AI Weekly Malaysia

The Register's Systems Editor Tobias Mann moved from Denver suburbs to the Colorado Rockies and found DSL (20 Mbps down / 1.5 Mbps up) unusable for remote work. He now bonds T-Mobile 5G home internet (near-gigabit down, 30-40 Mbps up) with Starlink (100-400 Mbps down, ~30 Mbps up) using SD-WAN, after experiencing a dozen connection drops in two months on either service alone during working hours.

Why: If you're building or working from areas with unreliable single-link broadband — a real issue outside Klang Valley in Malaysia — bonding 5G home internet with Starlink or a second cellular link via SD-WAN is a proven pattern worth evaluating. The article's concrete drop-frequency data (12 outages in 2 months, lasting 1-5 minutes, mostly during work hours) gives you a realistic reliability baseline to plan around rather than trusting either provider's marketing.

Amazon has received authorization for a custom 35-turbine gas power plant in Texas to supply a 7.65GW AI data center, permitted to emit 33 million tons of greenhouse gases annually—potentially making it the largest single source of CO₂ pollution in the US. The scale of the plant reflects the massive power demands of AI training and inference workloads.

Why: For builders relying on AWS AI compute, this signals that hyperscaler capacity expansion is increasingly gated by energy infrastructure, not chips. Malaysian founders running AI workloads on US-region cloud should expect continued price pressure on GPU instances as power costs get internalized, and should evaluate whether Singapore or other APAC regions with different energy economics offer better long-term cost stability for inference workloads.

A blog post details the process of discovering abnormal magic hexagons (hexagonal grids where every line sums to the same value) of arbitrary order, a problem with no known formulaic construction. The author uses AI-assisted mathematics to find solutions beyond the previously known order n=9 found by Klaus Meffert in 2024, and walks through the human-plus-AI discovery process including observations, search space reduction, and construction.

Why: For builders experimenting with AI-assisted problem-solving, this is a rare first-hand walkthrough of how to combine human mathematical intuition with AI search to crack a combinatorial problem that had no deterministic algorithm. The concrete takeaway is the methodology: making structural observations to shrink the search space before applying AI, rather than brute-forcing it.

Puspakom deployed Malaysia's first AI-assisted vehicle undercarriage inspection system, developed with local tech company Keymag Sdn Bhd, cutting commercial vehicle inspection time from ~10 minutes to ~1 minute (90% reduction). The system uses image analysis to detect corrosion, oil leaks, brake faults, axle defects, and structural damage, but human Vehicle Examiners retain sole authority for pass/fail decisions. It is currently live on Puspakom's Mobile Truck Service for heavy commercial vehicles.

Why: This is a concrete Malaysian example of AI deployed as decision-support rather than autonomous decision-maker in a regulated, statutory process—a pattern directly relevant to anyone building AI systems for government or GLC clients where legal liability cannot be automated away. The human-in-the-loop architecture here is the takeaway: if Puspakom won't let AI make pass/fail calls on vehicle safety, your SaaS probably shouldn't either for high-stakes decisions.

Amazon is investing in an on-site natural gas power plant for a planned data center in Pecos County, Texas, permitted to release 33 million tons of CO2 per year — more than any other U.S. power plant. Amazon's carbon emissions rose 16% last year, moving away from its 2040 net-zero pledge, with a spokesperson acknowledging 'the world looks different now.'

Why: If you build on AWS and your company or clients have ESG or carbon-neutrality commitments, Amazon's rising emissions and reliance on fossil-fuel-powered data centers directly affect your Scope 3 reporting. Founders pitching to investors with sustainability mandates should be prepared to explain their cloud infrastructure's carbon footprint, especially as AI workloads push hyperscalers toward dirtier energy sources.

Nvidia's RTX 5090 is reportedly being sold by retailers in bundles that include 8 motherboards, forcing buyers to purchase unwanted components to get the GPU — a tactic reminiscent of the crypto mining boom's GPU scarcity era. The practice effectively inflates the real cost of acquiring an RTX 5090 well beyond its sticker price.

Why: If you're planning to buy an RTX 5090 for local AI/ML workloads or agent development, budget for bundle-inflated pricing rather than MSRP. Malaysian builders importing these cards may face significantly higher effective costs, which affects ROI calculations for on-prem GPU inference setups versus cloud alternatives.

Senko Rašić pushes back against the popular narrative that 'coding was never the hard part' of software development, arguing it dismisses the skill, stress, and craft programmers have invested for years. He challenges both sides of the claim with rhetorical questions: if coding is easy, why do buggy software, burnout, and deep technical literature exist; and if figuring out what to build is the hard part, why aren't product managers and business analysts the highest-paid rockstars in tech companies.

