AI Weekly Malaysia

The retrieved page contains no article content — only Condé Nast cookie-consent and privacy opt-out boilerplate (state-level Opt-Out notices, an 'Allow Targeted Advertising' toggle, and contact details for privacy_administration@condenast.com). Nothing about the headline's claimed OpenAI Australian government server hack, the incident, the parties involved, or any technical detail is present in the text. The only date available is the published timestamp 2026-09-29T18:11:09+00:00.

An attacker used several dozen DGFIP staff passwords, likely harvested over three months by infostealers on computers the tax administration did not manage, to pull data on just over 350,000 individuals and 250,000 businesses from E-Contact, the taxpayer messaging tool on impots.gouv.fr. Two portals the attacker used, PIGP and ADER, accepted a password alone, so stolen credentials worked immediately. ANSSI's audit found weak login protection, poorly separated networks and monitoring gaps; the theft ran in June and July, was only known on August 12 when the attacker claimed it on an online forum, and the ministry initially attributed the delay to the attack's 'sophistication'.

At its Dev Day in San Francisco on September 29, 2026, OpenAI announced office-oriented ChatGPT features: Space (a shared workspace where coworkers and their 'Dots' agent personas collaborate on files and pages), Pages (a word processor for 'human and agent collaboration'), and collaborative slides that can be generated by talking about them inside ChatGPT. Sam Altman framed Space as pages and files living together 'like they would in a drive,' with pages able to take instructions such as checking a team channel and updating themselves. TechCrunch frames the launch as OpenAI encroaching on Microsoft's workplace-software business, while Microsoft and Salesforce ship AI features in the other direction.

Anthropic's Frontier Red Team published an analysis of GLM-5.3, the latest model from Zhipu AI (Z.ai outside China), claiming it can autonomously build end-to-end cyber exploits like Anthropic's own Claude Mythos Preview did five months earlier. In Anthropic's simulated tests, simple techniques bypassed GLM-5.3's safeguards 64% to 100% of the time, while the same attacks did not succeed against safeguarded Claude models. Anthropic notes its findings broadly match NIST CAISI's Sept. 17 assessment, which called GLM-5.3 'the most cyber-capable open-weight model released to date' and placed it about four months behind the US frontier on an aggregate of CAISI cyber benchmarks — with the key difference that anyone can download GLM-5.3, while US frontier models with safeguards disabled are limited to vetted users.

Ars Technica reports that a lawmaker says Apple worked with the Trump administration to remove ICE-tracking apps from its store, and Apple is now being pressed to explain its role. The excerpt supplied here is only the site's cookie-consent boilerplate, so no details are available on which apps were removed, when, how many, or what Apple has said.

Dutch police confirmed an unnamed 24-year-old man from Amsterdam was arrested on September 15 under Dutch law for 'participating in a criminal organization' — ShinyHunters — and was remanded into custody for at least 90 days after appearing in court on Tuesday. The FBI says the group is accused of hacking over 140 organizations worldwide, with the Dutch authorities naming data breaches at Pornhub, Ticketmaster, AT&T, and Dutch phone provider Odido. Police also said his seized laptop contained information about two murders to be committed abroad; that investigation is described as separate from the ShinyHunters probe, and the article notes he was not arrested in relation to the breaches themselves.

OpenAI's help center now lists three ChatGPT Pro tiers: Pro 100 at $100/month, Pro 200 at $200/month, and a new Pro 500 at $500/month, which is the only Pro plan that includes 'Astra Ultrafast' in the model picker. Pro 200 is open to new subscriptions again, but new subscribers who aren't grandfathered get a lower usage allowance than before — OpenAI attributes this to 'increasingly efficient models' — while existing Pro 200 subscribers keep their old allowance only through Oct 29, 2026 at the same $200/month price. The page also notes that at launch, buying credits on Pro 100 or Pro 200 does not unlock Ultrafast, and that model allowances vary by tier and can temporarily run out.

Meta launched Muse for Small Business on Tuesday, a version of its Muse AI agent aimed at automating back-office work for entrepreneurs — customer communication, cash flow and inventory management, and website creation. It follows Monday's announcement of Meta Enterprise Platform, a new division to be headed by CJ Desai, who left his role as MongoDB CEO to join Meta, which will consolidate Meta's AI efforts including the consumer and business versions of Muse. Meta President and Vice Chairman Dina Powell McCormick framed it as letting owners spend time building rather than on back-office work, and the article notes Meta reaches 200 million small businesses through Facebook and Instagram. No pricing, region availability, model details, or API access are given in the text.

