Is AUTOGEN Microsoft's Langchain alternative or much BIGGER???

  Рет қаралды 12,927

1littlecoder

1littlecoder

10 ай бұрын

Autogen Github: github.com/microsoft/autogen
Autogen Announcement: www.microsoft.com/en-us/resea...
❤️ If you want to support the channel ❤️
Support here:
Patreon - / 1littlecoder
Ko-Fi - ko-fi.com/1littlecoder
🧭 Follow me on 🧭
Twitter - / 1littlecoder
Linkedin - / amrrs

Пікірлер: 54
@HB-kl5ik
@HB-kl5ik 10 ай бұрын
Holy hell! The intros are getting smooth as butter now (literally watched 3-4 videos you posted last). To the point I don't even notice it's an intro and I get straight into the video, wow!
@1littlecoder
@1littlecoder 10 ай бұрын
I actually didn't record that separately as an intro but yeah kept that in mind to give the users something to watch the video. How was the speed btw?
@HB-kl5ik
@HB-kl5ik 10 ай бұрын
@@1littlecoder yeah that's what I meant, in short I wanted to say delivery was awesome (perfect) :D
@1littlecoder
@1littlecoder 10 ай бұрын
@@HB-kl5ikThanks for the feedback and the support!
@n3cr0ph4g1st
@n3cr0ph4g1st 10 ай бұрын
Great channel bro keep it up. This looks very interesting
@1littlecoder
@1littlecoder 10 ай бұрын
Thanks bro!
@Tyronne_
@Tyronne_ 9 ай бұрын
Great video man. Please keep them coming!
@blender_wiki
@blender_wiki 10 ай бұрын
Very nice but at the moment too OpenAI centric, the avantages with langchain is that you can easily mix different models for specific tasks
@TheLibin27
@TheLibin27 9 ай бұрын
Thats true. But in the process, Langchain have abstracted way too much. We had to use QdrantDB in one of our projects and we use langchain too. We started to hit lot of limitations with data retrieval API and finally decided to ditch Langchain and call the API's directly. But honestly, Langchain helped us to get started. If Langchain was not there it would have been very difficult to started with lot of these new concepts.
@arjenvanputten9744
@arjenvanputten9744 9 ай бұрын
Is it? From what I saw in the examples you can also configure other model endpoints. In one of the examples I saw a local llama model being used. Didn't look in to it yet.
@JanMuller-qj5im
@JanMuller-qj5im 9 ай бұрын
​@@TheLibin27 I had the exact same problem with chromadb. In the end I had to use both, the original api as well as the langchain version of it to get a reasonable amount of functionality.
@surfbort
@surfbort 9 ай бұрын
thanks for your feedback@@TheLibin27
@aaronward9140
@aaronward9140 9 ай бұрын
For those asking Langchain vs Autogen - heres my take: - Langhchain is a more general purpose LLM application framework, so it has more features out of the box. However its documentation, code and general direction is inconsistent and confusing at times. There are open source contributions being added all the time, but i find that they're often buggy or don't work in their entirety. - Autogen is a agent framework, it provides less general features but is much more powerful and simplified way of implementing agents. One cool features is that it allows users to create chat interactions with AI agents while they're breaking down a task and solving it, allowing for "human-in-the-loop". Chances are that you wont use either/or but you will use both of them in conjunction with each other. You may need stuff in both, but in my opinion Autogen is better at least for agent based workloads.
@atchutram9894
@atchutram9894 9 ай бұрын
Kindly give an example of a feature that langchain potentially has which Autogen does not possess.
@aaronward9140
@aaronward9140 9 ай бұрын
@@atchutram9894 documents loaders, chains, text splitting utils, other tools and custom use case add-ons. The way i see it its like langchsin takes the philosophy of "we need to defined ways to guide the LLMs in the application" while autogen takes the philosophy of "let the agents figure out what they need to do"
@jawadmansoor6064
@jawadmansoor6064 10 ай бұрын
don't you need openai api key for this? can autogen be used with llama.cpp (guff models) if so please make a tutorial
@d_b_
@d_b_ 10 ай бұрын
Would you recommend this in production, or are you better off building up your own agentagent workflow. I haven't seen much positive experience with LangChain in prod, and people are prototyping with it then extracting the prompts and optimizing from there
@1littlecoder
@1littlecoder 10 ай бұрын
I guess Microsoft's target audience would more of someone who moves it to PROD than Hobbyists, so I guess it's worth a shot!
@VinayLobo1986
@VinayLobo1986 10 ай бұрын
Can you cover Qwen 14b?! Especially if it can be fine tuned and how?!
@leafdriving
@leafdriving 10 ай бұрын
Time is precious. When I have some, I will try this. (I bet you wear a lot of hats too) Thank you for pointing this out. At a minimum, I want to understand how this works :)
@tijendersingh5363
@tijendersingh5363 10 ай бұрын
can you also make a video how to use it thanks
@KevinKreger
@KevinKreger 10 ай бұрын
It's an extension of the work they were doing in semantic kernel, which lacked support for other LLMs just like Autogen. They should have just added the workflow definitions to semantic kernel IMO.
