I wish every AI Engineer could watch this.
VloĆŸit
- Äas pĆidĂĄn 19. 06. 2024
- đ Timestamp đ
00:00 Intro
00:02 Understanding the framework for using LLMs in various applications
02:15 Question answering with LLM
06:54 Chatbots need more than short-term memory for effective use.
09:13 LLM is central to leveraging prompt, short-term, and long-term memory
13:46 Importance of Context Window in Language Models
15:55 Implement retrieval augmented generation for chatbots
20:06 Leveraging LLM for NLP tasks
22:20 Function calling in AI models enables structured responses.
26:18 Understanding the concept of Agents in AI
28:30 Agents are the next Frontier in AI development.
32:24 AI developing with extended tools and memory capabilities.
đ Links đ
This video talks about gives a blueprint for LLM Apps. It talks about 5 different LLM levels framework for building AI Apps.
â€ïž If you want to support the channel â€ïž
Support here:
Patreon - / 1littlecoder
Ko-Fi - ko-fi.com/1littlecoder
đ§ Follow me on đ§
Twitter - / 1littlecoder
Linkedin - / amrrs - VÄda a technologie
An Indian man explaining something rather complex but so simply and smoothly on CZcams
Always a classic!
Cheers from Brazil đ§đ·đ€đźđł
the legacy continues
You just taught an entire course in 34 minutes. We Want more! We want more!
Thank you
Aye! Didn't he!
@@1littlecoder No, thank you for the amazing job.
This guy deserves serious attention, content is amazing
Thank you
This was excellent. A series would be much appreciated.
True, much needed
You have so beautifully distinguished and clarified these similar yet different concepts .
I hope this blows up and you continue the series
Extremely delightful and your examples were right on point. By placing topic breakdown in comments has helped me immensely the topics I am very weak on. Thank you so much
Thank you for this amazing course. I would be fantastic if you do more of them.
Excellent!! Thank you for the comprehensive tour.
Finally some content that puts everything so well together, like a delicious home made recipe, thank you and keep up the good work!
Wonderful !!
For last 6 months, I have been going from video to video ... blogs ... articles etc ... just getting more and more confused as to how all these terms relate to each other. This has put a lot more structure to those terms. Of course, I need to still read up a lot before even deep-diving into these topics.
Nice! This pyramid way of looking at the progression really helped me see the bigger picture and where we're headed. đđŸ
Wow. Now I know. You got exceptional teaching skills bro. Thank you for explaining.
Kind words thank you!
Thanks for this video. I knew every piece of it but your lecture connected them in my brain. Super good one. Looking forward for the next.
This is perfect! Please, more. And thank you for making everything so accessible!!
I really appreciate this video thank you for sharing your knowledge and providing context and clarity
A complete tour of current LLMs world!đ€©Really well crafted educational content! Thank you for sharing
Love the video looking forward to the rest of the series. Itâs very hard to understand all this but this video really helped me understand the concepts. It would be awesome if you could just put a bunch of personal information in a PDF and feed it to the LLM as a mini rag, but it always has to refer to.
very nice articulation, loved on how you covered overall things in simple way
Simply awesome clarity of thoughts and knowledge. Looking fwd to see more like this. Great
yes!! loved the simplecity yet effectiveness.
Well illustrated! Thanks for sharing.
This is awesome! Please keep more coming.
Loved the content! Looking forward to hearing your deep dive into agents!
Awesome, thank you!
Excellent video, thanks for sharing, very well done
Dude, seriously.. amazing job đđ»đđ» you just taught an entire course!
Glad you liked it!
Thank you sir! I like how you simplify things even for no-coders ppl like me, also I like your humor and examples.
Loved how much you have simplified the concept, given a structure to a lot of floating bits and pieces. You have definitely earned a subscribe. If I was forced to give a feedback, maybe have some live demos the next time, but this was more than perfect. Thank you!
Thank you for the feedback!
Thank you! You are concise and I appreciate you brother!đ„đŻđ
Thanks a lot. Expecting more videos like this
Excellent video content! Thank you very much!
amazing ability to explain complex concepts, thank you!!
