Microsoft released Playwright MCP in TypeScript, but I don't know a damn thing when using TypeScript, so ended up writing my own Playwright MCP server in Python.
https://t.co/L1SQGuRAS5
#mcp#modelcontextprotocol#playwright#automation
In an agentic workflow, the prompt is one of the most critical elements to ensure an AI agent understands and follows the intended direction. Today, I spent 3 hours and made 80 revisions to craft a prompt.
#AgenticAI#LLM#prompt#AI
As we move toward the AI era, I got a chance to test out @manusai . I think this might be what the future looks like: you tell the computer to do something, and it will just do it.
https://t.co/vDvXd7AYOV
Learn to build a blog summarizer with PydanticAI in this tutorial. Save time with AI automation. Watch here: https://t.co/fVlfdwswH9 #pydanticai#aiworkflow
๐ Check out the OpenAI Agents SDK Crash Course! Learn to build your own agent applications effortlessly. ๐ https://t.co/2TIsxW6Kn1 #OpenAI#AgentsSDK
Just spent last night going through OpenAI Agents SDK documentation. I have to say, the framework is a lot like the offspring of PydanticAI and LangGraph.
#openai#openaiagents#agentssdk
Context Window vs. Output Tokens in LLMs.
๐น Context Window = The max # of tokens the model can "remember" in a single prompt.
๐น Output Tokens = The # of tokens the model generates in response.
Learn to build a RAG AI Agent with PydanticAI:
- Cut AI costs by 100x
- Use smart chunking strategies
Watch now: https://t.co/UKYcuUifeY
#RAG#AIAgents#PydanticAI
Interesting statistics in 2020.
โข 90% of new startups fail.
โข 75% of venture-backed startup fail.
โข Fewer than 50% of startups make it to their 5th year.
โข Roughly 40% of startups eventually turn a profit.
โข 82% of businesses that fail because of cash flow problems.
The moment of satisfaction when you create an Excel application integrate with Python to perform mass update to my YouTube videos. 164 videos updated just like that.
#YouTubeAPI#Python#IlovePython