрын

absolutely! Taha is actually coming down to my city next week to film some stuff IRL!!

@kk-nw2rx 2 ай бұрын

@@Mark_Kashef awesome duo

@RedCloudServices Ай бұрын

how does Pinecone deal with tables inside the PDF?

@Mark_Kashef Ай бұрын

It’s decent but given that it’s still in Beta, definitely could use some work on more complex PDFs that would otherwise need something like LlamaParse to better extract them

@daveinpublic 2 ай бұрын

But how many tokens does it take? When you have a few docs being used, does it balloon up? I’ve seen the forums at OpenAI and it looks like the token use becomes unreal after just a few messages.

@Mark_Kashef 2 ай бұрын

Pricing is a bit amorphous at the moment, but I imagine it will be cheaper if they don’t provide configuration control to the user to augment the prompt, chunk size etc, and use a consistent system prompt across the board for retrieval. Unlike OpenAI, I don’t expect storage to be much more expensive either as they’re likely solely storing the vectors of your documents instead of the actual PDFs/docs. Not fully clear to me so will keep everyone updated once I figure that out!