Apple picks Google's Gemini to run AI-powered Siri coming this year (4 minute read) Apple and Google have formed a multiyear partnership to power Apple's AI features with Google's Gemini model. Apple's future foundation models will lean on Google's cloud technology. Models will continue to run on Apple devices and the company's private cloud compute. The terms of the deal have not been made public, but in August, Apple was reportedly planning to pay about $1 billion a year to utilize Google's AI. | Meta Creates High-Powered Team to Oversee AI Infrastructure Buildout (4 minute read) Meta Compute is a new initiative by Meta aimed at securing the enormous amounts of computing power the company's AI models need. Meta plans to build tens of gigawatts this decade as it races to develop superhuman untell agency. The company named Dina Powell McCormick as the new president and vice chair on Monday. Powell McCormick will work on developing partnerships with governments to finance and deploy data centers across the world. | | Science & Futuristic Technology | Nvidia to Invest $1 Billion in AI Drug Laboratory With Eli Lilly (4 minute read) Nvidia plans to invest $1 billion over five years in a new laboratory with Eli Lilly. The facility, to be built in Silicon Valley, will bring Lilly's lab expertise close to the center of AI innovation. It will help AI engineers learn what it takes to run lab equipment and research tasks. Nvidia plans to make its suite of AI models and agents tailored for the healthcare industry open source, so anyone will be able to tinker with the technology and adapt it to their needs. | 1X unveils 1XWM world model for NEO robot platform (1 minute read) 1X has integrated its new video-pretrained world model, 1XWM, into its NEO robot platform. The 1XWM model predicts robot actions by generating text-conditioned video rollouts and then translating them into motion commands through an Inverse Dynamics Model. It uses a 14B parameter generative video model as a backbone, with inference currently taking about 11 seconds per rollout. The method enables faster adaptation to new tasks as it doesn't require tens of thousands of robot demonstration hours. | | Programming, Design & Data Science | Why We Built Our Own Background Agent (17 minute read) Inspect is a coding agent that closes the loop on verifying its work by having all of the context and tools needed to prove it. It can run tests, review telemetry, and query feature flags, and it visually verifies its work and gives users screenshots and live previews. Each session runs in a sandboxed VM with everything an engineer would have locally. This post contains the spec for Inspect so anyone can replicate the tool. | First impressions of Claude Cowork, Anthropic's general agent (8 minute read) Claude Cowork is a general agent with a UI now available as a research preview to Max subscribers as part of the updated Claude Desktop macOS application. It looks very similar to the desktop interface for regular Claude Code. The general agent is designed to bring the powerful capabilities of Claude Code to a wider audience with a less intimidating interface. Screenshots of the feature are available in the article. | | The AI-powered Veblen economy (9 minute read) We will still want human-sourced goods and services in the future. Humans have been shown to be incredibly creative when manufacturing scarcity for the purpose of playing status games. This ability is increasing, and the trend is accelerating. Humans will be able to create enough scarcity to employ people to produce artisan goods. | The AI revolution is here. Will the economy survive the transition? (37 minute read) This post contains a conversation between Michael Burry, Dwarkesh Patel, Patrick McKenzie, and Jack Clark on the AI revolution and its impact on the economy. Burry, who called the subprime mortgage crisis when everyone else was buying in, is skeptical of the trillions powering into AI infrastructure. Clark is one of Anthropic's co-founders, and McKenzie and Patel have interviewed many people across the industry. | | What is an Engineer? (22 minute read) Engineers make the right observations, make the right decisions, and take the right actions, in the right sequence, to accomplish the right intentions. | | | Love TLDR? Tell your friends and get rewards! | | Share your referral link below with friends to get free TLDR swag! | | | | Track your referrals here. | | | |
No comments:
Post a Comment