This week: we explore how engagement became the backbone of intelligence through OpenAI’s Sora; and how every prompt, replay, and regeneration quietly fuels the system itself.

AI in Action

The Quiet Business Behind OpenAI’s Flashiest Model [Multi-Modal AI/Engagement]

OpenAI’s new model, Sora, can turn a sentence into a scene, text into moving, cinematic video. A few words now produce sequences with steady camera pans, believable motion, and lighting that feels almost filmed. Text-to-video tools have existed for years, yet Sora crosses three thresholds at once: it sustains motion over time without warping, it understands how a camera should move through space, and it does all of this at scale using the same infrastructure that powers GPT-4. The result is a system that composes shots with story and physics intact. Within days of launch, Sora reached the number one spot on the U.S. App Store, a sign that imagination, once limited to text, is now something people expect to see.

Beneath the spectacle lies strategy. Sora is more than a technical feat, it’s part of OpenAI’s quiet shift from research breakthroughs to sustained engagement. Every prompt, edit, and replay teaches the model what people find believable or beautiful, turning interaction into training data. In doing so, OpenAI has entered the same attention economy it once claimed to transcend, one where engagement is currency, and every tap, replay, and reshare becomes a business signal. Sora is designed not only to inspire creation, but to hold attention, keeping users within OpenAI’s ecosystem just as Meta does with reels or TikTok with loops. The commercial path follows naturally: faster rendering, longer scenes, enterprise licenses, creative partnerships. What looks like experimentation today is market research in disguise, each prompt revealing what users value enough to pay for.

Sora’s realism outpaces the world’s ability to verify what’s real, widening the gap between creation and credibility. When video can be generated faster than it can be checked, imagination and evidence begin to overlap. The same energy-intensive systems that could train surgeons or visualize climate risks could just as easily flood the web with convincing but meaningless content, what critics call AI slop. If OpenAI’s goal is to capture imagination, it must also protect attention—the scarcest resource its models now compete for. The real test of Sora’s legacy will be how it’s used: to illuminate or to overwhelm. If it succeeds, it could make visual storytelling as universal as language itself; if it fails, it will remind us how easily creativity turns into noise

Bits of Brilliance

Engagement as Infrastructure [AI Economics]

In the generative era, engagement is not a metric, it is the machine. Every time you ask a question, refine a prompt, or tweak an AI-generated clip, you are not just using the system, you are teaching it. Those micro-interactions flow back into the model, sharpening its sense of what looks realistic, helpful, or creative. When millions of people correct phrasing, adjust lighting, or retry a scene, systems like Sora quietly learn what “good” looks like. The more we engage, the smarter these models become. Activity has become the new infrastructure of intelligence.

The irony is that engagement was once the goal of social platforms, but now it is the engine of intelligence itself. The same patterns that kept people scrolling are now teaching machines how to think, write, and create. In this new feedback loop, attention is no longer monetized through ads, but through learning. Each prompt becomes both instruction and investment.

But the loop has limits. More engagement brings more noise, and not all data makes a model wiser. The task ahead is learning to engage with intent, to know when our curiosity is adding clarity and when it is simply feeding the machine. Each interaction is a vote for the kind of intelligence we want to shape. Maybe the future of AI won’t be decided by scale alone, but by how thoughtfully we choose to participate. The most powerful thing we can do is not simply engage more, but engage better.

Curiosity in Clicks

How to train the trainer [chatbot/experiment]

This week, try a small experiment in co-training.

Pick your favorite AI tool, ChatGPT, Claude, Gemini, whichever you use most, and ask it to write a short paragraph in your voice. Then, spend five minutes refining it: adjust the tone, swap phrases, and tell it what sounds more natural to you. Do this three times.

By the third round, you’ll notice something uncanny, it starts to sound more like you. You just trained your own mini-model. Every bit of feedback, every nudge toward “better,” teaches the system what your version of good looks like. That’s engagement as infrastructure in action.

Bonus Click: Ask your chatbot to remember your writing style for future drafts. You’ll start to notice how much faster it learns from you, and how much you’re training it back.

_{Byte-Sized Intelligence is a personal newsletter created for educational and informational purposes only. The content reflects the personal views of the author and does not represent the opinions of any employer or affiliated organization. This publication does not offer financial, investment, legal, or professional advice. Any references to tools, technologies, or companies are for illustrative purposes only and do not constitute endorsements. Readers should independently verify any information before acting on it. All AI-generated content or tool usage should be approached critically. Always apply human judgment and discretion when using or interpreting AI outputs.}