Why: If you're using AI coding tools or building AI-assisted products, this piece is a useful counterweight to the prevailing discourse that coding skill is becoming irrelevant. It's worth reading before you repeat 'code was never the hard part' in a team discussion or hiring context—the argument has logical gaps that the author exposes, and dismissing implementation difficulty may lead you to undervalue experienced engineers precisely when AI-generated code needs the most careful review.

UTM has released Triton, a Windows DirectX 11 DDI-level driver for QEMU that, combined with their Neptune protocol forwarding layer, enables modern Windows games and GPU-accelerated applications to run in QEMU VMs. Unlike previous DLL-replacement approaches, Triton implements the DirectX Device Driver Interface directly, avoiding CPU blitting performance penalties and anti-cheat detection issues.

Why: If you run Windows VMs on macOS or Linux for testing or compatibility, Triton means you can now get native-grade DirectX 11 acceleration without per-application DLL hacks. Developers doing cross-platform testing on Apple Silicon should evaluate UTM/QEMU as a more viable alternative to paid solutions for Windows-on-ARM workloads.

Microsoft Edge is ending support for Manifest V2 (MV2) extensions, disabling older ad blockers like uBlock Origin starting this month. Only 58 extensions with meaningful usage still rely on MV2, and just 3 lack MV3 equivalents. Consumer transition completes by end of 2026, with enterprise following in early 2027.

Why: If you or your team relies on uBlock Origin in Edge, it will be gradually turned off—switch to uBlock Origin Lite or another MV3-compatible blocker now, or move to Firefox or Opera if you need full MV2 capabilities.

A new Amazon data center in Texas is raising concerns because it would be powered by what could become the most polluting power plant in the United States. The NYT report highlights the tension between hyperscaler AI/compute expansion and environmental costs of the energy infrastructure backing it.

Why: Malaysia is experiencing its own data center boom with AWS, Google, and Microsoft investing billions locally. If environmental scrutiny on hyperscaler power sourcing intensifies globally, Malaysian builders should expect tighter sustainability reporting requirements and potential cost pass-throughs on cloud compute as providers internalize cleaner energy costs. Founders running AI workloads on AWS should factor in that ESG pressure on data center power may eventually affect pricing or available capacity in their region.

The Nixpkgs core team has disbanded, announced via a GitHub PR. Over their 10-month tenure they reformed the committer delegation process, onboarded 19 new committers, and extended the merge bot to empower maintainers.

Why: If you rely on Nixpkgs/NixOS for reproducible builds or dev environments, governance is now in flux with no core team steering package maintenance and committer delegation. Monitor how merge-bot permissions and committer onboarding continue without the team before committing further infrastructure to Nix.

ShinyHunters used a vishing (voice phishing) attack to trick staff at Abbott's cancer diagnostics unit Exact Sciences into granting access, then dumped 10.9 million email addresses plus personal health data after Abbott refused to pay ransom. The crew claims to have exfiltrated 30M+ rows of customer info, 22M+ rows of doctor-patient notes, 20M+ medical-order records, and 425M+ rows from Databricks. Abbott disclosed the breach on July 16 and confirmed it began with a phone-based social engineering attack, not malware.

Why: The Databricks exfiltration of 425M+ rows is the detail builders should note: if your analytics warehouse is accessible via credentials a phone call can extract, your data lake is one social-engineering attempt away from a breach. Review whether production Databricks access requires MFA and whether human-operated credential resets have an approval chain that a convincing caller can't bypass.

MIT CSAIL researchers Daniël Trujillo and Mengjia Yan will present TONTOU at DEF CON 34, a new speculative execution attack that bypasses Spectre v2 defenses on Intel and AMD CPUs by using precisely timed timer interrupts to re-poison branch predictor state during the post-neutralization window. They demonstrated a working exploit on AMD Zen 2 using a stock Linux kernel with all default mitigations enabled, showing that both Intel's eIBRS (entry neutralization) and AMD's Safe RET (in-place neutralization) are vulnerable.

Why: If you operate multi-tenant infrastructure on Intel or AMD CPUs—especially Zen 2—unprivileged code with timer access can re-poison branch predictor state after mitigations run, meaning existing Spectre v2 defenses are insufficient until kernel or microcode patches land. Cloud providers will likely patch transparently, but anyone self-hosting shared workloads should track vendor advisories and plan for kernel updates rather than assuming current eIBRS or Safe RET protections are adequate.

Cloudflare's Web Integrity & Trust team outlines their approach to classifying traffic on the 'Agentic Internet,' where sessions can shift between human and automated agent mid-session (e.g., a user browses a store, then hands checkout to a shopping assistant). They distinguish Risk (ephemeral likelihood of harm) from Trust (reputation built over time) and argue that static point-in-time checks are insufficient—continuous behavioral analysis is needed. The post previews findings from their Precursor product and teases upcoming launch updates.

