วีดีโอ

You people are pathetic

What temperature do you use at the experiments?

Great video! Thanks. Only challenge I faced is I questioned my eyesight when I heard 68% instead of 86%, and caused me a mild anxiety 😅

Video is very good. We need more independent benchmarks! But you need to prove it's actually calling the models.

Good video, I like that you were strict and ran the tests many times. However without the code being in the video or linked in description, we don't really know how well the tests were executed.

I’m trying to make my game multiplayer but when I go on a different account and look at the game it doesn’t have the people symbol

Nice content , I want to ask what is happening to coding now since current ai tools able to code by their own, so as an software developer we have to provide pseudo code to these agents or something else ?

This was really helpful. Thanks. Are there any differences in the tools available between them? For example when I tried crewAI - it couldn’t do instagram scraping. Wondering if others offer better more rich tools that can invoked easily. For reference - l’m a newbie coder.

Good idea! But you didn't talk about what actually happened. How many iterations? How long did it take? What are the exit criteria? Would have liked to see the final output. Also, what if the Web page is longer than the screen? Will a screenshot capture a long page?

Don't worry about all the hateful comments. I found the video useful! . Web development will change dramatically for sure. There are a lot of people who will resist the change. I'm waiting for the follow up video. Thx

please, do a tutorial...I need to do the same with Buffet and more knowledge

Thank you for sharing this experiment! I have an idea about using claude models with swarm. OperRouter service wraps different providers (including anthropic) and gives us API casted to the OpenAI API (with custom endpoint). I hope, it works with images too.

now do authentication, now do multi-user accounts, now do multi-tenanting, now do subscriptions, now do SEO, now do A11Y, now do cookie acceptance, now do ads integration and metrics. You can't build an aeroplane by just drawing an aeroplane.

are you AI?

you should not put your video over the slide contents ;-)

No, its not dead. You are living in another world maybe.

Spoken like a another AI youtuber that hasn't actually coded anything even remotely complex.

I have an idea for your next title: AI is slaughtering fake-news-youtubers with German accents. Im just getting sick and tired of this everyone tries to make a penny on AI but lacks the quality. The TH-cam algorithm gives me strange conspiracy theories just by searching for AI models just to give you an idea hoe bad it is

that is it. we are all obsolete!

dudue, why is your head in the graph? 👎

I prefer LangGraph 🦜🎡

Eyo, Florenz, can we setup a meeting to talk about AI agents? I have some questions and you seem like someone who could teach me the best way to procide

That's an interesting get_weather function you have there 🤔🤔🤔🤔🤔🤔

bro you shared you api keys

:)

Thanks for the video but there is something I don’t understand in this « benchmark ». What about the frameworks « output » ?? Does all the frameworks generated exactly the same final report ? I guess not ? Then for a « benchmark » to be meaningful I think it should includes a careful rating of the results. Why would I pick a framework that it slightly faster or easier to install if the produced results are poorer ?? Or did I miss something ?? Many thanks for your thoughts on that…

I was lost and then I found it. Super helpful and down to earth explanation. May the Almighty bless you, man!

great video!

Please stop stomping or maybe hitting with your hand. It is very disturbing.

Fyi, you can combine langgraph with crewai too

Where is agency swarm ? If you want fast, use groq, if you want cheap use ollama, if you need something done like this kind of research, speed does not matter, but insightfull results do. So improve prompt and or use better model. If you cant code, dont use ai frameworks, there zare plenty of other options now, or was opensource a requirement ?

You are comparing apples and oranges here. Agent zero is a single autonomous agent, not a multi agent agentic framework. There are a bunch of others now popping up in that space. Prompts and which llm you use, determine everything. Agent zero prompts can be easily improved, very visible in the code, not hidden deep diwn in pip package ...

Really liked the series but you have the playlist the wrong way around. Currently the results are the first video :D

Your ease of use VS instalation complexity is misleading. The ease of use trait should have it's maximum on the right as that's the established way to present it, yet "easy" is on the left side

good content!

Thank you for this review, I always like seeing different perspectives. That being said, I don't agree with your final assessment of Agent Zero. For example benchmarks step 1 through 6 could use a little more analysis. I'm thinking that cost and execution time depends on desired goal as interpreted by the framework. So if normalization of data before comparison wasn't taken into account, I'd probably reconsider what the data is actually telling us if we don't adjust or adapt for discrepancies between the different frameworks. Finally, step 7 and 8 contradict each other and so do steps 9 and 10. Step 11 which is the summary, declares the framework not to be "agentic". I think this surprised me the most and made me question what "agentic" really means when referring autonomous AI agent frameworks. This is all IMHO of course, so any feedback or insight would be appreciated!

Good video, thanks for sharing the code. Can you please make font in the video a bit bigger to understand what I see in sync with your explanations?

Btw, now gemini (in android) supports asking about the youtube video, which i thinks a very cool and practical feature

Thank you. What do you think about agentzero? I want the framework to do the work for me but i dont want it to reinvent the wheel when a similar task is asked

Thank you! I love this series.

This was a great video. You're a little ahead of the crowd on this one so I bet this video stacks up views in teh months to come.

My conclusion on LangChain is the same as yours: very very powerfull tool, but kind of hard to understand because it can do so many things in so many ways (I've tested it for RAG implementation). Did you try Flowise? It's a "low code" interface to create flows in Langchain (and others). Just a suggestion for your videos: when you explain code, please zoom your IDE and collapse the terminal panel. It will be more readable for the viewers. Thank you for you videos.

That's the kind of benchmark I wanted to make. Thank you for avoiding me the pain of reading a ton of documentation 👍 ps: you deserve more views

I'm following A.I closely and I found this website a few months ago, I think it's really cool what it can do and I can finally start on my childhood dream to create my own games without any coding knowledge. Only downside is both of my games I have been working on just totally breaks after a while, I always seem to come to a point where tha a.i simply can't add anything else without breaking the rest of the game's functions. And bow recently the context menu(right click on PC) for mobiles has dissapeared, I always find the main prompt bar at the top gets progressively worse the longer you work on a project so I always use the edit function instead

Beautiful nice idea, 😊 Keep going ❤👍

Good tutorial, Nice work....... Please add Ollama Integration as another option to new videos about crewai 🙏

Awesome ❤🙏 Very great idea and tutorial Please, Go Go Go for more 🎉

Nice content, thx!