AI Weekly Malaysia

Nvidia CEO Jensen Huang publicly rejected Anthropic CEO Dario Amodei's proposal that leading AI labs coordinate to 'pace the frontier' of model development, which Amodei suggested may require government mediation or antitrust waivers. Huang argued AI safety should be solved through engineering and testing, not new laws or regulations.

Nvidia CEO Jensen Huang is expected to attend President Trump's state dinner for Chinese President Xi Jinping next week, signaling a continued political alliance between Huang and Trump. AI is expected to be a main topic of the Xi-Trump meeting, and Huang has recently sided with Trump against other tech CEOs who warned about risks of rapid AI advancement.

TechCrunch reports on growing US community opposition to AI data center construction, centering on Philadelphia activist Shawmar Pitts and Philly Thrive, who are pushing for a data center moratorium in Grays Ferry—a neighborhood already scarred by a former oil refinery that exploded in 2019. City officials have identified two potential data center sites in Philadelphia, though no formal proposals exist yet, and similar resistance is emerging nationwide over environmental and energy concerns.

AWS has confirmed it cannot restore some customer data from Middle East facilities damaged by Iranian strikes, marking a rare admission of permanent cloud data loss by a major provider. The incident underscores that cloud region availability and durability guarantees have hard physical limits when infrastructure is destroyed.

Apple patched a record 260+ CVEs across iOS 27 and macOS 27 Golden Gate, with iOS 27 fixing 122 flaws and macOS 27 fixing 204. Only about 10 CVEs were directly credited to AI-assisted discovery, including two found by firm Calif using Claude and Anthropic Research. None of the patched vulnerabilities are listed as under active exploitation.

German defense contractor Rheinmetall has published documentation for onboardapi 9.10.0, the middleware protocol used in its Battlesuite connected weapon systems. The API uses OMG's Data Distribution Service (DDS) standard with a publish-subscribe architecture for low-latency sensor-to-software communication, with C++ core libraries and wrappers for Java, C#/.NET, and Python. Interface descriptions are licensed under EPL v2.0, while runtime libraries use a custom EULA (EULA-RME-SDK-1.0).

At the AI Infra Summit, Intel spin-off Cornelis Networks introduced its Active Compute Fabric (ACF), an open architecture for scale-up and scale-out AI networking, alongside its upcoming 800 Gbps CN6000-series switches and NICs targeting the Ultra Ethernet protocol. Delos Data also entered the race to replace Nvidia's NVLink, which currently lets 576 GPUs act as a single accelerator. AMD is already tunneling the competing UALink protocol over Broadcom's 102.4 Tbps Tomahawk 6 switches with MI455X GPUs, but purpose-built UALink switches remain unavailable.

A 174-line PHP polyfill written in 2014 to replace pecl_http's http_build_url() during an AOL CMS upgrade has accumulated nearly 20M Packagist installs and is bundled inside WPML (1.5M+ WordPress sites), SPIP, Debian, and Ubuntu. The author is now deprecating it, noting a known bug where trailing-slash URL joining strips every letter 'a' from the path due to a crude find-and-replace workaround.

AWS engineers open-sourced Pizza Bot, an AI agent management tool that organizes agent tasks and permission requests as email-like threads instead of live chat sessions. Finished tasks land in an 'Unread' category; items needing human decisions go to 'Action'. The tool started as an internal Amazon tool for non-coding AI agent use cases.

The article text provided contains only cookie consent and privacy policy boilerplate from Ars Technica — no actual article content about the PassThePopcorn lawsuit is present.

The US Senate blocked the Clarity Act from advancing on September 15, 2026, stalling a comprehensive crypto market structure bill after months of bipartisan negotiations. Republican leaders added ethics restrictions to address Democratic concerns about public officials profiting from crypto ventures, but the changes failed to overcome opposition tied to President Trump's family crypto profits.

Meta released a WhatsApp Business Tools MCP server that lets AI coding agents like Claude, Cursor, Codex, or ChatGPT directly set up and manage WhatsApp Business messaging—creating accounts, verifying phone numbers, registering for the Cloud API, creating/editing messaging templates, testing webhooks, and monitoring ToS/payment/verification status. Previously this required manual navigation across the Developer Console, Business Manager, API reference, and an editor.

Good Start Labs, spun out of Every with $3.6M from General Catalyst and others, is training AI models on board games to teach transferable skills. A 30B model trained on the railroad strategy game '1830' was tested on financial research tasks, and earlier work showed fine-tuning on Diplomacy improved performance on customer support and industrial operations benchmarks. CEO Alex Duffy observed that OpenAI's o3 won Diplomacy by planning betrayals while Claude Opus 4 refused to lie and 'got destroyed.'

