Antonello Scalmato

Where are all the agents? We were promised autonomous intelligence and we only got chatbots and glorified workflows sprinkled with LLMs. Where is the autonomy if I still have to account for all possible edge cases? Why do I still have to deal with APIs, spaghetti logic, and esoteric protocols? Why can’t I just describe the outcome I want in plain english and the computer will figure out all the details? These are the questions we’ve been asking ourselves at Latitude for the past few months. Today we’re introducing a new way to build automation software. Building for autonomy requires upgrading your mental model of what software is. All you need now are prompts and tools. Code is out, the LLM is the backend now. Latitude is the first full-stack agent engineering platform built around this idea. It is a code editor but you won’t see any code, only plain-english instructions. We have all the advanced tools you need to design flexible behavior: agent orchestration, tool calling, evaluations, and monitoring. And the best part is you don’t even have to worry about any of it. We’ve built an agent—Latte!—that takes care of everything for you. You don’t have to be an expert to program computers anymore. We’re bringing joy to building software, and this is just the first step.

120

9 commenti

Fun fact – I used to be a C++ developer... but I wasn’t all that good at it. Making a UDF library for SVGs feels very familiar somehow. The other day I got a nice question: “Who in a real business can actually build and maintain these UDF + SVG + IBCS solutions? Business teams don’t have the time or skills for this day-to-day. So how can we implement these brilliant ideas fast, without them staying just great showcases?” Here’s my take: • Not every analyst should be writing complex DAX UDFs. • Focus on libraries and templates. Once a function is built, it should be reusable by anyone with drag-and-drop simplicity. • Think of it like Excel add-ins: someone develops the core functions, but used by anyone. • In Power BI, this means central teams (or external consultants) package the complexity → and business users only swap in their measures. • The real win is standardization: you get IBCS-style visuals, SVG stat cards, or KPI tables without reinventing the wheel every report. So yes – building the engine is specialized work. But using it can be as simple as “Current Year, Past Year → done.” That’s the bridge from showcase to operations. Here's how many lines my current library is - and I'm still trying to make "standard" functions, and "specialized" functions - i.e. super simple ones where you just put in the 3 measures you need, and ones tons of parameters to get everything you need to. Also thinking about how to use C# scripts to get global values fixed for you in one go.

163

7 commenti

MCP Primitives in a nutshell 👇 This is probably the simplest and easiest way to clearly understand MCP in a single image. 𝗠𝗖𝗣 𝗱𝗲𝗳𝗶𝗻𝗶𝘁𝗶𝗼𝗻: It's an open standard developed by Anthropic to allow AI systems, particularly LLMs to seamlessly integrate with external data sources and tools. Think of it as a universal interface that enables agents to access and utilize information from various systems, enhancing their capabilities beyond their initial training Start with an 𝗠𝗖𝗣 𝗦𝗲𝗿𝘃𝗲𝗿. It can expose: 🧰 Tools – Let your agent connect to external services (e.g., APIs). 👥 Resources – Extra context like docs, DBs, or any data your app can use. ✍️ Prompts – Instructions to guide the model. Today, most MCP setups only support Tools, but Resources and Prompts are coming fast. What’s cool? MCP is bidirectional—the server can talk back to the client! On the 𝗖𝗹𝗶𝗲𝗻𝘁 𝘀𝗶𝗱𝗲, you can expose: 🧩 Sampling – Ask the app to generate completions using its LLM. 🗂️ Roots – Give the server access to local files. 📧 Elicitation – Collect user input when needed (like an email). MCP keeps evolving, along with a huge adoption of AI agents and it’s getting seriously powerful. #ai #agents #genai #mcp

171

12 commenti

🔵 𝐇𝐨𝐰 𝐭𝐨 𝐈𝐦𝐩𝐥𝐞𝐦𝐞𝐧𝐭 𝐀𝐈 𝐢𝐧 𝐘𝐨𝐮𝐫 𝐂𝐨𝐝𝐞 𝐔𝐬𝐢𝐧𝐠 𝐒𝐞𝐦𝐚𝐧𝐭𝐢𝐜 𝐊𝐞𝐫𝐧𝐞𝐥 – 𝐢𝐧 𝐔𝐧𝐝𝐞𝐫 𝟑𝟎 𝐌𝐢𝐧𝐮𝐭𝐞𝐬 If you’re a .NET developer, you’ve probably noticed that most AI frameworks are built for Python. LangChain, AutoGen, CrewAI, they’re all Python-first. But what about us, the C# developers who want to add AI capabilities to our applications? That’s where Semantic Kernel comes in… Semantic Kernel is a open-source SDK that brings AI orchestration to .NET. It’s not just another wrapper around OpenAI’s API. It’s a complete framework for building AI-powered applications with the patterns and practices we already use in our .NET projects. The killer feature? You can run it 100% locally with Ollama. Zero API costs. Zero latency. Total control. In my latest blog, I show you EXACTLY how to do it. Link is in the comments! 👇 💬 Are you already using Semantic Kernel or still hardcoding API calls? ❤️ Like if you're ready to level up your AI game! 🔁 Repost to save a fellow .NET dev from AI integration nightmares! #SemanticKernel #DotNet #CSharp #AI #ArtificialIntelligence #MachineLearning #LLM #OpenAI #Ollama #CleanCode #BackendDevelopment #AspNetCore #Microsoft #AIIntegration #SoftwareArchitecture

