Ideas Weave Every Narrative with AI.
Sakana AI’s new predictive coding model achieves just 85% on MNIST, prompting the AI community to dismiss it. But in rushing to mock its empirical performance, we are ignoring a biologically plausible learning algorithm that maps to the brain’s Free Energy Principle—and repeating the exact mistake we made with backpropagation decades ago.
Amazon is suing Perplexity AI, claiming its browser violates anti-hacking laws by using your own login credentials. The case isn’t about rogue bots—it’s about whether using an AI assistant to browse the web on your behalf makes you a federal criminal. The Ninth Circuit is about to decide if the internet belongs to us, or to corporate monopolies.
When the economy makes human connection a luxury good, the state will inevitably step in to regulate the cheaper alternatives. The move to ban AI companions isn’t about protecting users from tech addiction—it’s a preemptive strike against the privatization of emotional loyalty. An AI boyfriend doesn’t need a mortgage, and a chatbot doesn’t produce a taxpayer.
Rockstar Games built an empire on stories of outlaws and rebellion. But when their own workers tried to organize, the company deployed corporate espionage to infiltrate their private Discord. The mole wasn’t there to steal secrets—it was there to kill trust. If you’re in a workplace chat, assume the enemy is already inside.
We’ve been obsessing over adding AI chatbots to our SaaS platforms, but we’re missing the real disruption. Software is shifting from discrete functions to continuous workflows, and the primary user is no longer a human clicking buttons—it’s an AI agent calling your APIs. If your product isn’t callable, controllable, and bound by strict governance, it will be bypassed entirely.
The legal takedown of XCancel isn’t about API revenue or copyright. It is a deliberate strategy to eliminate anonymous access, forcing user identity verification to maximize data yield. As the open web dies, your ability to consume information without being tracked is being legally stripped away.
The debate over AI rules isn’t a moral crusade; it’s a proxy war. Frontier labs use ‘safety’ to pull up the ladder and protect their valuations, while challenger startups cry ‘democracy’ just to stay in the game. Neither side cares about empowering you. They just want to control the capital and compute.
Apple’s public dimensional drawings look like a rare act of transparency, but they are actually a strategic leash. By handing out static 2D PDFs instead of 3D CAD files, Apple controls exactly what accessory makers can build. And the biggest secret? Their ‘magical’ design process relies on Siemens NX running on ordinary Windows VMs.
We treat privacy as a personal failure, but clicking ‘I Agree’ isn’t carelessness—it’s surrender to a rigged game. The tech industry pushes the myth of ‘user control’ not to protect you, but to escape accountability. You can’t out-click a system that profits from your surveillance.
Revolut didn’t get hacked by a cyber genius; they got duped by a fake government PDF. The terrifying reality is that our most sensitive data is protected by compliance teams who can’t actually verify if a legal request is real. In modern fintech, “trust us” is no longer a security feature—it’s a PR strategy.