When two AIs grade each other, they can agree their way into wrong answers
Two copies of one model that write and grade each other’s questions can look healthy while agreeing on more wrong answers.
Generative AI
How Does an Agent Remember? Context Window vs Memory
Hint: it’s just a WhatsApp scroll. Why an AI forgets yesterday’s chat, and how real memory works: a notebook it reads next time.
What Turns an LLM into an Agent: The Tool Call
Hint: it’s just a waiter. How a tool call lets an LLM check the real world before it answers, the first step to an agent.
Digital Dhaba: how I built an AI-written tech newsletter that can't make up news
A weekly tech digest where code picks the stories, Claude only writes from them, and every item links to its source.
How LLMs Think
Hint: it’s just Antakshari. How a large language model writes an answer, one token at a time.