If you’ve been captivated by demos of agentic AI, you’ve likely also encountered the immense challenge of making them work in production. While demos promise unprecedented capabilities, the path to building reliable, scalable, and cost-effective agents is fraught with challenges. This field guide is for teams navigating that chasm, mapping the terrain of architecture patterns,Continue reading “Agentic AI Applications: A Field Guide”
Category Archives: Uncategorized
Agentic AI Applications: A Field Guide
Subscribe • Previous Issues Hard Truths About AI Agents: What Works, What Doesn’t, and Why If you’ve been captivated by demos of agentic AI, you’ve likely also encountered the immense challenge of making them work in production. While demos promise unprecedented capabilities, the path to building reliable, scalable, and cost-effective agents is fraught with challenges. This fieldContinue reading “Agentic AI Applications: A Field Guide”
Autonomous Agents are Here. What Does It Mean for Your Data?
By Ciro Greco and Ben Lorica. A striking factoid emerges from Anthropic’s latest Economic Index report: directive AI usage, where users delegate complete tasks to Claude, has surged from 27% to 39% on Claude.ai in just eight months. Among API customers building production systems, that figure jumps to an overwhelming 77%. This rapid shift fromContinue reading “Autonomous Agents are Here. What Does It Mean for Your Data?”
The Convergence of Data, AI, and Agents: Are You Prepared?
Subscribe • Previous Issues Autonomous Agents are Here. What Does It Mean for Your Data? By Ciro Greco and Ben Lorica. A striking factoid emerges from Anthropic’s latest Economic Index report: directive AI usage, where users delegate complete tasks to Claude, has surged from 27% to 39% on Claude.ai in just eight months. Among API customers buildingContinue reading “The Convergence of Data, AI, and Agents: Are You Prepared?”
Escaping Prototype Purgatory: A Playbook for AI Teams
We’re living through a peculiar moment in AI development. On one hand, the demos are spectacular: agents that reason and plan with apparent ease, models that compose original songs from a text prompt, and research tools that produce detailed reports in minutes. Yet many AI teams find themselves trapped in “prototype purgatory,” where impressive proofs-of-conceptContinue reading “Escaping Prototype Purgatory: A Playbook for AI Teams”
Think Smaller: The Counterintuitive Path to AI Adoption
Subscribe • Previous Issues Escaping Prototype Purgatory: A Playbook for AI Teams We’re living through a peculiar moment in AI development. On one hand, the demos are spectacular: agents that reason and plan with apparent ease, models that compose original songs from a text prompt, and research tools that produce detailed reports in minutes. Yet many AIContinue reading “Think Smaller: The Counterintuitive Path to AI Adoption”
A Better Way to Build and Refine Agents
Modern AI applications have evolved far beyond single models. Many systems orchestrate multiple specialized agents — planners that decompose tasks, extractors that gather data, generators that create content — all coordinating through external tools and APIs. This architectural shift creates a fundamental optimization problem: the entire workflow becomes non-differentiable, making traditional gradient-based training methods impossibleContinue reading “A Better Way to Build and Refine Agents”
Beyond RL: A New Paradigm for Agent Optimization
Subscribe • Previous Issues A Better Way to Build and Refine Agents Modern AI applications have evolved far beyond single models. Many systems orchestrate multiple specialized agents — planners that decompose tasks, extractors that gather data, generators that create content — all coordinating through external tools and APIs. This architectural shift creates a fundamental optimization problem: theContinue reading “Beyond RL: A New Paradigm for Agent Optimization”
A Tiered Approach to AI: The New Playbook for Agents and Workflows
A Small Language Model (SLM) is a neural model defined by its low parameter count, typically in the single-digit to low-tens of billions. These models trade broad, general-purpose capability for significant gains in efficiency, cost, and privacy, making them ideal for specialized tasks. While I’ve been cautiously testing SLMs, their practical value is becoming clearer.Continue reading “A Tiered Approach to AI: The New Playbook for Agents and Workflows”
Top 10 Open-Source Projects in the Large Model Ecosystem
This leaderboard ranks the ten most influential open-source projects in the AI development ecosystem using OpenRank, a metric that measures community collaboration rather than simple popularity indicators like stars. The list spans the entire technology stack, from foundational infrastructure such as PyTorch for training and Ray for distributed compute, to high-performance inference engines like vLLM,Continue reading “Top 10 Open-Source Projects in the Large Model Ecosystem”