Thank you very much
I love this lecture. It is all about LLM. Good to go for beginner.
I loved this! I think you're absolutely right about this and if we extrapolate, when the LLM OS can run on smaller, more mobile machines, we get really really intelligent robots. Not that we've not already seen some, but I don't think we will be quite ready for when that day comes.
Awesome. Love this kind of videos.
Thank you
This topic is very helpful to me, looking forward to this series of content
Glad it was helpful!
Great job as always my dear, Abdul. Looking forward to your next topics on this series.
-Arun
Very lucid with clarity and amazing simplicity. Thank you
Thank you đđŸ
I'm not gonna lie this has been the singular most useful AI video I've watched on the internet till date. Please keep uploading more content you're amazing at this
Super Video , hopefully we will get more about building - agents - thanks a lot
You are just wonderful at explaining complex concepts. Thank you a lot!
Thanks very much
Great vid, very helpful, thank you, subscribed.
Amazing explanation. Just got my view better at how to look at these things
This tutorial is *_Exceptional!_* It's like taking a class by a Professor that can communicate complex ideas as "common sense," easily understandable concepts in very logical, building-block progression. The presentation style is engaging, compelling, and informative.
May I suggest; if you decide to expand out this tutorial with further instruction with each "building-block", would you consider an entire program based around creating this "LLM OS Stack" on your _Local Hardware?_ The reason I suggest this is from my experience over the past 19 Months. From knowing nothing about LLM's to creating an Open-Source based private LLM for an organization. Now, at a personal project level, I am attempting to build a local full "LLM OS stack" on my hardware using ChatRTX with one Nvidia RTX A6000. If you are not familiar with ChatRTX, its something I think you would see great promise in.
Thank you for the kind words. I could never try ChatRTX given I don't have an Nvidia machine. I think the definition of LLM OS is still somewhat vague!
Exacly!!!!
@@1littlecoder I interpret your statement to mean you don't have a PC? You use a Mac? As you know, Nvidia has a stranglehold on underlying AI hardware (Proprietary Tensor)...it is what it is. Obviously, there is no "Nvidia "machine", but Nvidia hardware (GPU). An _Investment_ in a workstation-class PC + some RTX or better GPU would allow a ground-up Private LLM build tutorial (as it is simply impossible to escape the hardware layer when teaching LLM). I can not stress enough the importance of local hardware, and a SECURE private LLM platform. I learned this the hard way - we realized that sharing very proprietary info with PUBLIC hyper-scale LLMs (ChatGPT, Gemini, etc.) because it was ostensibly "cheaper," "expedient," and convenient meant our secrets were ultimately shared with the world - as these platforms train their LLM's on user data.. Our laziness allowed the huge investment in our business methods to be shared and profited on by the hyper-scale LLM's. At the end of the day, they OWN all information. This is evidenced by recent showcasing of both ChatGPT4o and that google I/O thing, where a presenter spoke about "privacy". If these platforms were not already training their foundational LLM's on user data, there would have been ZERO reason for both of them to suggest some "element" of privacy.
Let me put it to you in no uncertain terms:
"Confidential Computing" is a Cloud provider term which includes methods that seek to provide greater assurance for cloud customers that their sensitive data in the cloud is protected and confidential. Cryptography at the hardware level would ensure a customers data is accessible only to authorized programming codes. Allowing for computation in a Cloud that is invisible and unknowable to anything or anyone else, *_including the cloud provider._* But NO hyper-scale LLM offers this.
Why would anyone trust their most secretive information to a cloud (or a commercial AI platform) provider that has no security protections for their data? A totally secure on-site environment, IMO, is the only way to apply - and guarantee security - the LLM OS stack is on local hardware. Otherwise, we have seen how over the past 30 years how hyper-scalers hijack customer data, and use it for their own goals. I do understand if such an "Investment" in hardware is not an expenditure you have a prioritized interest in pursuing. However you would be a great teacher for this. Again, and with all due respect, thanks for responding.
Keep doing this. Content is excellent.
Excellent video with crisp and clear explanation !
Thank you
Thank you for very good summary of the very timely concepts!
