The XY Space blogAI, funding & shipping

Field notes

Field notes on building,shipping, and funding AI.

Practical writing from the XY Space team. The Academy is our category of hands-on tutorials for non-developers building production-grade systems, alongside notes on funding, hiring, and the craft of shipping production AI.

All posts

01AI Explainers

How to Reduce LLM Hallucination

Hallucination is an incentive and system-design problem, not a mysterious glitch. You reduce it by grounding, constraining, verifying, and measuring, not by waiting for a perfect model.

Jul 14, 202610 min read

02AI Comparisons

Loop Engineering vs Subagent-Driven Execution

Two ways to push an agent past a single pass: keep one context iterating in a loop, or fan the work out to fresh subagents. They optimize for depth versus breadth.

Jul 14, 20269 min read

03Guides

Migrating from Claude Code to Codex

A practical map for moving from Claude Code to OpenAI's Codex CLI: what the permission flags, headless mode, skills, and config actually translate to, and what you don't need to migrate at all.

Jul 14, 202611 min read

04AI Comparisons

Superpowers vs Opusplan

Superpowers and opusplan both aim at better output from Claude Code, but one is a disciplined workflow and the other routes a smarter model to planning.

Jul 8, 20269 min read

05AI Explainers

What is claude -p?

claude -p runs Claude Code without the interactive terminal: it takes a prompt, does the work, prints the result, and exits, which makes it scriptable.

Jul 8, 20268 min read

06AI Explainers

What is LLM-as-a-judge?

LLM-as-a-judge uses one language model to score another's output against a rubric, scaling quality evaluation that exact-match checks cannot express.

Jun 28, 20269 min read

07AI Explainers

What is an AI agent?

An AI agent is a system that uses a language model to decide and act in a loop, calling tools to accomplish goals rather than only generating text.

Jun 25, 20269 min read

08AI Explainers

What is RAG (retrieval-augmented generation)?

RAG retrieves relevant documents at query time and feeds them to a language model, grounding answers in your data instead of pretrained memory.

Jun 22, 20269 min read

09AI Explainers

What is human-in-the-loop AI?

Human-in-the-loop AI keeps a person in control of consequential decisions, letting an agent do the work while a human approves anything irreversible.

Jun 19, 20269 min read

10AI Explainers

What is an agentic workflow?

An agentic workflow blends deterministic steps with model-driven decisions, giving an AI freedom where judgment is needed and structure everywhere else.

Jun 16, 20269 min read

11AI Comparisons

CrewAI vs LangGraph

CrewAI and LangGraph both build agent systems, but one models a crew of role-based agents while the other models explicit stateful graphs.

Jun 13, 20269 min read

12AI Comparisons

n8n vs custom AI agents

n8n offers fast visual automation with AI nodes; a custom AI agent trades that speed for control, testability, and production durability.

Jun 10, 20269 min read

13AI Comparisons

OpenAI Agents SDK vs LangGraph

The OpenAI Agents SDK offers a lightweight, fast path to agents; LangGraph offers explicit graph-based control for stateful, durable workflows.

Jun 7, 20269 min read

14AI Comparisons

MCP vs A2A

MCP and A2A solve different interoperability problems: one connects agents to context and tools, the other connects agents to agents.

Jun 4, 20268 min read

15AI Comparisons

LangGraph vs Pydantic AI

LangGraph and Pydantic AI both help build agent systems, but one centers graph orchestration while the other centers typed Python agents.

Jun 1, 20269 min read

16AI Comparisons

Mastra vs Flue

Mastra and Flue are TypeScript-first agent frameworks, but they emphasize different paths to durable, production-grade AI systems.

May 29, 20269 min read

17AI Comparisons

LangGraph vs LangChain

LangGraph and LangChain are related but not interchangeable; one focuses on stateful graph orchestration, the other on LLM app building blocks.

May 26, 20269 min read

18AI Explainers

What is LangGraph?

LangGraph is a graph-based framework for building stateful, controllable agents and long-running AI workflows.

May 23, 20268 min read

19AI Explainers

What is Mastra?

Mastra is a TypeScript agent framework for building AI agents, workflows, memory, evaluations, and production-oriented AI applications.

May 20, 20268 min read

20AI Explainers

What is AG-UI?

AG-UI is a protocol for connecting AI agents to user interfaces through events, state updates, messages, and tool-call streams.

May 17, 20268 min read

21AI Explainers

What is an MCP server?

An MCP server exposes tools, resources, and prompts to AI clients through a standard protocol instead of one-off integrations.

May 14, 20268 min read

22AI Explainers

What are agent tools?

Agent tools are the functions, APIs, data sources, and actions that let AI systems move from conversation to controlled execution.

May 11, 20268 min read

23AI Explainers

What is the A2A protocol?

A practical guide to the Agent2Agent protocol, how A2A lets agents discover capabilities, exchange tasks, and coordinate across systems.

May 8, 20268 min read

24Funding

Hiring students in Ontario in 2026 is cheaper than you think

Three stackable Ontario and federal programs (SWPP, the Co-operative Education Tax Credit, and SR&ED) can cover well over half of a co-op student's salary. Here's the worked math.

May 5, 20268 min read

25Academy Video

Host your Lovable app on Cloudflare

Move your Lovable app off the default preview URL and onto Cloudflare, a fast, free, production-ready edge for your custom domain.

May 2, 20266 min read

New posts,the moment they ship.

Drop your email and we'll let you know when new courses and lessons drop. No spam, just the next release.

Loading form…