TechCrunch has selected its Startup Battlefield Top 20 for the 2026 Disrupt conference, chosen from thousands of applicant founders. The finalists will pitch on the Disrupt Stage in six-minute sessions beginning October 13, competing for a $100,000 prize and the Disrupt Cup. Companies in the group span space logistics, physical AI, biotech, robotics, and fraud detection, including Longshot Space's hydrogen-powered cargo launcher and Veritas's brand-authentication microchips.
Microsoft is forming a new internal group called Xbox Labs, led by Scott (a games executive), tasked with turning emerging AI techniques—like decompiling and rebuilding games for new hardware, porting PC titles to browsers via WebAssembly and WebGPU, and remixing one game's code into another—into experiences for players and developers. The initiative follows recent community experiments that rapidly ported and transformed major titles using AI tools.
Framework has refreshed its budget-oriented, repairable Laptop 12 with Intel's new Wildcat Lake processors, keeping the 2-in-1 convertible design, touchscreen with optional stylus, and multiple color options. The base model starts at $549 without RAM or storage, but a fully configured unit costs roughly $200-$300 more than rival laptops like Dell's updated XPS 13.
Microsoft has disclosed details of its Maia 300 AI chip, a custom processor designed for data center workloads. The chip is part of Microsoft's ongoing effort to develop in-house silicon for AI computing tasks.
Framework has updated its budget-focused Laptop 12 with Intel's new Core Series 3 processors, offering configurations from a Core 3 304 chip with 8GB RAM and a 512GB SSD at $799 up to a Core 7 350 model with 32GB RAM at $1,549. Buyers can save $100 by choosing Fedora Linux instead of Windows, and pricing on all tiers has reportedly risen beyond the company's original targets.
Anthropic will disclose its third-quarter financial results for the first time at a meeting with institutional investors on Wednesday, according to people familiar with the matter. The company, which has not commented publicly, is reportedly preparing for what could become the largest IPO on record.
Canadian software engineer and YouTuber Anthony Sistilli built his own automatic license plate reader using a 4K surveillance camera and AI coding tools, configuring it to flag only police cruisers rather than ordinary vehicles. He tested the dashboard-mounted prototype in real traffic, refining it after early lag issues, and detailed the project in a YouTube video. Local police subsequently visited his home in September to question him about the device.
At its 'Welcome Home' event, Apple introduced the HomePod Mini 2, a four-inch smart speaker powered by the new S10 chip that optimizes music playback in real time. The speaker starts at $149, ships in five new colors on October 30, and will later gain expanded Siri AI capabilities through a software update. Basic Siri functions like reminders, questions, and Apple Home device control are available at launch.
Apple's website confirms that HomePod mini 2, shipping October 30, will not include the Siri AI assistant features advertised as a key selling point at launch. A footnote on Apple's site states Siri AI is 'coming in a future software update,' with the delay also confirmed in Apple's press release.
A hands-on test of Meta's Muse, a personal AI assistant, shows it can research products and complete purchases on a user's behalf. In doing so, it bypasses the review websites and retailers whose content and listings informed its recommendations, meaning those sites receive no traffic or referral credit from the sale.
A review piece assesses whether the Apple Watch Series 10, released in 2024, still makes sense to purchase in 2026. It notes the watch has already received three major watchOS updates and compares its chip to the newer S11 found in the Series 12, which reportedly doubles RAM to 4GB and improves sensor accuracy.
A developer's blog post explains a 'two-stack' method for computing sliding-window aggregations—such as minimum, maximum, or quantiles—that works even when the aggregation operator has no inverse. The technique, drawn from a paper by Tangwongsan et al. on low-latency sliding-window aggregation, achieves amortized constant-time updates using two stacks instead of a double-ended queue, and the author says it renders their own earlier six-year-old min/max algorithm obsolete.
GoKawiil