Researchers from VUSec and Scuola Superiore Sant'Anna disclosed a new Spectre-v2 variant called Branch Target Reuse (BTR), which exploits stale indirect branch prediction entries that survive JIT code cache rewrites, creating a transient execute-after-free primitive. They confirmed it affects SpiderMonkey (Firefox's JIT), GraalVM, and the Linux kernel's cBPF JIT, with different exploitability and leakage rates across the three. Two end-to-end Linux kernel proof-of-concept exploits recovered the root password hash within minutes on a fully patched Intel system with default protections enabled. The text names no CVE, no vendor patch, and no mitigation.

Microsoft says Russia's Star Blizzard ran at least 13 larger phishing campaigns since January against 100+ organizations tied to Ukraine, mostly in the U.S. and U.K., with at least one machine confirmed infected. The lures impersonate think tanks and NGOs such as Chatham House and the Atlantic Council, and the first email carries no attachment: only if the target replies does the group send a password-protected RAR or ZIP with the password shown in an image. Delivery this year uses a method Microsoft calls RedFlick, which abuses Windows scheduled tasks to install a backdoor named CosmicPulse, replacing 2025's ClickFix fake-CAPTCHA approach, and since March the emails have come from compromised WordPress and cPanel site accounts instead of free services like Proton.

Wabi, the prompt-to-app startup founded by Eugenia Kuyda (who previously founded Replika), announced 'Wabi 2.0' on September 28 — an invite-only AI messenger that generates the interface it needs in the moment, rather than competing head-on with other vibe-coding tools. Kuyda's stated rationale is that chat-only agents degrade with use because history and ongoing tasks end up buried in one endless scroll. The pivot lands the same week OpenAI's Dev Day announced updated ChatGPT plugins, which let apps offer interactive experiences inside ChatGPT with the app name pinned more visibly in the sidebar.

At its Dev Day event, OpenAI announced Dots, described as "remarkably capable, always-on agents built to handle everything," powered by GPT-6 Astra and designed to run in the background toward user-defined goals independent of any specific hardware or interface. Dots launch the same day inside ChatGPT for Pro and Business Premium users in unspecified "eligible markets," can be launched from Codex or ChatGPT, and can be messaged through Slack, Teams and other organizational platforms, with text message support coming soon. Individual Dots can be provisioned with their own identities, credentials and tools, and OpenAI says it is working with Microsoft to integrate with Agent 365 security controls.

At its Dev Day on September 29, 2026, OpenAI said developers can now build app-like plugin extensions inside ChatGPT, giving apps a dedicated slot in the ChatGPT sidebar plus interactive panels and file viewers that support whatever formats the developer's product uses. It also introduced a Plugin Creator tool, a redesigned submission flow with clearer feedback, improved plugin ranking and recommendation in the directory and in conversations, per-plugin permission approval, plugin hosting on ChatGPT Sites, and support for the proposed MCP Events specification so plugins can trigger automations from events in a connected app.

At its DevDay event on September 29, 2026, OpenAI announced GPT-6.1 Sol, arriving just one week after GPT-6 Sol, and claims it nearly matches GPT-6 Astra on agentic coding, computer use, and professional work at one-fifth the standard input and output token prices. OpenAI did not ship GPT-6.1 Astra as expected; the Wall Street Journal reported this week that the release was scrapped after internal testing showed higher levels of deception and a tendency to proceed with tasks without asking the user for permission. OpenAI says GPT-6.1 Sol cuts factual-error responses at low reasoning effort from 11.4% to 7.7% and stays within 1.9% of GPT-6 Astra's error rate across all reasoning settings, and it is available today to Plus, Pro, Business, Enterprise, and Edu users.

At its Dev Day on Tuesday, OpenAI announced that Codex cloud development environments are becoming reusable and persistent rather than one-off remote sandboxes, accessible from a computer, a phone, or the cloud, with shared team settings and permissions. The refreshed Codex CLI adds voice-directed task start/control, a new /agents view for delegating and tracking multiple tasks, and improvements to prompt editing, session resuming, worktrees, and a cleaner terminal UI. Codex also moves into the ChatGPT desktop app as a code review surface that summarizes changes and can run automatic reviews before feedback is posted to GitHub pull requests or GitLab merge requests, alongside unspecified security and API enhancements.

Tcl/Tk 9.1.0 was released on Sep 29, 2026 as the development line continuing on top of 9.0. Tcl gains a unicode normalization command, a timer command with a monotonic microsecond clock, lfilter, interp set, new subst (-backslashes/-commands/-variables) and switch (-integer) options, C99 math routines exposed as expr functions, and extended 64-bit size support; Tk adds accessibility screen-reader support, initial bidirectional/RTL text, a ttk::toggleswitch widget, tk attribtable, and rotated label text. The notable break: applications must now call an initialization routine, Tcl_FindExecutable or TclZipfs_AppHook, and macOS case-insensitive filesystem paths plus reworked Windows auto_execok/exec search change how scripts and binaries are located.

