90eights👊🏾🇰🇪

A Chinese hardware team just mass-democratized AI agents. They took a 430,000-line AI assistant that needs a $599 Mac Mini and 1GB of RAM — and rewrote it in Go so it runs on a $9.9 dev board with less than 10MB of memory. Boot time: from 500 seconds to 1 second. Cost: from $599 to $9.9. Memory: from 1GB to 10MB. Same features: code generation, web search, Discord/Telegram chat, memory system, scheduled tasks, security sandbox. The wildest part? They claim 95% of the new codebase was written by AI agents themselves. The humans just guided the architecture. It's an AI assistant that literally rebuilt itself to be smaller. Launched February 9th. Four days later: 7,400+ GitHub stars. This is the pattern no one's talking about enough. Every AI capability that starts expensive gets commoditized within months. GPT-4 level models went open source in 6 months. Now the hardware floor for running a personal AI agent just dropped 60x in weeks. The infrastructure moat in AI isn't sustainable. The only defensible advantage is what you do with these tools — not access to them.

This Wait But Why illustration becomes more relevant each day in AI.

Tesla’s decision to focus on pure vision will go down as the most important strategic decision in their history, and it’s crazy that so few people have grasped that yet. Just go take a ride in both. The Tesla is smooth and human like. Once you’ve experienced it, you’ll start to notice how slow and jerky the Waymo is. The reason the Tesla drives better is because of the massive dataset enabled by computer vision sensors in every car. To the extent LIDAR is useful, you can always layer that on top of the pure vision system that is already safe enough to drive the car on its own. You’ll mop the floor with anyone who requires multiple sensor modalities and can’t drive if any of them go down. Waymo is giving millions of rides, but they’ve driven about 200 million miles. Tesla Self-Driving has been used for 8.5 billion miles, making it by far the most used self-driving software in the world. They’re also now operating without a driver in Austin. The goal isn’t millions, it’s trillions of miles traveled a year. Good luck getting there with a car that is 3x more expensive and has yet to turn a profit. Good luck scaling to South America, Asia, and Africa with a car that costs $75,000 and requires you to shoot out laser beams constantly. Less is more. Layer on whatever additional sensor modalities you want to the extent it improves anything, but people don’t realize how powerful it is to have a single end to end model that can drive more safely than a human even if all the sensors go out except the cameras. If the cameras go out you can’t drive the car anyway, even with LIDAR, because you need to be able to see traffic lights, signs, etc. Computer vision is both necessary and sufficient for fully autonomous driving. Adding anything extra as a requirement just holds back the technology from scaling and stopping many preventable deaths and injuries.