Why: If you build AI agents that browse, scrape, or transact on third-party websites, Cloudflare's shift from 'bot = block' to behavioral trust scoring means your agents may be evaluated on accumulated reputation rather than per-request checks. This affects whether your agent traffic gets through Cloudflare-protected sites—start thinking about session-level behavior patterns and handoff design now, before the upcoming product updates land.

Cloudflare is merging Workers AI (its GPU-hosted inference service) and AI Gateway (its proxy for any model provider with observability, logging, and security) into a single control plane. Developers now use one unified Workers binding (`env.AI.run()`) and one REST API endpoint (`/ai/`) to route to any provider, including Cloudflare's own hosted models like `@cf/zai-org/glm-5.2`, with a built-in 'default' gateway for automatic logging.

Why: If you build AI agents or inference pipelines on Cloudflare Workers, you no longer need to choose between the Workers AI binding and AI Gateway setup—both paths are now one, and you get observability and logging by default without configuring a separate gateway. This simplifies multi-provider routing and billing consolidation, but it is a platform-specific convenience, not a reason to migrate if you are not already on Cloudflare.

Cloudflare beta-launched Radar Researcher, an AI tool that lets users query Cloudflare Radar's internet traffic datasets in plain language and receive interactive charts in response. The tool is built on Cloudflare's developer platform and grounds its LLM responses in Radar's free API data, covering DNS queries from 1.1.1.1, HTTP traffic, network quality data, and more. It is available on every Radar page and was announced as part of Cloudflare's Agents Week.

Why: Radar Researcher is a working example of the pattern where an LLM agent is grounded in a real API rather than free-form generation—useful both as a tool you can try today for internet traffic analysis and as a reference architecture if you are building AI agents over your own datasets. If you currently use Radar's API manually for network or outage research, you can now skip filter selection and API documentation and ask questions directly, though it is in beta so verify outputs against raw API calls for anything critical.

An 18-year-old use-after-free bug in Linux's SCTP networking code (CVE-2026-64564, dubbed 'SCTPhantom') allows local users to gain root and escape containers. Tencent Zhuque Lab demonstrated container escape on Debian 13, Ubuntu 24.04, Rocky Linux 9, RHEL 9, and OpenCloudOS using default seccomp without CAP_NET_ADMIN. Fixes shipped August 3 in kernels 7.1.6, 6.18.42, 6.12.101, and 6.6.148; no public exploit code exists yet and it's not in CISA's KEV catalog.

Why: If you run multi-tenant container platforms or shared Linux hosts with SCTP enabled, patch your kernels now—the bug has been present since Linux 2.6.25 (2008) and the container escape path works without special capabilities. If SCTP isn't reachable on your hosts, exposure is limited since the flaw is local-only, not remote.

Security researcher Malcolm Stagg disclosed NatJack, a new attack class presented at Black Hat USA 2026 that manipulates NAT connection-tracking state to hijack active TCP sessions, spoof DNS responses, expose mapped ports, and exhaust NAT tables. Two CVEs were assigned: CVE-2026-56181 (CVSS 8.3) in Windows NAT used by Hyper-V and CVE-2026-63913 (CVSS 8.2) in Linux Netfilter conntrack, with patches available for both.

Why: If you run multi-tenant workloads behind shared NAT (e.g., Hyper-V VMs or Linux containers on the same host), apply the Windows and Linux updates now and stop assuming hosts behind the same NAT cannot interfere with each other's connections. The attack requires privileged access on a system behind the same NAT, so isolate untrusted workloads from trusted systems sharing NAT infrastructure and encrypt internal traffic rather than relying on NAT boundaries for trust.

Arctic Wolf Labs reports a widespread AitM phishing campaign hijacking Microsoft 365 accounts via voicemail-themed emails, using a six-stage redirect chain through Google Meet, Google Ads, and Amazon S3 to bypass reputation filters. Compromised sessions are maintained at roughly eight-hour intervals using residential proxies, and attackers target employees in financial workflows to reroute payroll payments. Hundreds of organizations were hit last month across healthcare, education, manufacturing, government, and professional services in the U.S., Canada, and Europe.

Why: If your organization runs on Microsoft 365, MFA alone does not stop this attack—AitM proxies capture credentials and MFA codes in real time. Founders and IT leads should verify whether conditional access policies (device trust, impossible-travel detection, session length limits) are enforced, since the attackers maintain sessions for eight hours and use residential proxies to mimic legitimate sign-ins. The abuse of Google Meet redirect URLs and S3-hosted HTML as redirect hops means email filters that trust Google/S3 domains will not catch the initial lure.

