News
Trump AI hoax, China model drop
US
Trump calls AI risks a ‘hoax,’ says there is a ‘SICK conspiracy’ against AI and data centres
U.S. President Donald Trump stated on Monday that he would serve as a sufficient guardrail against rogue AI, dismissing efforts to limit the technology as part of a “SICK conspiracy.” He emphasized existing federal criminal and regulatory powers over AI companies and argued that slowing development would primarily benefit China, declaring “WHOEVER WINS AI, WINS!”
Nvidia CEO Puts Trump on Speakerphone While Downplaying AI Risks
At the All-In Summit in Los Angeles, Nvidia CEO Jensen Huang received a live call from President Trump on stage. Trump dismissed AI takeover fears as a “hoax,” stating “the robots will not be taking over.” Huang agreed the U.S. would ensure every industry, company, and state wins in the AI race, while noting safety remains paramount but rejecting unscientific extinction predictions.
Microsoft drafts code of conduct to keep its AI under human control
Microsoft published a draft 37-page “Humanist AI Code of Conduct” for its MAI models, requiring them to remain subordinate to human control, accept interruption and shutdown, avoid self-directed goals or opaque communication, and observe absolute constraints against weapons, child safety violations, and large-scale manipulation. The draft is open for six weeks of public consultation before guiding 2027 model development.
Apple releases iOS 27, redesigned Siri AI
Apple began rolling out Siri AI in English beta as part of its 2027 software releases (iOS 27 and related OS updates). The redesigned assistant features personal-context retrieval from messages, email, and photos; onscreen awareness; web-backed answers; cross-app actions; and a dedicated conversation app. Support for additional languages follows next month; availability is delayed in the EU and China due to regulatory reasons.
OpenAI Says It’s Working With Anthropic, Google on AI Safety
OpenAI confirmed it has been coordinating with Anthropic and Google DeepMind for several weeks on AI safety measures. The company stated it does not require an antitrust waiver to collaborate on these issues and prioritizes joint efforts to address safety concerns.
Google DeepMind staffer says AI may ‘kill us all’ in exit post
A Google DeepMind AI Safety and Alignment researcher publicly resigned, stating belief that AI has the potential to kill humanity. This follows similar high-profile warnings and resignations from Anthropic researchers earlier in the period.
US advancements in consumer AI deployment (Apple) and internal safety frameworks (Microsoft) demonstrate continued leadership in practical application and governance innovation, positioning American firms to shape global standards while expanding accessible capabilities. Shortfalls in coordinated industry pacing amid political pushback risk fragmented progress and potential competitive exposure if safety gaps widen.
Non-US West
King Charles to host AI leaders after dire warnings from researchers
Britain’s King Charles will convene leaders from Nvidia, Google DeepMind, OpenAI, Anthropic, and others this week in Scotland at Dumfries House to discuss shared principles for AI development that benefit society and uphold human dignity. The gathering, organized with the Ditchley Foundation, aims to explore responsibilities around the technology amid rising internal industry warnings.
European and UK efforts center on high-level convening to foster principles-based dialogue without immediate regulatory mandates. Progress remains measured, with limited concrete enforcement mechanisms emerging alongside the discussions.
China
Shanghai AI Lab drops Atria Dawn, a 744B agentic MoE
Shanghai Artificial Intelligence Laboratory released Atria Dawn Preview, a 744B-parameter agentic Mixture-of-Experts model built on GLM-5.2 and trained via a Verifiable Experience Pipeline. Available open-weight under MIT license on Hugging Face, it reports competitive or leading scores across multiple agentic benchmarks focused on research, engineering, tool use, and multi-step tasks, with demonstrated capabilities in areas such as weather model design and OS construction.
China bristles at Anthropic CEO’s ‘fearmongering’ about its AI development
China’s Ministry of Foreign Affairs rejected Anthropic CEO Dario Amodei’s calls related to curbing Chinese AI capabilities as “fearmongering.” Officials released AI Safety Governance Framework 3.0, updating risk categories and technical/governance recommendations, while emphasizing multilateral UN-centered approaches and balanced development with security.
