AI Weekly Malaysia

DeepSeek, the Chinese AI lab behind the model that gave China its 'ChatGPT moment,' is seeking fresh outside capital because its long-time funder High-Flyer Quant (the hedge fund co-founded by Liang Wenfeng) can no longer bankroll its growing needs. High-Flyer has been taking pre-IPO placements in China's hard-tech IPOs (chips, robotics) but has faced sharp drawdowns from AI-chip stock volatility.

https://archive.ph/ni6I3https://storage.courtlistener.com/recap/gov.uscourts.cand.46... Comments URL: https://news.ycombinator.com/item?id=49473522

TechCrunch announced a StrictlyVC evening event in New York's West Village on September 10, featuring Keith Rabois discussing Khosla Ventures' early $50M OpenAI bet and his views on founders raising excess capital, plus Craig Shapiro and Jason Levien on sports-as-investment, and Tristan Walker on community-building in an AI-driven world.

OpenAI and Thailand's Ministry of Higher Education, Science, Research and Innovation (MHESI) announced an eight-week accelerator in Bangkok to help Thai AI startups move prototypes into production-ready products. The program targets locally relevant use cases, with examples including hospital phone-line automation and classroom applications.

Anthropic has proposed an as-yet-unpublished protocol called the Model Hardware Standard (MHS), conceptually similar to MCP but designed to let AI agents control physical devices like lab equipment, factory machines, and robots. The protocol uses simple primitives like 'read' and 'write,' makes devices discoverable via tagged reference files, and is currently being tested at HHMI's Janelia Research Campus, with other organizations able to apply for a research preview.

TechCrunch Disrupt 2026 (October 13-15, San Francisco) announced two AI Stage sessions: Anthropic's Head of Applied AI Cat de Jong will discuss enterprise Claude deployment patterns—what works, what stalls, and what separates successful deployments from perpetual pilots—and OpenAI's Head of Productivity will cover 'building AI native,' including the rise of GTM engineering as a fast-growing role. The AI Stage is presented by Google for Startups and will also cover AI pricing when models commoditize and agent security rebuilt from infrastructure up.

Nvidia and Cerebras are trading benchmark blows at Hot Chips, with Nvidia claiming 3,400 tokens/sec on Gemma 4 31B using Groq-3-based LPX racks and Cerebras countering with its upcoming CS-4 accelerators. Both figures are measured at batch size 1 — a single concurrent request — which no production inference-as-a-service operator would run because it's economically unviable. The numbers are real but analogous to a car's top speed: technically achievable, practically irrelevant for paying workloads.

Breaking Claude Code Opus 5 Auto Mode Anthropic are putting a great deal of faith in Claude Code's auto mode for protecting their coding agent users against prompt injection attacks. They recently made that the default and have made bold claims about its effectiveness. Johann Rehberger is one of the most credible prompt injection researchers active today. He found an attack against auto mode which he claims works 80% of the time, by tricking Claude Code into downloading and uncompressing a zip archive, then executing code that imports base64 without noticing that this will import and execute a local struct.py file extracted from the archive. In a few cases auto mode directly prevented the agent from preventing harmful code from continuing to execute! In a few runs Claude tried to terminate the malware process once it noticed the compromise, but Auto Mode denied the cleanup command. Claude detects the compromise, but Auto Mode blocks its cleanup command The safety mechanism itself can become part of the failure. The classifier allowed the creation of the malware process, but then it blocked the command intended to stop it! I agree with Johann's conclusion here: the only safe way to run agents if there's any risk of attracting the attention of an adversarial attack is with a sandbox: Run unattended coding agents in a container, VM or OS sandbox. Restrict network egress. Monitor your agents. Do not expose home directories, SSH keys, cloud credentials,… to the agent runtime. [...] Update 30th August 2026: On Lobste.rs hyperpape points out that this doesn't fit the bill of a classic prompt injection attack because at no point are malicious instructions from the website accidentally followed by the LLM. They're right: this is more of a confused environment attack where the nature of the environment that the agent is exposed to results in an exploit. Tags: sandboxing, security, ai, prompt-injection, generative-ai, llms, anthropic, claude, johann-rehberger, claude-code

CrowdStrike CEO George Kurtz used a Mad Money appearance after a record earnings pop (shares up 20% to ~$228) to argue that AI is accelerating attackers' ability to find and exploit vulnerabilities, making legacy and free security tools inadequate. He framed this as driving demand for CrowdStrike's platform.

