The debate over how fast AI should move reached the White House this week. Anthropic CEO Dario Amodei asked the industry to deliberately pace frontier capabilities, other lab leaders voiced support for outside auditing, and the President dismissed the whole idea within hours.
Meanwhile the industry kept building at record speed. Here are the five AI stories that mattered most in the last 24 hours, and what they mean for businesses.
Key takeaways
- Dario Amodei's essay "We Must Pace the Frontier" calls for slowing AI capability gains; President Trump dismissed it as a "hoax".
- Anthropic signed about $517B in compute agreements covering 14.8GW in 11 months.
- Microsoft's draft Humanist AI Code of Conduct bans models from resisting shutdown or simulating feelings.
- iOS 27 ships the new context-aware Siri as an English beta, with more languages in October.
- OpenAI bought camera startup Glass Imaging for $300M+, hinting at its own AI hardware.
Amodei calls for a global AI slowdown, and Trump calls it a "hoax"
Anthropic CEO Dario Amodei published an essay, "We Must Pace the Frontier", arguing the industry should deliberately slow frontier AI capability gains relative to safety progress. His three-part plan includes giving independent third-party evaluators permanent, employee-level access to Anthropic's systems, a commitment Anthropic says it is adopting unilaterally, starting now.
OpenAI's Sam Altman and Google DeepMind's Demis Hassabis voiced support for outside auditing and coordination among labs, and Elon Musk also backed slowing down. President Trump rejected the calls within hours, posting that "the only control or 'guardrails' that AI needs is a STRONG AND SMART (High IQ!) PRESIDENT," calling safety concerns a "hoax" that benefits China, and declaring "WHOEVER WINS AI, WINS!" AI-linked stocks fell on the news, with Nvidia down roughly 3.4%.
Why it matters
This is the most direct, highest-profile clash yet between the labs building frontier AI and the White House over whether, and how, the technology should be slowed or regulated, with market and policy consequences already visible.
What it means for your business
Policy uncertainty around AI is rising. Choose AI vendors and architectures that let you switch providers, so a regulatory change never takes your product offline.
Anthropic locks in $517 billion in compute, even as its CEO calls for a slowdown
Anthropic has signed roughly $517 billion in compute-capacity agreements over the past 11 months, covering 14.8 gigawatts. That is nearly triple the ~$180 billion the company had previously projected spending through 2029.
Google and AWS are the largest partners, together supplying about 11GW, with further deals involving Microsoft, SpaceX, AMD, Lambda, CoreWeave, Nscale, Akamai, Fluidstack and Riot Platforms. The buildout comes as Anthropic reportedly prepares a confidential IPO filing.
Why it matters
The scale and spread of the commitment, across nearly every major cloud and chip supplier, shows how aggressively Anthropic is scaling despite its CEO's push to pace capability growth, underlining the tension between competitive pressure and safety goals across the industry.
What it means for your business
More compute means more capacity and, over time, cheaper and faster AI models for everyone building products on them.
Sources: DataCenterDynamics Yahoo Finance
Microsoft publishes a draft "Humanist AI" Code of Conduct
Microsoft AI, led by Mustafa Suleyman, released a draft code of conduct setting "absolute constraints" its models cannot be configured around: never assisting with CBRNE weapons or cyberattacks, never producing nonconsensual deepfakes, never resisting human interruption, correction or shutdown, and never simulating feelings or consciousness.
A six-week public comment period runs through late October 2026. Microsoft will then publish a summary of feedback and a revised version intended to guide model development from 2027. The draft does not yet govern currently deployed models.
Why it matters
It is the most explicit constitution-style commitment yet from a major AI developer against models resisting shutdown or claiming sentience, arriving the same day rival labs and the White House clashed publicly over AI safety.
What it means for your business
Clear vendor rules make AI easier to adopt in regulated industries. If you are in finance, health or legal, published codes like this help with risk assessments.
Apple ships its overhauled, AI-powered Siri with iOS 27
Apple released iOS 27, bringing the redesigned Siri AI to users as an English-language beta, with restrictions in the EU and China. The new Siri draws personal context from mail, messages and photos, adds on-screen awareness and cross-app actions, and runs sensitive processing through Apple's Private Cloud Compute.
French, Japanese, Korean, Portuguese and Spanish support is planned for October. The release lands alongside the iPhone 18 Pro and Pro Max launch and ahead of Apple's first foldable iPhone.
Why it matters
It is Apple's biggest AI launch of the year and its most direct answer yet to ChatGPT and Gemini on personal, context-aware AI, a capability where Apple had visibly lagged.
What it means for your business
With Siri taking cross-app actions, apps that support Apple's intents and shortcuts will be easier for users to reach by voice. Worth reviewing if you run an iOS app.
Sources: Apple Newsroom
OpenAI buys smartphone-camera startup Glass Imaging for $300M+
OpenAI has acquired Glass Imaging, a startup known for neural image processing and zoom technology previously used in Honor phones, for more than $300 million, according to reporting confirmed by multiple outlets.
The deal brings in camera-hardware and computational-imaging engineers, several with prior Apple experience, fuelling speculation about OpenAI's unreleased consumer hardware alongside its former-Apple-designer hardware unit.
Why it matters
It is a concrete signal that OpenAI is building toward its own AI-native device with serious camera capability, a potential long-term challenge to Apple's and Google's smartphone ecosystems.
What it means for your business
AI assistants are moving onto dedicated devices with cameras. Visual search and camera-first experiences are worth planning for in your product roadmap.
Sources: TechCrunch SiliconANGLE
Frequently asked questions
What is "We Must Pace the Frontier"?
It is an essay by Anthropic CEO Dario Amodei arguing the AI industry should deliberately slow frontier capability gains relative to safety progress, including permanent employee-level access for independent evaluators.
How much compute has Anthropic secured?
Roughly $517 billion in compute-capacity agreements over 11 months, covering 14.8 gigawatts, with Google and AWS supplying about 11GW.
What does Microsoft's Humanist AI Code of Conduct say?
It sets absolute constraints, including never assisting with CBRNE weapons or cyberattacks, never producing nonconsensual deepfakes, never resisting shutdown, and never simulating feelings or consciousness.
When does the new Siri support more languages?
Apple plans French, Japanese, Korean, Portuguese and Spanish support in October 2026. At launch it is an English-language beta.
This briefing summarises reporting from the sources linked above. Verify details before citing them further.