AI Weekly Malaysia

SoyaCincau documented repeated Gentari Go DC charging failures across three Malaysian locations (Petronas Bayan Baru, MINI Tian Siang Ipoh, XPark Sunway Serene) during Merdeka weekend, with 180kW chargers showing errors or 'Charging completed' at 0kWh. The app also failed to release sessions promptly after a failed attempt, forcing multi-minute waits between retries, and customer service offered only scripted troubleshooting.

OpenAI customer story for Australian law firm Gilbert + Tobin, reporting 87% active usage of ChatGPT among enabled users. The firm reduced recruitment research from 4 hours to 20 minutes and KYC/AML checks from 3-8 hours to 5 minutes using ChatGPT and Codex.

John Ternus has succeeded Tim Cook as Apple CEO, inheriting Apple's hardware business and its lagging AI capabilities. The article frames Ternus's challenge as advancing Apple Intelligence while navigating U.S.-China geopolitical tensions and domestic political pressures.

Tim Cook has stepped down as Apple CEO, with John Ternus succeeding him as only the company's third leader. Cook, who led Apple to a market value exceeding US$4.5 trillion since taking over from Steve Jobs in 2011, will move to executive chairman.

Bengaluru startup Alteon raised a $2.5M pre-seed led by Lachy Groom to develop autonomous fixed-wing aircraft that use dynamic soaring—harvesting energy from ocean wind shear—to stay airborne for over a year. The startup has not yet demonstrated sustained energy-neutral flight but completed an autonomous test over the Bay of Bengal with seven O-shaped cycles at 62 mph within one meter of the water's surface.

Apple filed new evidence in its lawsuit against OpenAI, alleging former employee Chang Liu used a confidential Apple circuit schematic and an internal Apple engineering tool at OpenAI. Apple also claims Liu enlisted OpenAI colleague Yu-Ting Peng to destroy evidence in June, and that Liu's continued access to Apple systems came from exploiting a 'rare, previously unknown authentication bug' rather than routine offboarding failures as OpenAI contends.

MyKad printing services in Malaysia are experiencing disruptions that are expected to last until 7 September 2026. The article itself contains no further detail beyond the headline, as the source text is almost entirely site navigation boilerplate.

Lowyat.NET reports that Tim Cook has marked his final day as Apple CEO, but the article body was not captured—only the site's navigation chrome is available, so substantive details about the transition, successor, or implications are missing from the source text.

Google DeepMind announced WeatherNext 3, claiming it is their most advanced and accurate global weather AI model. The provided article text contains only navigation boilerplate and the headline, with no technical details, benchmarks, architecture notes, or availability information.

Google announced Gemini 3.8 Flash and a cybersecurity-focused variant called 3.8 Flash Cyber, but the provided article text contains only site navigation and boilerplate—no technical details, benchmarks, pricing, availability dates, or feature descriptions are present.

Google announced 'agentic video understanding' for Gemini, but the provided article text contains only site navigation boilerplate and no substantive details about the feature, capabilities, pricing, or availability.

Anthropic announced Enterprise Frontier Safeguards (EFS), which combines zero data retention (ZDR) privacy with cross-session misuse detection by storing data in customer-controlled cloud infrastructure rather than Anthropic's. Until EFS rolls out in phases 'later this fall,' eligible customers get ZDR on Fable 5 and Fable 5.1. EFS will be supported on Claude Code, Claude Enterprise, Claude Platform, Amazon Bedrock, Google's Agent Platform, and Microsoft Foundry.

Hugging Face released @huggingface/kernels, a JavaScript library for loading and running 207 Apache-2.0 licensed WebGPU kernels directly from the HF Hub, each published as a versioned package with WGSL shader templates, correctness tests, and benchmark cases. They also launched Fleet, a browser-based benchmarking tool that crowdsources kernel performance and correctness data across real-world GPUs, letting users contribute evidence that helps identify failures and improve kernel variants.