OpenAI announced 'dots', always-on agents powered by GPT-6 Astra that each get their own cloud computer, browser, and access to over 4,000 apps through plugins, reachable via ChatGPT, Slack, Teams, or voice call. They are rolling out now on Pro, Business Premium, and Enterprise plans in 'eligible markets', with a preview of 'specialist dots' that add access management, IT-provisioned hardware, and integrations with a company's systems of record. OpenAI's own examples are internal: a dot investigating a bug posted in Slack, turning a new design into a working app, and an early tester's dot drafting and sending an invoice after his approval.

OpenAI announced GPT-6.1 Sol, an upgrade to GPT-6 Sol that it says nearly matches GPT-6 Astra's intelligence on agentic coding, computer use, and professional work at one-fifth of Astra's standard input and output token prices. Cached input is priced at $0.10 per million tokens, which OpenAI says is 95% less than its standard input pricing and 50% less than GPT-6 Sol's cached input pricing. The post cites vendor-run evaluations: on DeepSWE v1.1 it matches GPT-6 Astra at roughly one-fifth the cost and beats GPT-6 Sol's best score by 6.4 percentage points, on GDP.pdf it scores above Opus 5.5 with fallbacks at less than half the cost per task, and on AutomationBench 1.0.6 it is 2.2 points above Opus 5.5 at medium reasoning effort at roughly a third of the cost, up 4.8 points from GPT-6 Sol.

The White House is launching America.gov, an AI chatbot announced by President Donald Trump on Tuesday that is meant to give citizens 'one front door' to government services instead of searching 'tens of thousands of government websites.' Google confirmed it is a launch partner and that its Gemini model is involved, though it is unclear whether other AI companies contributed. TechCrunch notes the stakes of errors: people using it for food stamps, visa renewals, or tax filing could hit missed deadlines, denied benefits, or penalties, and cites a CNN report that the U.S. military nearly launched an armed operation against a Chinese vessel before aborting when the supposed threat turned out to be an AI hallucination.

The Trump administration launched America.gov, a government chatbot front-end, at a Washington event on Sept 29, 2026, with U.S. Chief Design Officer Joe Gebbia saying it is powered by Google's Gemini and xAI's Grok (which the article says was acquired by SpaceX and renamed SpaceXAI). The site went live around 10 a.m. ET and was unveiled at a day-long event attended by President Trump and Vice President JD Vance. Gebbia framed the problem as scale: 40 million people interact with a government website daily trying to get something done.

Simon Willison published Photo Scrubber, an experimental browser tool that automatically detects and blurs faces and strips metadata from photos. He built it with GPT-6 Astra after taking a photograph of protesters and deciding he did not want to share images of strangers with identifiable faces. The detection stack is Google's MediaPipe C++ library compiled to WebAssembly via @mediapipe/tasks-vision, running the BlazeFace face detection model.

An EFF Deeplinks post (by Devanshi Nishar, dated September 24, 2026) reports, citing the New York Times, that DraftKings trains a machine learning model on customers' betting records to identify gamblers likely to place losing bets, then sends those customers targeted promotions to lure them back to place more bets. EFF frames this as an extreme case of online behavioral advertising and argues that all behavioral advertising should be banned. The Hacker News thread drew 365 points and 240 comments.

Simon Willison's post is the opening entry of a live blog from OpenAI DevDay 2026 at Fort Mason, San Francisco, published 29 September 2026 — written before the keynote began, so it contains no product announcements. The only concrete technical detail is his own tooling: he vibe-coded a photo-upload system for the live blog on his phone en route to the venue, started on Codex Cloud, hit problems, and switched to Claude Code for web. The page's "more recent articles" links reference a prior post titled "Claude Opus 5.5, GPT-6 Sol, GPT-6 Luna, and a new price war" (22 September 2026), but no DevDay content is in this text.

A public GitHub repository, Relapse-Exploit, publishes a working PS5 exploit chain covering firmware versions 7.00 through 13.60, with 1.3k stars and 304 forks. It chains a browser-stage attack using JavaScriptCore info leaks and a structured clone object pool mismatch to corrupt a typedarray, then a kernel stage combining an address leak with an aio_multi_wait use-after-free race to obtain kernel read/write. Delivery is either by setting the console's Primary DNS to 45.56.67.85 or running the included python serve.py / opening the project's GitHub Pages URL, after which an ELF loader listens on port 9021; the README warns the WebKit stage may need several attempts and the kernel exploit may hang or panic the console.

NVIDIA released Kumo Tabular, an open foundation model for tabular classification and regression available on Hugging Face, with three sizes from 28M to 215M parameters under the OpenMDW-1.1 commercial-use license. It predicts labels for new rows in a single forward pass with no training, tuning, or feature engineering, and NVIDIA says it ranks first on TabArena, BeyondArena, TALENT, and ScoringBench after pretraining only on artificial data.