Do you want to join my team? Apply here: Access my code, templates and prompts in my community: Follow me on Twitter – My Google Colab: Andrei Karpati Talk: Sam Altman for GPT-5: Please subscribe. Credits: @maya-akim @matthew_berman @BuildNewThings AI Agents are the next big thing β and how to build your own AI agent with Crew AI.
π» All the code, prompts, resources are now available in my community. Join now! https://www.skool.com/new-society
Hi, I'm.a creator in Vermont. For months I've been working to simulate a 5000 person corporation, but it is just me and cascading LLMs π
got it to work! thanks!
People need to stop over complicating these by calling them "agents" all that an "agent" is, is basically creating multiple zapier "chat gpt" steps and giving them different prompts to achieve the desired result.
Lol. "System 1 thinking". Thats what Zapier is for bro.
When it comes to coding, you rush ahead far too quickly. No beginner understands what you're doing and why you're doing it that way. So your statement that even a beginner could learn from you is completely wrong. Try to calm down a bit more before recording. If, for example, as a beginner I have to rewind over and over again because I get the feeling that you're skipping over things that are clear to you but not to me, then it's no fun to watch because you're taking up a lot of my time unnecessarily.
Dude are you the voice behind 2 minute papers?
I think the most of the companies are facing problems in the work it is better to use AI Agent that is SmythOS Ai Agent !
The future AI Agent was the SmythOS AI agent it Simplicity gives best results!
Can someone just give a working definition of what is an AI Agent and then give their talk?
This is pretty recent though in running today using colab, seeing Errors due to dependency:
Building wheel for pypika (pyproject.toml) … done
ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
tensorflow-metadata 1.15.0 requires protobuf<4.21,>=3.20.3; python_version < "3.11", but you have protobuf 4.25.4 which is incompatible.
Excellent content bud.
PSA: If you have an error downloading crewai with pip, try adding –use-deprecated=legacy-resolver at the end of the pip install crewai command. this resolved the dependency error for me
so we have to learn out own way to program the ai to get it to work for us. seems backwards
Thanks. This is the first video I've found that clearly explains what AI agents are and how to make one
"form factor"
Hey David i am from India & have no IT background. Can i start an AAA or SaaS?
Before I start, I have used ai agents, and I just see the agent is used for normal chat of Claude or ChatGPT.
I hope this video can show how agents does different.
Guys, does anyone know what to do from the minute 20:19 its terrible how fast from this minute and opening .env file I got lost after I started the course from the beginning.. I like what he does though… but I need help in steps- can someone help?
You really smart man!
I think you mean 7p's (prior proper planning prevents piss poor performance!)
I am getting 'insufficient_quota' errors from ChatGPT. Tried switching the model to "gpt-3.5-turbo", which is supposed to be in the free tier, but the error remains.
May have to pay for ChatGPT to get this to work.
What is the minimum pc requirement for this kind of stuff
Thankyou algorithm for this! Great vid btw!
I'm also interested in Gemini/Google AI agents, as Google offerings can always search the internet for current news/research.
[Edit:
OMG!
Why can these agents get current news?
I got a 1st July 2024 paper summarisation on 17th July 2024.
Serper?
Yes. It does Google Searches.
Nice one David et al.]
API keys:
Colab doesn't allow .env any more because it doesn't like things starting with ".".
There's another option with Colab Secrets, which is apparently more secure.
That's the key on the left side near where you upload files.
You put in key:value pairs for the API keys like this:
OPENAI_API_KEY weor82989io3nro1k
Then the llm line of the agent becomes
llm = ChatOpenAI(model_name="gpt-4-turbo", temperature = 0.2, π¨π©ππ§ππ’_ππ©π’_π€ππ²=ππππππ_πππ_πππ)
I'm trying to get this working.
It doesn't like quotes around the API key, so weor82989io3nro1k as the API key
NOT 'weor82989io3nro1k'.
This is not a real API key, just random stuff I typed in for an example.
Why am I getting a error when I try to upload the .env file. I try to drag it and upload but doesn't show up. When I try to load it, it says cannot find file. What am I doing wrong?
I have a weird problem. It seems task1 is recognized but task2 is not. Found it. stupid typo.
Absolute NOOB here (I learned on a 086 processor – 5 1/4 floppy in university!) – Question – can AI Agents literally access your systems (virtual mouse/ keystrokes) at your command? So they can log into your apps and enter data/ navigate command fields etc?)
Sorry if dumb question! I just stumbled on this video, and can't help myself – just ordered a new computer powerful enough to allow me to make my own AI Agent… Would like to know limitations (such as auto keystrokes and taught app navigation etc).
Thanks mate, greetings from Australia
17:47 thank me laterπ
I need your help
Google colab is not letting me upload the .env file. It just won't show up are there any alternatives?
what is the smartest ai agent that an electronic musician can create who wants to get 10,000,000 monthly listeners on Spotify