@rickymacharm4426
@rickymacharm4426 9 ай бұрын
I think we need to see your face more as you inform us on AI news. This is great video. Thanks for tsking the time to shoot it.
@TheReferrer72
@TheReferrer72 10 ай бұрын
Little Coder back to his best! Cool they have colabs.
@1littlecoder
@1littlecoder 10 ай бұрын
Did you not like any of the recent videos ?
@TheReferrer72
@TheReferrer72 10 ай бұрын
@@1littlecoder All your video's are good sir.
@1littlecoder
@1littlecoder 10 ай бұрын
@@TheReferrer72 Thank you, just asked to improve further 😊🙏🏽
@forcanadaru
@forcanadaru 10 ай бұрын
Those agents can play the roles of the different organs of brain and, working together, can compose AGI
@benshums
@benshums 9 ай бұрын
I am the hippocampus accessing memory...
@benshums
@benshums 9 ай бұрын
I am the prefrontal cortex... asking the hippocampus for memory... asking the amygdala for panic emotion
@markupshaw
@markupshaw 10 ай бұрын
Thank you for your rundown on Autogen. I just hit subscribe to all. Keep up the good work.
@1littlecoder
@1littlecoder 10 ай бұрын
Thanks for the sub!
@faridullahkhan1
@faridullahkhan1 10 ай бұрын
is it possible to run autogen in windows 10 locally and not in jupyter notebook ? can the examples be in python files format for easy execution locally ?
@DeonBands
@DeonBands 10 ай бұрын
You will not believe it, this weekend I sat and thought to setup a chatroom, with exactly this type of thinking, using Laravel and api's. This solves so much if it works. There goes my chatroom idea, for something much better and seemingly simpler. Tanx a zillion.
@1littlecoder
@1littlecoder 10 ай бұрын
Woah, When you build that please let me know (if it's public) would love to see it!
@webdancer
@webdancer 9 ай бұрын
Been dreaming of this particular application of llms. This is a wonderful gift from Microsoft.
@sgramstrup
@sgramstrup 8 ай бұрын
MS doesn't give 'gifts'. IT tech are constantly trying to box developers into their 'open' frameworks. They are only semi-open oc, so IT tech controls the development environment for years and years. That selfish control is bad for the dev community, so don't ever use 'open' frameworks from corporations. It's like pissing your pants on a cold winter day, and we all lose..
@henkhbit5748
@henkhbit5748 10 ай бұрын
Seems promising. Its free right? Thanks for updating us with new tools in the LLM offering👍
@1littlecoder
@1littlecoder 10 ай бұрын
Yes it is. Although you'd need OpenAI key to engage the LLMs
@JimMendenhall
@JimMendenhall 10 ай бұрын
Thanks for sharing this find!
@1littlecoder
@1littlecoder 10 ай бұрын
My pleasure!
@abagatelle
@abagatelle 10 ай бұрын
I guess Microsoft could swallow up Langchain if they wanted to.
@sgramstrup
@sgramstrup 8 ай бұрын
..instead of actually joining Langchain and further real open source projects. That shows that MS doesn't do this for the OS community, or for developers, but for their own benefit and control of the dev environment. Stay away from MS and their scamming frameworks..
@IronMechanic7110
@IronMechanic7110 10 ай бұрын
Impressive
@riyayadav8468
@riyayadav8468 10 ай бұрын
Hey what you think about show where multiple AI talking to each other like Big Boss and they will talk about those thing which the show host will ask him just a idea
@aitarun
@aitarun 9 ай бұрын
nice
@gRosh08
@gRosh08 9 ай бұрын
Uncle Amos said he is going to make a sandwich and try again later.
@MoiezIzmail
@MoiezIzmail 12 күн бұрын
bro you are all over the place(not a good thing). It seems like you didn't really know what AutoGen is, so you were just reading and trying to understand it while explaining. Your energy and flow is all over the place so calm down.
@1littlecoder
@1littlecoder 12 күн бұрын
Thanks for sharing the feedback
@s.dotmedia
@s.dotmedia 9 ай бұрын
A friend asked me what their son should learn coming out of high school to start a development career path. This.
@KevinKreger
@KevinKreger 10 ай бұрын
If anything, it looks like Microsoft is willing to play the open source agent game, while OpenAI, for obvious regulatory reasons, stops just shy of giving us agents.
@shiccup
@shiccup 9 ай бұрын
lol this is insane
@generativeresearch
@generativeresearch 9 ай бұрын
Langchain doesn't come close to Autogen
Google Releases AI AGENT BUILDER! 🤖 Worth The Wait?
34:21
Matthew Berman
Рет қаралды 225 М.
AutoGen Tutorial 🚀 Create Custom AI Agents EASILY (Incredible)
20:10
Зачем он туда залез?
00:25
Vlad Samokatchik
Рет қаралды 3,2 МЛН
Самый Молодой Актёр Без Оскара 😂
00:13
Глеб Рандалайнен
Рет қаралды 7 МЛН
Zuckerberg cooked a beast LLM - LLama 3.1 405B!!!!
14:49
1littlecoder
Рет қаралды 3,4 М.
LangGraph: Multi-Agent Workflows
24:02
LangChain
Рет қаралды 55 М.
What is LangChain?
8:08
IBM Technology
Рет қаралды 173 М.
AI business ideas funded by YCombinator
16:31
Harshit Tyagi
Рет қаралды 15 М.
AutoGen Advanced Tutorial - Build Incredible AI AGENT Teams
38:08
Matthew Berman
Рет қаралды 111 М.
AutoGen Studio: Build Self-Improving AI Agents With No-Code
27:06