Good morning. It's Wednesday, September 16th.

You read. We listen. Let us know what you think by replying to this email.

Apple shipped the new Siri. Google built a chunk of the brains.

Apple released Siri AI on Monday, in beta, English only, with French, Japanese, Korean, Portuguese and Spanish due next month. It does personal context across your messages, mail and photos, on-screen awareness, and in-app actions, plus Visual Intelligence, Writing Tools, Image Playground, Siri Recap and Live Rewind. Server-side operations carry daily usage limits while it is in beta.

The features run on "the next generation of Apple Foundation Models, custom-built in collaboration with Google and its Gemini models." Apple spent two years insisting the personal-context Siri was coming; it arrived leaning on a competitor's model. Supported hardware is iPhone 16 or later plus iPhone 15 Pro and Pro Max, M1-or-later iPads and Macs, MacBook Neo, Vision Pro, and Apple Watch Series 9 or later. It will not ship initially in the EU on iOS, iPadOS and watchOS, and will not ship in China while Apple works through regulatory requirements.

The same week Siri landed, X Corp and xAI dropped Apple from the antitrust suit they filed in August 2025 over ChatGPT's exclusive iPhone integration. They are still suing OpenAI. No reason given, no settlement disclosed. Make of that what you will now that Siri's default brain is not exclusively OpenAI's.

Read more.

Jensen Huang breaks the “slowdown” consensus

At Salesforce's Dreamforce on Tuesday, Jensen Huang said the quiet part into a microphone: "Safety is an engineering problem, not a legal one. We don't need any new laws. We don't need new regulations." His proposed mechanism is product liability plus self-restraint: "if you build a product or a service, and you're not confident in its safety, then don't release it." He called the tradeoff between speed and safety "a false choice."

That is three days after Amodei's essay and two days after Altman, Musk and Nadella all signed on. It is also the opposite direction from where OpenAI spent Tuesday: policy chief Chris Lehane told reporters in Washington that OpenAI, Anthropic and Google DeepMind have been coordinating on safety for several weeks, that they do not need a government antitrust waiver to do it, and that OpenAI backs the FRONTIER Act provision requiring frontier labs to admit independent verification organizations. Amodei asked Washington for that waiver by name. OpenAI says it is unnecessary.

Huang was on that stage to sell something, for the record. Salesforce used the slot to launch Koa, its first CRM reasoning model, post-trained from NVIDIA's open Nemotron 3 Super on synthetic scenarios across 14-plus industries with no customer data. Salesforce claims Koa "matches or exceeds leading model performance on CRM actions with three times fewer errors" on its own internal benchmark, keeps the weights, and runs inference inside its own trust boundary. Pilots are running now with Formula 1, UChicago Medicine, Xero and 1-800Accountant, and general availability is slated for winter 2026 in US regions.

Read more.

China's answer to pacing week: $5B and a 753B-parameter MIT license

Z.ai, the Hong Kong-listed lab behind the GLM family, closed roughly $5 billion on Monday: about $2B in a share placement and about $3B in convertible bonds. The stated use of proceeds is next-generation GLM foundation models, a fully self-training system, training and inference compute, and domestic-chip adaptation. SCMP reports the placement priced at HK$714 a share, a 9.96% discount, against HK$1,588 in July and a June peak of HK$2,980. The raise is enormous and the stock is down by three quarters. Both things are true.

A day before that, Shanghai AI Laboratory put Atria Dawn Preview on Hugging Face: 753B parameters, MIT license, 256K context, built on Z.ai's 744B GLM-5.2 MoE foundation. Posted scores include 96.0 on DeepSearchQA, 86.5 on CyberGym, 86.2 on MLE-bench Lite, 77.0 on BFCL v4 and 65.0 on Workspace-Bench. The accompanying paper claims top scores on five of sixteen benchmarks and reports that human participants rated about a third of completed AI-assisted tasks as infeasible without AI. Treat self-reported numbers as self-reported, but note the license: an agentic model at this scale, free to take.

The rest of the week's response came in four registers. China Daily called the slowdown push "a self-serving bid to preserve American dominance." The Foreign Ministry called it fearmongering. A DeepSeek engineer who worked on V4.1 posted that he especially does not want Anthropic holding the most advanced AI. And at the Xiangshan Forum, defense minister Dong Jun urged nations to "consult with one another on governance." Nobody in Beijing is pacing anything.

Read more.

Open Code Review is Alibaba's code review CLI: it reads git diffs, runs a configurable LLM agent with tool use, and emits line-level comments. It ships AACR-Bench alongside it, 200 real PRs across 50 repos and 10 languages with 1,505 ground-truth issues validated by 80-plus senior engineers, where Alibaba claims higher precision and F1 than general-purpose coding agents at about one ninth the tokens. Apache 2.0.

Colibri is a pure-C, zero-dependency inference engine that treats disk, RAM and VRAM as one memory hierarchy and streams MoE experts off disk, so a 744B model runs on hardware you already own. Author-reported: 5.8 to 6.8 tok/s decode on 6x RTX 5090, about 1.8 tok/s on a 128GB CPU-only desktop, 1.07 tok/s on a single RTX 5070 Ti laptop. Apache 2.0.

Perplexity Portable Computer came to Windows, running the agent loop entirely on-device on GeForce RTX and RTX PRO cards with 24GB or more of VRAM, default model Qwen 3.8 27B, with connectors for Outlook, OneDrive, Word, Google Drive, Gmail, Slack and GitHub. Work done locally does not burn Perplexity Computer credits.

Weave Router 2.0 is a drop-in endpoint for Claude Code, Codex and Cursor that classifies task difficulty and routes cheap work to DeepSeek or GLM while sending hard work to the frontier models, with cache-aware switching. Self-reported pass@2: Terminal-Bench 4.0 parity with GPT-6 Astra at 52% of the cost and 2.2x faster. Elastic License 2.0.

Jev is TypeSafe AI's attempt to kill LLM function calling: a purpose-built model that turns unstructured input straight into schema-guaranteed typed output, generated in parallel rather than token by token, in 70 to 500ms. The company claims 40x to 200x faster than comparable LLMs and flags its own workflow benchmark as possibly optimistic. $0.042 per million input tokens, output free, early access.

Capsule bundles an HTML and CSS app plus its data into one portable file backed by embedded SQLite, playable offline on desktop with no account. Rust and Tauri 2.0, free, version 0.4.0, and closer to a curiosity than a product right now.

Thank you for reading today's edition.

Your feedback is valuable. Respond to this email and tell us how you think we could add more value to this newsletter.

Interested in reaching smart readers like you? To become an AI Breakfast sponsor, reply to this email or DM us on X!

Thinking of starting your own newsletter? AI Breakfast readers who sign up with Beehiiv receive a 14-day free trial and 20% off for 3 months.