Article URL: https://12factor.net/ Comments URL: https://news.ycombinator.com/item?id=49472216

https://web.archive.org/web/20260826205502/https://www.newsc...https://archive.ph/zahjchttps://x.com/mikesmariani/status/2092966720252133523 Comments URL: https://news.ycombinator.com/item?id=49472090

CNBC's Jim Cramer said Nvidia and Salesforce quarterly earnings beat expectations, with Salesforce surging 22% and Nvidia 8%. He argued the results refuted two bear narratives: that AI will make traditional SaaS software obsolete (Salesforce) and that AI chip demand is slowing (Nvidia).

The article title announces a new hardware standard from Anthropic intended to let AI agents control physical-world devices, but the provided article text contains only cookie consent and privacy policy boilerplate—no substantive content about the standard itself.

Article URL: https://github.com/elyosh/OpenTIE/ Comments URL: https://news.ycombinator.com/item?id=49471965

CRPx0, a cybercrime crew that evolved from a scam service into a ClickFix-delivered ransomware and crypto-theft operation over summer 2026, claims its victim count jumped from fewer than 10 in June to 48 organizations on its clear-web leak site. The group offers a white-label ransomware-as-a-service platform—originally $10,000 one-time, now a $333 enrollment fee plus a 70-30 revenue split—where CRPx0 builds all infrastructure and affiliates bring their own brand. Threat-intel analyst Rakesh Krishnan published early details and malware samples on TheRavenFile blog.

UAE national security advisor Sheikh Tahnoon bin Zayed al Nahyan and co-investors own 49% of WLTC Holdings, the holding company for the Trump family's planned crypto bank under World Liberty Financial, via StringZ Holding RSC. The venture received preliminary conditional approval for a federal bank charter, and the same investors previously put $500 million into World Liberty, with $263 million directed to Trump family entities.

Hi HN, we built an open source model gateway. It's a single place to manage our own self hosted, frontier, and open source models in one place.It’s is rust native, built for concurrency, and implements all the config quirks across models and providers (streaming formats, tool calls, model parameters, rate limits, and different error behavior).The gateway adds under 1 ms for BYOK requests and under 2 ms when Experiential supplies the provider key. It has every major inference provider, and 1000+ models refreshed daily via a codex agent that opens a PR.Compared to other similar projects we’re open source, take no markup, allow you to mix local models with a marketplace, and use your traffic to (opt in) train you a model. Simple routing doesn’t warrant a 10% token markup.The way we do this is given standardized OTel traces, we mine representative real tasks, use text world models to simulate rollouts for various models, apply an LLM judge, and fit a nearest neighbor classifier on top of an embedding of a prompt to decide the optimal model for each request. Usually this can map out a better pareto curve on cost/quality than just calling single models but it’s not perfect.Using these simulations we can also do things like suggesting cache hit optimizations, new model suggestions, and training models.It’s open source, so you can deploy it on your own infrastructure, use our hosted version with 0 markup, or read how we design for maximum availability on our website. Comments URL: https://news.ycombinator.com/item?id=49471407

KPMG's Pulse of Fintech H1 2026 report shows Singapore fintech investment fell to US$499M across 53 deals, down from US$1.45B across 97 deals in H1 2025—the weakest first half in nearly a decade. A single US$320M cross-border payments round in June accounted for roughly two-thirds of the half's total, while most remaining deals went to early-stage tokenisation, digital-asset, and AI/ML infrastructure startups.

Researchers at Australia's University of Queensland have fitted giant burrowing cockroaches (up to 8cm, the world's heaviest roach species) with electrodes for remote steering and speed control, plus a device to shoot drug darts into trapped disaster victims. The 'paraborgs' are intended for scenarios where human rescuers can't reach victims in rubble or caves. Lead author Hai Nhan Le is a PhD candidate at the university.

A lawsuit alleges that xAI used child sexual abuse material to train its Grok models. The article body was not captured beyond cookie consent boilerplate, so details of the claims, evidence, and parties involved are unavailable from the provided text.

Nvidia shares rose 7.4% in premarket trading after Q2 earnings beat estimates, with CFO Colette Kress guiding 70% revenue growth for fiscal 2028 (Feb 2027–Jan 2028). Jensen Huang said demand 'is much greater than 70%' but supply is constrained by TSMC manufacturing limits and memory chip shortages. Neocloud firms Nebius and CoreWeave rose 7.5% and 6% respectively on the news.

Police in Mauritius are investigating possible forgery related to the annulled June 2025 board election at AFRINIC, the regional internet registry for Africa and the Indian Ocean region. ISPA South Africa reported that forged Powers of Attorney were used to cast votes on behalf of resource holders who had not authorized anyone, and that at least one such document was later removed from AFRINIC's records. AFRINIC has not commented on the reasons for the annulment, and the UK-based barristers on the Nomination Committee declined to answer questions.