AI researchers are finding that today’s most advanced models are superhuman in some ways and stumbly in others. Take OpenAI’s new “o3” model, for example: it aced challenging business tasks in seconds but tripped over a simple children’s riddle. This paradox – uneven, “jagged” performance across tasks – has led Wharton professor Ethan Mollick to...
Author: Martin Treiber
Model Context Protocol Comparison: MCP vs Function Calling, Plugins, APIs
How does the Model Context Protocol (MCP) compare to other ways of connecting AI to tools? In this article, we break down how MCP stacks up against function calling, plugins, direct APIs, and agent frameworks. We’ll also explore real-world use cases for MCP and where this emerging standard might take us in the future. MCP...
CaMeL Prompt Injection Defense Explained
Prompt injection attacks pose a major threat to AI assistants, but a breakthrough system called CaMeL offers a powerful defense by design. Developed by researchers from Google, DeepMind, and ETH Zurich, CaMeL uses proven software security principles to protect large language models where others fail. Here’s how it works—and why it matters. The Problem: Why...
Model Context Protocol: Inside the MCP Architecture
The Model Context Protocol (MCP) is redefining how AI systems interact with external tools and data sources. Building on simple function calling, MCP introduces a powerful, modular client-server architecture that enables dynamic tool discovery, secure integration, and efficient session management. In this post, we’ll explore the core components of MCP, including lifecycle phases, primitives like...
A2A vs MCP: Comparing AI Standards for Agent Interoperability
As AI agents increasingly automate and enhance complex workflows, the technology landscape is seeing the rise of crucial interoperability standards. Google's Agent2Agent (A2A) protocol and Anthropic's Model Context Protocol (MCP) stand out as prominent initiatives that, while complementary, target distinct aspects of AI integration. Here we take a deeper look at both, examining their designs,...
Model Context Protocol (MCP): A New Standard for Tool and Data Integration
Large language models are evolving fast, but their real power comes from connecting to the outside world. Enter Anthropic’s Model Context Protocol (MCP)—a universal standard that makes it easier for AI to call external tools, access real-time data, and interact fluidly with services. In this first article of our series, we unpack what MCP is,...
The Rise of Vibe Coding and Hyper-Personalized Apps (N=1 Software)
Vibe coding is a new approach to software creation where you describe what you want in natural language and let AI generate the code – essentially coding by feel or “vibe” rather than explicit syntax . The term was coined by AI researcher Andrej Karpathy in early 2025 to describe “a new kind of coding...
The Shifting Landscape: How LLMs Are Becoming the New Content Interface
In a thought-provoking statement on X, former OpenAI researcher Andrej Karpathy has painted a vision of the future where large language models (LLMs) become the primary interface for digital content. This perspective suggests a fundamental shift in how we create, structure, and optimize information in the coming years. The 99.9% Optimization Shift While Karpathy acknowledges...
Vibe Coding: A Successor to literate Programming?
Literate programming and vibe coding represent two distinct paradigms in software development, emerging in different eras with different motivations. Literate programming was introduced by Donald Knuth in the 1980s as a way to write programs as works of literature – focusing on explanation and human readability over machine-oriented structure (Donald Knuth). In contrast, vibe coding...
Vibe Coding in Software Engineering
Vibe coding is a programming approach where developers rely on intuition and an organic “flow” state rather than formal planning or strict methodologies. In practice, this often means jumping straight into coding with minimal upfront design or documentation. Advocates argue that this free-form style can boost creativity and speed, but it also raises concerns in...
