Відео

How has he become a monster?

Hey, thanks for watching! I'm glad the video was helpful. To clarify, RAG (Retrieval-Augmented Generation) is a feature within LangChain. LangChain lets you coordinate various workflows and LLMs (Large Language Models), including more advanced tasks by connecting multiple LLMs. RAG is just one option among many in this setup. Think of LangChain as having different nodes for different functions, and RAG is one of those nodes. So, RAG is a tool or technique used within a LangChain workflow for more complex tasks. I hope that clears it up. Keep the questions coming-I appreciate them. Cheers!

To get a better sense of this, you can try out LangChain, which is more code-heavy, or FlowWise, which is a low-code, no-code option for doing what you would in LangChain. I suggest checking out FlowWise, which is open-source and free, to see how this works in practice.

@@MannyBernabe ahh that makes a lot of sense thank you so much! one more quick question how does OpenAI's new update (specifically the File Search tool within the assistants) compare to Langchain? Are they basically serving the same purpose -- allowing you to add context to your LLMs with vector stores. when would you use langchain then and when would you use OpenAI's vector store, and what would be the difference within each approach? thank you so much for answering all these questions btw! really helps!

@@krishras23 You got it. Add files to a chatgpt chat is a RAG. Use langchain when you have more complicated workflows.

Actually, no. The beauty of RAG (Retrieval-Augmented Generation) is that you don’t have to fine-tune a model on domain-specific data. Fine-tuning is difficult and requires technical expertise. Instead, RAG allows an LLM to search a data corpus and include the relevant information in its prompt, making it an easy way to add domain expertise to an AI service. Before fine-tuning, try RAG. There are great options like Stack.ai, BotPress, and VoiceFlow where you can easily implement RAG in a no-code, low-code way. Check them out and let me know your findings. I’m glad the video was helpful-keep the questions coming! I really appreciate them.

@@MannyBernabe Hey thanks for the reply. Really appreciate it. I'm new to all this but I'm very interested in learning more and hoping to get to the point of developing AI chatbots that can be used commercially for small companies. Would Vertex AI be one of the RAG options too? Since it's developed by Google I'm thinking of using it to develop virtual agents.

@@fascinatingstories101I'd start with botpress, voiceflow and/or Stack AI. No/code code, easy to use, a lot of great content and community out there. Good luck and keep me posted.

@@MannyBernabe Thank you! Yes will do. Keep up the good work with your channel.

That was the goal. Glad it was helpful! :)

I value your time! :)

Very kind of you, Max. Thank you!

Thank you, Zack. Means a lot coming from you! :)

Great to hear!

Agree. Thanks for feedback. 😊

You got it. First will ping the corpus for relevant data, retrieve and insert into prompt. If none, then you just get the standard LLM output. Hope that helps.

Thanks!

🤣

Hey Jason, the current best practice is to first try RAG with a stock LLM and see if that works. If not, then consider fine-tuning, because it requires more effort than RAG. Hope that helps.

Book of revelations should be good too!

Correct. The retrieved context still needs to fit into the context window with the original prompt. In terms of compression, we can summarize the retrieved context, saving space as well. Hope that helps.

indeeded!

See my reaction video. Unbelievable at first glance.

@@MannyBernabe yes

Stunning!

Hope this helps!

Hey there. Sometimes the file to large. It gets throttled in transit.

oh, that's a good idea. Thank you. will have to try it out next time on I'm on it.

Thanks, Sathya. I really like the batch function.

Hey There, they're still rolling out. I completed the below form that I got from @mattVidPro AI and I got access within 4 hours. Let me know if works for you! docs.google.com/forms/d/e/1FAIpQLSdiJK7uFzjfIqbH9wGeXnmR3rwjnLUMwih_W5Cet0FVdv8-aQ/viewform

Yes. Bing as well. You use it there for free too. May just take a while.

why do u think so

He's talked about this many times, written about it in books, etc. They absolutely do this.

There's no way they're not doing this

I was at Amazon (2000-2011) when this happened. It's absolutely true

I worked for Amazon 2020-2022. We 100% used this on any design review. Just started implementing it at my current company. It's amazing. Especially for tech design reviews.

Perfect. Read at go.mannybernabe.com/ai-start. Would love your feedback! -Manny

LLM's are alien technology brother. Will help with SO MANY things. Exciting times!

Likely only US traded assets for now.

Thank you! Glad you found it helpful.

I still like them big and front and center on the videos. Does the captions hurt your viewing experience?

@@MannyBernabe it does.

Thanks, Ray! Adding to the list. Will you tell me more? What type of metrics/insights in particular were you looking for? Asset type? Stocks, Mutual funds, ETfs, etc?

Thank Manny, I’m looking forward to that video. I actually used your video to create an account and added the PortfolioPilot plug. I then went to Vanguard downloaded my portfolio in a CSV file then cut and paste the ticker symbols for all my holding and current value into ChatGPT. I ran the full assessment and produced some interesting result. I’m currently implementing several of the recommendations to diversify my portfolio.

@@raydockery9402 Nice. What insights were the most valuable/actionable for you?

Hey Robin, good luck with the exam. A lot good free youtube videos to look at.

@@MannyBernabewhere, please? Thanks ! I’m Brazil.

Databricks, Hugging Face,

PLTR's game to lose. Very well positioned to own the LLM for Enterprise category.

Indeed. I offer my humble suggestions. ☺