GoKawiil
Tech News
Apple is overhauling the Calendar app's interface in iOS 27, focusing on the event details view and the event creation/editing screens. The updated layouts add icons, group information under clearer headings, and reorganize content so more details fit on screen at once.
Nvidia has released PAIR, a free open-source beta tool that connects compatible devices on a home network to pool their unused processing power for local AI inference and agentic tasks. It works with GeForce RTX 20-series and newer GPUs, RTX Pro cards, DGX Spark systems, and Apple M4 or later chips, running on Windows, Linux, and macOS.
Tottenham Hotspur has replaced its stadium's VMware virtualization software with HPE's Morpheus VM Essentials, cutting licensing fees by more than 85 percent. The club is nearing completion of a broader overhaul of its stadium data center, now running on HPE ProLiant Gen12 servers and Alletra storage, managed through HPE GreenLake and OpsRamp.
On the morning of Sept. 3, multiple leading AI chatbots—OpenAI's ChatGPT, Anthropic's Claude, Google's Gemini and xAI's Grok—went down around the same time, roughly 9:30 a.m. ET. Anthropic confirmed an infrastructure issue affected Claude.ai, Claude Code, Claude Cowork and its API, restoring service by about 12:16 p.m. ET, while other providers' services also returned to normal later that day.
TikTok is expanding its comment section beyond text, letting users reply with voice clips, photo carousels, and polls. The update aims to make comment threads more interactive and expressive, treating them as a richer space for engagement rather than a simple text feed.
Beyerdynamic has released the Aventho Y headphones, priced at $150 and available in four colors, boasting up to 90 hours of playback with noise cancellation off and 60 hours with it on. The headphones feature a five-minute quick charge for six hours of use, Bluetooth 6.0 multipoint connectivity, and user-replaceable batteries and ear cushions for easier repairs.
OpenAI has unveiled GPT-6 Astra, a new flagship model succeeding last year's GPT-5, bringing major gains in computer use, browsing, software engineering, and scientific reasoning. The company reports near-perfect scores on benchmarks like FrontierMath Tier 4, ARC-AGI-3, and ExploitBench, and says Astra is its best model yet for coding tasks. Astra also introduces a new memory system for Codex that lets it retain notes across long coding sessions instead of relying solely on context summarization.
OpenAI tested its new Astra model on new benchmarks measuring exploit development and binary reverse engineering, finding it far outperformed the prior GPT-5.6 Sol model. During testing, Astra independently discovered two previously unknown zero-day vulnerabilities, which OpenAI is now disclosing to the affected software maintainers. Without safety restrictions, expert testers found Astra could achieve code execution in hardened browsers and craft privilege-escalation exploits against hardened operating systems.
A published recipe reconstructs a Babylonian lamb stew with beets dating to roughly 1750-1730 BCE, based on ancient cuneiform culinary tablets. The instructions detail searing lamb in rendered sheep fat, then simmering it with beets, arugula, shallots, cumin, beer and leeks, finished with a coriander and herb garnish. A companion recipe for an ancient barley-based sourdough 'beer-bread' called bappiru is also included.
Startup Abliteration.ai now hosts modified, open-weight AI models—including Z.ai's GLM-5.3—with safety refusals removed, letting users query them via browser or API. TechCrunch tested the free web version and got the model to produce password-stealing code and instructions for culturing a dangerous pathogen. The company, founded late last year and incorporated in March, frames this as a service for red-teaming and offensive security testing.
OpenAI launched GPT-6 Astra on Thursday, calling it the company's most capable model to date. President Greg Brockman declared the release the beginning of the 'AGI era,' though the model's advanced cybersecurity capabilities and reasoning processes that are harder to monitor have sparked new safety questions.
Cerebras has added the 27-billion-parameter Qwen 3.8 model to its public API endpoints, offering inference speeds of roughly 1500 tokens per second. The model supports 64k context on free tier and 128k on paid tier, and is available under Cerebras's free trial and pay-as-you-go pricing, subject to rate limits.