xAI raises $20 billion to expand Grok AI models and enterprise tools (1 minute read) xAI has successfully closed a $20 billion Series E funding round. The capital will be used to accelerate the development of the company's compute infrastructure. xAI has introduced several business-oriented offerings, including Grok Business, a team plan priced at $30 per seat per month, and Grok Enterprise, which includes Custom SSO and SCIM features. It has also released a Collections API that features built-in RAG, OCR, layout-aware parsing, and more. | How Google Got Its Groove Back and Edged Ahead of OpenAI (13 minute read) Google was behind in the AI race in August, but by September, its Gemini AI app had become the most downloaded app in Apple's App Store. The company launched its most powerful Gemini model yet two months later, leapfrogging over OpenAI to the front of the AI pack. Google's deep roots in science and research, willingness to invest in custom hardware, and leadership changes in recent years cleared the way for faster experimentation. The company's AI work has begun generating substantial revenue through search ads, paid versions of Gemini, and sales of new computer chips developed in-house. | | Science & Futuristic Technology | Commonwealth Fusion Systems installs reactor magnet, lands deal with Nvidia (3 minute read) Commonwealth Fusion Systems (CFS) has installed the first magnet in its Sparc fusion reactor, a demonstration device scheduled to turn on next year. The reactor will have 18 magnets in total in a donut-like shape. These magnets will create a powerful magnetic field for confining and compressing superheated plasma. If successful, the plasma will release more energy than it takes to heat and compress it. CFS is working with Nvidia and Siemens to develop a digital version of the reactor. | Dude, where's my supersonic jet? (16 minute read) Aviation innovation has been stuck in a rut in the 56 years since Concorde debuted due to noise, regulation, and cost. All three of these issues are at various stages of being solved. Regulatory walls are cracking, and the technology is better than ever. This article looks at the work being done by Boom Supersonic, Astro Mechanica, and Hermeus toward Supersonic 2.0, where anyone can catch a quick, affordable flight almost anywhere on Earth. | | Programming, Design & Data Science | OpenAI is rolling out GPT-5.2 "Codex-Max" for some users (2 minute read) OpenAI's GPT-5.2 Codex-Max is now rolling out to subscribers. There is little public information about the model available, but OpenAI may share more details in the coming days. The standard GPT-5.2 Codex can stay on track for long tasks, make large contexts useful through compaction, and handle heavy changes like refactors and migrations. It can understand screenshots, UI bugs, and diagrams. | Dynamic context discovery (6 minute read) Dynamic Context Discovery is a pattern that allows agents to pull relevant context on their own with fewer details provided up front. It optimizes token usage while still allowing agents to gather enough context for tasks. Data is stored in files, which can then be accessed as necessary. Cursor will release dynamic context discovery for all users in the coming weeks. | | OpenAI's countdown: monetization, ads, and a Google-shaped threat (9 minute read) OpenAI executives are debating whether to attend the Cannes Lions festival, a festival where the ad industry celebrates its ability to turn creativity into influence. The company has yet to find an ads boss, and it still needs to figure out how to best handle data privacy and ensure its commerce proposition proves successful. Competition is heating up, with Google now building the infrastructure needed to monetize its AI services. OpenAI isn't actively talking to advertisers about advertising yet. | | Mantic (GitHub Repo) Mantic is a structural code search engine for AI agents that provides sub-500ms file ranking across massive code bases without embeddings, vector databases, or external dependencies. | A field guide to sandboxes for AI (45 minute read) This post provides readers with a simple mental model for evaluating sandboxes and then walks through the boundaries that show up in real AI execution systems: containers, gVisor, microVMs, and runtime sandboxes. | | | Love TLDR? Tell your friends and get rewards! | | Share your referral link below with friends to get free TLDR swag! | | | | Track your referrals here. | | | |
0 Comments