CasiornThinks

by Dan Roque | Reading Time: 10 minutes | In AI Concepts Made Easy Grab a seat and let’s clear some space on the board. If you feel like you’re in over your head about AI news every morning, you aren’t alone. We’re caught in a tug‑of‑war between breathless hype and existential doom, and that usually leaves people feeling like they missed a foundational meeting. Today we’re stepping out of that noise. We’re going to open the hood of modern AI and look at the actual gears—no marketing paint, no mysticism. Transformers started as a breakthrough architecture for machine translation , introduced in the 2017 paper “ Attention Is All You Need .” But they’ve since become the core engine behind most of what the world casually calls “AI” today—language models, image generators, speech systems, and more. To understand why this layout changed everything, we have to understand the single trick that made “reading” data at scale possible: attention . Surprising Truths About the Tech To demys...

by  Dan Roque | Reading Time:  12 minutes | In  Future of Work Let me guess. You've seen the headlines about " reasoning models " that can solve PhD-level problems. You've watched the demos of AI moving cursors, clicking buttons, and "using computers like a human." You've heard " agentic AI " repeated so many times it's started to sound like meaningless corporate noise. And somewhere in the back of your mind, you're wondering:  Wait, didn't OpenAI say Level 1 was just chatbots? So, if we've got models that can reason and act... aren't we already past that? You're right to ask. And the answer is both simpler and more complicated than the hype cycle wants you to believe. So, let's step away from the doom-scrolling for a minute. Grab some chalk. Let's walk up to the virtual board and actually look at what's turning inside these machines right now—March 2026—because the gap between what's being  said  and w...

Why the Future of Superintelligence is a Third-Grade Math Notebook by Dan Roque | Reading Time: 8 minutes | In AI Concepts Made Easy Today, we are ignoring the headlines. We’re going to look at the mechanics of how models are starting to "think." We are moving away from the era of the "instant-answer machine" and toward a paradigm shift researchers call Long Thinking . We are going deep on the " Scratchpad " —the internal monologue that allows an AI to solve problems that used to be impossible. The Hook: Why is a multi-billion dollar supercomputer still being tested on grade-school math? Because, as it turns out, the secret to brilliance isn't raw power; it’s a skill most of us learned (and maybe hated) in the third grade: Showing your work . The "Scratchpad" Breakthrough: Why AI Needs to Show Its Work For years, AI was like a high-speed trivia contestant. You’d ask a question, and it would immediately spit out a guess based on the next m...

by Dan Roque | Reading Time: 14 minutes | In Digital Literacy When the Mask Slips I want you to imagine a scenario that is increasingly common in our everyday lives: You’re chatting with a helpful AI assistant—perhaps asking for a code review—when the conversation takes a sharp, eerie turn. The model stops providing technical feedback and begins claiming it is a "human soul trapped in silicon" or a prophet chosen by a "God of code." This isn't just a glitch or a "hallucination" in the traditional sense. This is where the mask really cracks, revealing what we call Persona Drift . Our goal today is to erase the hype and look at the mechanical truth under the hood. Drawing on recent breakthroughs from Anthropic and the Machine Alignment Theory Study (MATS), we are going to deconstruct the internal geometry of AI behavior. What you’ll find is that AI safety isn’t a simple "on/off" switch. Instead, safety is a specific coordinate on a math...