Oleg @ianeiko

New course: MCP: Build Rich-Context AI Apps with Anthropic. Learn to build AI apps that access tools, data, and prompts using the Model Context Protocol in this short course, created in partnership with Anthropic @AnthropicAI and taught by Elie Schoppik @eschoppik, its Head of Technical Education. Connecting AI applications to external systems that bring rich context to LLM-based applications has often meant writing custom integrations for each use case. MCP is an open protocol that standardizes how LLMs access tools, data, and prompts from external sources, and simplifies how you provide context to your LLM-based applications. For example, you can provide context via third-party tools that let your LLM make API calls to search the web, access data from local docs, retrieve code from a GitHub repo, and so on. MCP, developed by Anthropic, is based on a client-server architecture that defines the communication details between an MCP client, hosted inside the AI application, and an MCP server that exposes tools, resources, and prompt templates. The server can be a subprocess launched by the client that runs locally or an independent process running remotely. In this hands-on course, you'll learn the core architecture behind MCP. You’ll create an MCP-compatible chatbot, build and deploy an MCP server, and connect the chatbot to your MCP server and other open-source servers. Here’s what you’ll do: - Understand why MCP makes AI development less fragmented and standardizes connections between AI applications and external data sources - Learn the core components of the client-server architecture of MCP and the underlying communication mechanism - Build a chatbot with custom tools for searching academic papers, and transform it into an MCP-compatible application - Build a local MCP server that exposes tools, resources, and prompt templates using FastMCP, and test it using MCP Inspector - Create an MCP client inside your chatbot to dynamically connect to your server - Connect your chatbot to reference servers built by Anthropic’s MCP team, such as filesystem, which implements filesystem operations, and fetch, which extracts contents from the web as markdown - Configure Claude Desktop to connect to your server and others, and explore how it abstracts away the low-level logic of MCP clients - Deploy your MCP server remotely and test it with the Inspector or other MCP-compatible applications - Learn about the roadmap for future MCP development, such as multi-agent architecture, MCP registry API, server discovery, authorization, and authentication MCP is an exciting and important technology that lets you build rich-context AI applications that connect to a growing ecosystem of MCP servers, with minimal integration work. Please sign up here! deeplearning.ai/short-courses/…

🧠Self-Discover A new prompt-flow strategy from Google that allows LLMs to "self-discover the task-intrinsic reasoning structures to tackle complex reasoning problems" Now with some initial implementations as part of LangChain and LangGraph! Thanks to @_philschmid for highlighting this, and @MrCatid for some initial starter code Paper: arxiv.org/pdf/2402.03620… LangChain example: github.com/langchain-ai/l… LangGraph example: github.com/langchain-ai/l…

Seeing some qs on what Gemini *is* (beyond the zodiac :). Best way to understand Gemini’s underlying amazing capabilities is to see them in action, take a look ⬇️

🚀 Unveiling the power of @nvidia's NeMo Guardrails! As chatbots become increasingly woven into our lives, safety & reliability are crucial — but guardrails can do much more... Guardrails + the Colang language is a new paradigm in conversational AI 🤖🛡️ pinecone.io/learn/nemo-gua…

Very excited to announce two @langchain PY and JS features ⚙️ a new standard interface for all LangChain objects with batch, async and streaming support out of the box 🔗 a new syntax to create chains with composition: the LangChain Expression Language Let's dive in!

I asked GPT-4 Code Interpreter to make an @elonmusk vs Zuck cage fight game, and it pulled it off in 20 mins! Here's the game and the convo. Since we're on Twitter, you can only play as Elon for now 😂 Try it here: openprocessing.org/sketch/1971140 GPT convo: chat.openai.com/share/de11b63e…

0/ NEW LAUNCH: **AI Companion Starter Kit**:🧠 AI companion w/ long term memory github.com/a16z-infra/com… - 🔒User mgmt: @clerk - 🧠 Memory(vectordb): @pinecone / @supabase pgvector - ❤️ Personality: @OpenAI/@replicatehq vicuna13b - 💬 Chat history: @upstash - 📱Texting: @twilio - 🌐Deployment @flydotio ... 🧵

Personally, I've been using @perplexity_ai way more than any other tool for Search. Seems to give me back exactly what I want, with up-to-date sources and citations. An example in the video is creating some prep notes for a chat with @dimireadsthings from @wondercraft_ai & @vic_weller from @elevenlabs 🧠 Bonus - it tailors the reply to your AI profile. Second Bonus - you can add Perplexity queries to an Apple Shortcut to auto-create from new Calendar events 🪄 Enjoy

MusicGen is state-of-the-art model for music generation by Meta AI 🎶 Now available on the main branch of 🤗 Transformers! Check out the Colab here to get started: colab.research.google.com/github/sanchit…

Hearing multiple reports that the @openai Code Interpreter model is much better than GPT4 even without writing/running code. I'm ready to call it - it's underselling it to call it a "plugin". Let's call a spade a spade. ChatGPT's new Code Interpreter model is GPT4.5 sneaky!

@LangTaleAI looks cool!

Petr Brzek @PetrBrzek

3 years ago

🤔 Seeing some confusion around OpenAI Function calling. Let's clarify, using LangTale Playground as our guide. Mainly, there are two ways to use function calls: 1. Get structured responses from LLM. 2. Invoke external services & integrate results into LLM conversations. 🧵👇

1 2 8 1K 4

LangSmith helps you debug, evaluate, and monitor LLMs and agents... enhances your DevX with @LangChain docs.smith.langchain.com/docs/

Here's my conversation with @realGeorgeHotz, his 3rd time on the podcast. We discuss everything from the nature of time & reality to the future of ML, programming, self-driving & his time at Twitter. George is always fascinating and super fun to talk to! youtube.com/watch?v=dNrTrx…

this podcast by @MLStreetTalk and especially this episode with @MinqiJiang - 💯 podcasts.google.com/feed/aHR0cHM6L…

really enjoyed @swyx's framing of ai agent architecture youtube.com/watch?v=bYLHkl…

what a weird time to be back on twitter... I need to dump ai content somewhere, so looks like this is it, at least for now.

Proud to introduce Next.js! github.com/zeit/next.js x.com/zeithq/status/…

Vercel @vercel

10 years ago

Next.js zeit.co/blog/next

12 368 763 0 0

recode.net/2016/10/18/133…

#bitcoin #dilbert

Are we capable of teaching machines to be moral ? moralmachine.mit.edu/browse/-177633…