Anthropic is hiring engineers to co-design custom ASIC inference chips, with Samsung reported as the manufacturing partner, joining Amazon, Meta, OpenAI, and Google in building in-house AI silicon to reduce dependence on costly Nvidia GPUs. The move is framed around making inference workloads more efficient and potentially profitable as chip supply remains constrained.

Why: For builders shipping on Claude APIs, this signals that inference cost-per-token could drop materially over the next 2-3 years as custom silicon replaces Nvidia GPUs in Anthropic's data centers. If you're architecting AI agents or SaaS products with tight margins on API costs, factor in that pricing leverage may shift — but don't change anything today since this is early-stage hiring, not a shipping product.

SK Hynix will invest 54 trillion Korean won ($38.1 billion) to build two new memory chip plants: a 35.2 trillion won fab in Yongin called 'Y2' and a 19.1 trillion won facility in Cheongju called 'M17.' The expansion is driven by surging memory prices caused by supply shortages and high demand from AI infrastructure builders and chip firms like Nvidia for high-bandwidth memory (HBM).

Why: Memory prices are climbing due to HBM supply tightness, which directly affects GPU availability and cloud compute costs for AI/ML workloads. Builders scaling AI inference or training should expect continued upward pressure on hardware and cloud pricing through the multi-year buildout timeline, and factor this into infrastructure budgeting and vendor lock-in decisions.

U Mobile launched its Enterprise Innovation Platform (EIP) Hub, a sandbox facility built on its 5G-Advanced infrastructure where Malaysian enterprises can develop, test, and demonstrate 5G-A and AI solutions before scaling. The hub has 15 partners including AWS, Huawei Malaysia, Palo Alto Networks, Qualcomm, and Agibot, and is working with MRANTI and Cradle Fund to connect startups with funding pathways and testing environments.

Why: If you are building AI, robotics, IoT, or edge-compute solutions in Malaysia, the EIP Hub offers a controlled 5G-A testbed and a potential route to Cradle Fund and MRANTI support—worth exploring before investing in your own infrastructure. The partner list (AWS, Qualcomm, Huawei, Agibot) signals what tech stack the sandbox is optimized for.

A weekly security roundup covering 30+ threats, including an npm supply-chain campaign ('Flooding Dropper') flooding the registry with 850 malicious packages, SideWinder's new phishing chain abusing ClickOnce files to deliver Rust-based backdoors via Cloudflare Workers, a US Congressional report on Chinese telecom presence in US infrastructure, and various RCE and one-click takeover vulnerabilities in consumer-facing products.

Why: The 850-package npm flooding campaign is the most directly actionable item: if your projects pull from npm without lockfile pinning or provenance checks, you are exposed right now. The SideWinder chain's use of Cloudflare Workers for C2 is a reminder that serverless platforms you build on can also be abused for malware hosting — relevant if you're shipping agent or automation tooling that trusts external endpoints.

AmInvestment upgraded ViTrox Corporation to 'Buy' with a RM10.25 target after 2QFY2026 revenue hit RM375M (+104.8% YoY) and net profit tripled to RM85M, driven by AI infrastructure spending flowing through the semiconductor supply chain. The research house raised FY2026-FY2028 earnings forecasts by 96%-121%, noting ViTrox's largest customer is now a hyperscaler contributing ~10% of group revenue, and that demand appears structural rather than temporary front-loading. Capacity expansion continues at Batu Kawan, Penang (Campus 4 and 5), with a renewed Pioneer tax incentive pushing effective tax rate below 10%.

Why: For Malaysian builders and founders, this signals that AI infrastructure capex is creating durable demand in the local semiconductor equipment supply chain, not just a cyclical blip. If you're building or investing in Penang-region hardware, manufacturing tech, or supply-chain SaaS, ViTrox's hyperscaler relationship and capacity expansion at Batu Kawan are concrete indicators of where AI-driven spending is landing locally.

WIRED reports that Meta approved and ran over 50 ads containing AI-generated child sexual abuse material across Facebook, Instagram, Messenger, and Threads between November 2025 and early August 2026, with some reaching several thousand accounts in the US, UK, and Europe. Researchers at the Tech Transparency Project found the ads in Meta's own ad library, many linking to 'nudify' apps, with some specifically targeting men and running for several days before removal.

Why: If you build any platform with user-generated content, ads, or AI-generated media pipelines, this is a concrete demonstration that automated content review at Meta's scale still fails catastrophically on the most obviously prohibited content. The ads were not buried user posts—they passed Meta's ad review and collected ad revenue. Anyone shipping AI image generation features or ad marketplaces should treat trust-and-safety moderation as a hard dependency, not an afterthought, and should assume their automated filters will miss abuse that seems glaringly obvious to a human reviewer.