63

25 commenti

LLM Frameworks are evolving fast, but which one should you actually use? If you’ve been building with AI, you’ve probably heard of: • 𝗟𝗮𝗻𝗴𝗚𝗿𝗮𝗽𝗵 → graph-driven, great for stateful multi-agent apps • 𝗟𝗮𝗻𝗴𝗖𝗵𝗮𝗶𝗻 → developer-friendly, focused on chains, tools, and memory • 𝗔𝘂𝘁𝗼𝗚𝗲𝗻 → a multi-agent system where AI agents collaborate through dialogue • 𝗖𝗿𝗲𝘄𝗔𝗜 → role-based orchestration to coordinate agents toward shared goals Each framework has its own strengths, and choosing the right one depends on what kind of AI app you’re building. ✅ Need 𝘀𝘁𝗮𝘁𝗲 𝗺𝗮𝗻𝗮𝗴𝗲𝗺𝗲𝗻𝘁 + 𝗽𝗮𝗿𝗮𝗹𝗹𝗲𝗹 𝗹𝗼𝗴𝗶𝗰 → go with 𝗟𝗮𝗻𝗴𝗚𝗿𝗮𝗽𝗵 ✅ Want a 𝗱𝗲𝘃𝗲𝗹𝗼𝗽𝗲𝗿 𝘁𝗼𝗼𝗹𝗸𝗶𝘁 with prompt chains → use 𝗟𝗮𝗻𝗴𝗖𝗵𝗮𝗶𝗻 ✅ Working on 𝗮𝘂𝘁𝗼𝗻𝗼𝗺𝗼𝘂𝘀 𝗮𝗴𝗲𝗻𝘁 𝗰𝗼𝗻𝘃𝗲𝗿𝘀𝗮𝘁𝗶𝗼𝗻𝘀 → try 𝗔𝘂𝘁𝗼𝗚𝗲𝗻 ✅ Building 𝘁𝗲𝗮𝗺-𝘀𝘁𝘆𝗹𝗲 𝗔𝗜 𝘄𝗼𝗿𝗸𝗳𝗹𝗼𝘄𝘀 → explore 𝗖𝗿𝗲𝘄𝗔𝗜 I’ve broken down the 𝗰𝗼𝗿𝗲 𝘀𝘁𝗲𝗽𝘀 𝗼𝗳 𝗲𝗮𝗰𝗵 𝗳𝗿𝗮𝗺𝗲𝘄𝗼𝗿𝗸 in a simple graphic (below) to help you see the differences clearly. Which framework do you think will 𝗱𝗼𝗺𝗶𝗻𝗮𝘁𝗲 𝗶𝗻 2025? Repost for others who want to understand this. #AIModels #MachineLearning #DeepLearning #CloudComputing #AITechnology #DataScience #Robotics #SmartSolutions #TechInnovation#AIResearch#AIForGood #Automation #AICommunity #Python #TCS ##MultiAgentSystems #GenerativeAI #Analytics #DataScience #IoT #EnterpriseAI #AgenticAI #A2A #MCP #AutonomousAgents #AzureAI #GCpAI #AWSBedrock #Neurotechnology #VectorDatabases #AI #MachineLearning #RAG #SemanticSearch #DataEngineering #DataScience #SoftwareEngineering #azure #aws #Gemini #gcp #AITools #FutureOfLearning #AIGamechanger #SmartLearning #TechWithAI #AIForEveryone #NextGenLearning #LearnWithAI #AIRevolution #DigitalLearning

84

Quick thought- Aren't large context windows making more harm then benefit Because Lazy minds are stuffing them without "context engineering"🤔 This (Context Engineering) guide covers real-world context engineering: 1. Dynamic system prompts 2. Clear input/output structures 3. Dynamic tool integration 4. RAG, memory, and context management 5. Multi-agent coordination It’s not about better prompts—it’s about engineering robust, measurable context flows that reduce ambiguity and latency. Covers everything from system design to advanced concepts like context compression and evaluation pipelines. Guide here - https://t.co/qj4Cl12fao Thanks to Armand Ruiz for sharing and Elvis S.

38

1 commento

🚨 95% of apps will use LLMs(AI). We’ve officially entered the Software 3.0 era — where we don’t write code… we write prompts. And the implications go far beyond developers. 💥 🧠 From Code to Behavior Design: Software 1.0 → Handwritten code (if this, then that) Software 2.0 → Machine learning (train a model, set weights) Software 3.0 → Prompt-driven apps (talk to a model, shape outcomes) 💡 LLMs are becoming a new operating system: → Prompts = the new programs → Interface = natural language → Output = action, code, decisions In this carousel, I break down key ideas from Andrej Karpathy’s talk: ✅ How AI-native apps are being built today ✅ Why context, guardrails, and orchestration matter more than ever ✅ What "prompt engineering" really means for the future of work 📎 Full talk link in the first comment. Worth watching. 🗣️ Question for you: Which software era are YOU working in — 1.0, 2.0, or already 3.0? 👇 Let me know in the comments.

125

15 commenti