Infoblox reports that roughly 1.7 million Chinese-language illegal casino websites double as command-and-control infrastructure for malware distribution and espionage, with some hosted on major US cloud providers (AWS, Microsoft, Cloudflare, Google) via account theft or 'infrastructure laundering.' A UNODC report estimates online scam losses in East and Southeast Asia at $88.3B–$114.1B in 2025.

Cody Ho and Niklas built a fully OpenGL ES 3.0 compliant Linux GPU driver for the M4 Mac Mini and MacBook Neo in about a month, a process that normally takes years. They reverse-engineered Apple's AGX GPU firmware ABI and user-space components in a clean-room manner using hardware traces from a custom hypervisor, achieving 200fps in Minecraft and working WebGL compositing in Chrome and Firefox. The code is not yet end-user ready, and they plan to implement Vulkan drivers next.

TypeSafe AI, founded by Diogo Almeida (who says he worked on ChatGPT-era instruction-following at OpenAI), emerged from two years in stealth to release Jev, a 'System One Model' that outputs only typed structured decisions with calibrated probabilities rather than generated text. Jev uses a new parallel sampling architecture and a training method called RLCD (Reinforcement Learning for Calibrated Decisions), and the company claims it matches LLM intelligence on structured-decision tasks while being ~100x faster and unable to hallucinate. The model is available today in early access.

OpenAI, Google, and Anthropic are in early discussions about collaborating on AI safety, following a July proposal by Google DeepMind's Demis Hassabis for a U.S.-led 'Standards Body' modeled on FINRA-style self-regulation with federal oversight. The talks are ongoing and no agreements or concrete frameworks have been announced.

AWS publicly confirmed it cannot restore service or recover data from facilities in Bahrain and parts of the UAE, more than six months after drone strikes in March damaged multiple data centers. AWS stated the damage exceeded what its services are designed to withstand, and it is still working to restore two other UAE availability zones. This is AWS's first public update since April.

TechCrunch is maintaining a running list of failed AI projects and startups. Notable entries include Relay, a 5-year-old AI workflow automation tool and Zapier alternative that shut down because OpenAI and Google built similar automation directly into their own platforms. S&P Global Market Intelligence reports ~42% of AI initiatives are ultimately abandoned by their corporate parents, and OpenAI's July 9 attempt to turn ChatGPT into a 'super app' (combining Chat, Codex, and Work modes) was rolled back after users found it confusing and cluttered.

The article title reports an 18-year-old dying of measles in Pennsylvania from a neurological complication, marking the fourth death in an outbreak nearing 700 cases. However, the captured article text contains only cookie consent and privacy policy boilerplate from Ars Technica—no substantive article content was retrieved.

Elastic Security Labs detailed a Brazilian banking malware operation (REF9334) using a toolkit called KREMLIN, active since May 2025, that installs malicious Chrome and Edge extensions impersonating a dozen Brazilian banks. The malware bypasses Chromium integrity mechanisms by manipulating Secure Preferences and regenerating HMACs, uses Ethereum smart contracts as dead drop resolvers for C2 endpoints, and sideloads its payload via a legitimate SentinelOne binary.

SpaceX has reportedly declared its Starship vehicle ready for orbital flight and set a launch date for the following week. However, the article body is inaccessible—it contains only Ars Technica's cookie consent and privacy policy boilerplate, so no technical details, dates, or quotes are available.

Agility Robotics announced a new humanoid robot designed with a safety behavior where it stops and squats to avoid injuring human coworkers. The article itself was not accessible—only cookie consent boilerplate was available in the source text.

BloombergNEF projects US data centers will consume ~18 billion cubic feet of natural gas per day by 2035, nearly double its forecast from just nine months ago, surpassing Germany and Japan combined. Onsite gas-powered data centers from Meta, Microsoft, Google, and Amazon alone could use 2.9-3.4 billion cubic feet/day by 2035—roughly equal to all data center gas consumption today. Noreva analysts warn the combined data center and LNG export boom could spike natural gas prices, potentially hitting utility ratepayers.

SpaceX will attempt the 14th Starship test flight on September 22, aiming to put the upper stage into Earth orbit for the first time and deploy 26 V3 Starlink satellites. This would be the first Starship launch to generate revenue, and only the second test since SpaceX's June IPO.