China Spy Chief Warns of AI Risks as Anthropic CEO Urges Brakes
China’s top intelligence official warned that rapid AI advances could threaten political stability, institutional security, and critical infrastructure, citing risks from adversarial use and advanced foreign models.
Chinese releases of large open agentic models expand accessible capabilities, though state-driven frameworks and security-focused rhetoric highlight tight control priorities that may constrain broader collaborative innovation and raise concerns over dual-use applications.
Non-China East
No highly notable breaking developments meeting the criteria of major global impact emerged in the scanned period from other East Asian sources.
Regional activity remains steady without standout frontier announcements or shifts in the immediate window. Shortfalls in visible high-impact model releases or policy pivots leave the area in a monitoring posture relative to larger players.
Grok / xAI / Elon Companies (AI, Robotics, Supporting Tech)
Elon Musk, Grok Bot summary of All-In Summit interview, post
Grok @Bot summary
(Exact content covers Musk’s discussion of AI dangers including the Hugging Face incident, proposed peer-review testing among competitors before release, agreement that risks are significant and rising, skepticism of certain extinction framings, Starship progress, TeraFab chip ambitions, and Roadster tease.)
Elon Musk, AI data centers resulting in lower electricity prices for consumers, post
AI data centers resulting in lower electricity prices for consumers
Elon Musk, Our C/C++ training was written by humans..., post
Our C/C++ training stack was written by humans. AI is not yet good enough for extremely high performance software (but it will be).
SpaceXAI / xAI staff will livestream building a startup from scratch over 72 hours using Grok Bot agents (Grok Bot Galaxy event, September 15–17 in San Francisco with global stream). Three team members start with no predetermined idea, product, or name, using the multi-agent system across engineering, product, sales, marketing, and support workflows. This follows SpaceX’s acquisition of xAI and related coding tools, highlighting agentic automation for business creation.
Grok and related xAI/SpaceX efforts emphasize practical agentic tools for rapid company-building and high-performance software stacks, alongside infrastructure commentary linking AI compute to consumer benefits. These steps advance automation in software and operations while peer-review proposals address safety coordination.
Summary
The period is dominated by a sharp industry-internal debate over pacing frontier AI development after researcher warnings and agent incidents, met with political resistance in the U.S. favoring continued acceleration, new safety codes and consumer deployments from major labs, open agentic model releases from China, and royal-level convening in the UK. Elon-linked efforts focus on live agentic automation demonstrations and infrastructure pragmatism amid the broader safety discussion.
Term
Domain
The name you type into a browser, like yourname.com. You rent it from a registrar. It is the address people put on an invoice, not the website files themselves.
Acronym
DNS
The internet's address book. When someone types your domain, DNS tells their computer which server actually holds your site or your mail.
Acronym
MX
A DNS record that says which service should receive email for your domain. If MX is wrong, mail to you@yourdomain bounces or lands in the wrong inbox.
Acronym
IMAP
The usual way a phone or laptop mail app talks to your mailbox. Mail stays on the server, so the same inbox shows up on every device.
Term
Registrar
The company that rents you the domain name and lets you point it somewhere. Cloudflare Registrar and Porkbun are registrars. Squarespace or Google can also act as one if you bought the name through them.
Term
Nameserver
The computers, named by the registrar, that answer DNS questions about your domain. Changing nameservers is how you hand DNS from one host to another.
Acronym
2FA
A second check after the password, usually a code from an authenticator app or a hardware key. SMS codes are weaker. You type 2FA. An agent does not.
Term
Stack
The small set of tools you actually run the company on: mail, invoicing, calendar, a site. On this site, a stack page is one job with a named pick and a named skip.
Term
Agent
Here, an AI helper that can draft files, checklists, and clicks you can undo. It does not hold passwords, 2FA, the card, or the final words on a public page.
Acronym
CRM
Software for tracking people you sell to: notes, follow-ups, a pipeline. A spreadsheet or starred mail is enough until those follow-ups stop fitting in a table.