Glad you liked it. Thank you đ„
Thanks for making this much information Easy to grasp in no time
Glad it was helpful!
excellent bird's eye view. Can't wait for LLM OS experiments!
Amazing delivery..kudos to yiur efforts.
Thanks a lot
I would like to extend my deepest gratitude for the valuable information you provided, which has helped me develop a new and different perspective on the world of artificial intelligence. Every minute spent watching your videos was enjoyable and insightful, offering a glimpse into what the future holds.
I sincerely hope you continue to create such outstanding content, as its impact is significant and its value immeasurable.
Thank you once again
Thanks a lot. I was so confused with different terms used you solved it.đđ
Most welcome đ
Very well done. I will watch more if you do them.
Thanks. Let me know if this sounds good to your taste czcams.com/video/oWXTWqsSos4/video.html
Absolutely fantastic đ
Seamlessly explained..Thank you so much bhai, We need more content like this.
Thanks bro đđđŸ
Good framework, thank you.
Thanks for this session
Thanks a million, this is a great explanation
I loved this video. Excellent summary. Thank you so much.
A great overview, what happened the last couple of years. đđđđđ
Amazing visuals. Great job
Thank you
Really great thought process
very good video. great synthesis. keep them coming!!
Thank you
Excellent video
Truly
Glad you think so! Thank you!
Just amazing content đ
Very insightful video. Thankyou very much
You are gifted with a great teaching ability đđ»
Thank you
â„thanks brother, people like you are making our country famous when it comes to education in a very good way
Thanks bro!
Fantastic video sir. đź
Very nice job! Thank you
Thanks for watching!
this is an amazing video !
Thank you brother !! đđ
that's great! literally a great work..i was thinking the same that how llm os will solve my daily problems ....awesome
Well explained, thanks
You are an awesome teacher ..
Excellent tutorial. LLM brought a set of opportunities. We startups sometimes get lost with so many alternatives. Your tutorial put "order in the house" as we say in Brazil. For the end customer, everything is AI, it's up to us to know the differences to plan and charge the correct price of product and services. Tks !
I convinced Sir please continue please
Bro i could have pushed 100 likes. Wonderful explanation. Subscribed â€
amazing video, reallly loved the explanation, we want moree engineers like youu, one question: how does agents solve captcha's for doing some particular tasks, and make another video on how to get started with each of the layers like chatbots, RAGs etc practically to build few projects in them?
Excellent explanation. Keep posting more
Glad you liked it. I hope to create more content like this
Great explanation!
Pretty good content! Thanks
Glad it helped!
Valuable content and an insightful summary of llm applications. Could you add links to resources to further explore these verticals?
Thank you so much.
Defenitiely salute to you for sharing this knowledge
Thank you đđŸ
Great explanation
Appreciate your content on this subject. Which "agent" development system would you recommend today ?
thank you for this content.
You're welcome
Thank you â€
Please complete it, need more
Very nice overview of LLMs!! You said you will explain more about agents separately. More on agents as a separate series please.I have subscribed and will be waiting for it. Please continue doing the good work
Thank you. I published a kind of prequel to this czcams.com/video/oWXTWqsSos4/video.html let me know your thoughts on this!
Thanks for sharing the prequel! It helped me get an idea of what's an LLM, how it works, and what basic and fine tuned LLMs are , which helps me understand this overview of use cases better. It's really summarised well in you video. Looking forward to follow and learn more about LLMs, agents, their use cases and use them.
Thanks for the feedback
Great video. I really think you should produce more like this
I tried something similar czcams.com/video/0AfQXFj9m7Y/video.html
Amazing content Manâ€
Thanks man!
Fantastic!!!!!!!!!
Thanks, good one
Great video thanks
You're welcome
Thanks, Well-made content! Do one on NVIDIA NIMS, and where does it fit in your framework?
It's very useful đ thank you
You're welcome
Well explained mate. Can you talk through Multi function calling LLMs like Gorilla open functions vs Agents and link your opinion probably to a challenging questions from integration arena around how Orchestrated workflows work vs Choreography. Will agents evolve to be orchestrated systems with central system of orchestration or work as a choreographed system independent of control and work based on conversation.
Very good explanation â€â€
Thank you đ