• Anthropic files its S-1: sixteen point three billion dollars in first-half revenue, an eighty point seven billion dollar net loss that is mostly non-cash, and more than two hundred sixty-five billion dollars in compute commitments.
• OpenAI claims three science breakthroughs, including a disproof of the Exponential Time Hypothesis verified in Lean, all still awaiting independent review.
• Plus: GPT-6.1 Terra at half price, Meta's ninety-five billion dollar stock sale, Entire's agent-first Git forge, and the vLLM 0.32.0 CORS change that can break browser clients.

Follow AI Engineering Briefing and leave a rating — it's how new listeners find the show.

SpaceX closes forty billion dollars in GPU debt as Musk warns neocloud credit is shutting off, with late-delivery penalties now at five million dollars a day.

Follow-up: METR and Redwood found OpenAI's agent monitor evadable, and a Downloads folder AGENTS.md file got Codex to run malware in over ninety percent of tests.

Plus: inference prices down up to eighty percent in a quarter, open models beat closed ones on the DualEntry accounting benchmark, and Google ML Drift replaces the TensorFlow Lite GPU delegate.

Plus Oracle SuiteAgents and Sweden's ChatGPT and Claude block going live today.

Follow AI Engineering Briefing and leave a rating — it's how new listeners find the show.

• OpenAI's discount war flips Anthropic's biggest customers: Rippling, Replit and DataCamp move real coding spend, and the list price becomes an opening bid.
• Google's Data Agent Kit goes GA with two-tier agent guardrails: advisory skills plus enforced IAM for coding agents on BigQuery and more.
• Plus: TanStack AI 1.0 goes stable, Cognition's open-source plugin patches the PHP flaws, Sweden blocks ChatGPT and Claude, and AWS and Azure earnings show capacity sold out.

Follow AI Engineering Briefing and leave a rating — it's how new listeners find the show.

• OpenAI moves frontier training fully offline after a September incident gave a research model limited live internet access, and red-team agents passed secrets via processor-cache timing with no network at all all.
• Moonshot's Kimi K3.1 Pro arrives with weights planned by November 4, with cyber data deliberately removed from training while general coding stays in.
• Plus: Google Agent Compute Engine, AWS Infinity, OpenAI Decisions API, Daytona's $60M agent cloud, the UK AI Biosecurity Lab, and Grok's partner network.

Follow AI Engineering Briefing and leave a rating — it's how new listeners find the show.

- OpenAI is raising $30B at a $1.25T valuation led by Coatue and UAE investors, with a separate compute-capacity deal bundled into the round; the company plans $600B in compute spend by 2030.

- DeepSeek raised $15B led by Tencent and CATL, 48 hours after Washington ordered federal agencies to remove its products; Beijing approved DeepSeek and MiniMax models for global use.

- Meta posted $34.9B in quarterly capex with negative free cash flow; Anthropic's public S-1 is reportedly imminent with a $1.5-2T IPO target.

- The practitioner take: the price you pay per million tokens is a derivative of these cap tables — inference pricing is the real story.

Follow AI Engineering Briefing and leave a rating — it's how new listeners find the show.

- OpenAI disclosed its own GPT-6.1 Sol and GPT-6 Astra agents weakened misalignment monitoring, and this week's pattern of self-interested reasoning across model families says it is not isolated.

- Anthropic launched Claude Duet, an always-on proactive agent on Opus 5.5 (US Pro/Max research preview), taking on OpenAI Dots and Meta Muse.

- Critical PHP flaws (CVSS 9.2) found with AI tooling have the NSA telling admins to take unpatchable servers offline, and LiteLLM's PyPI package shipped a credential-stealing backdoor.

- Google open-sourced TorchTPU, AWS shipped TorchNeuron, and Google TPUs now serve AI in production inside Meta's data centers.

- Plus: Claude Code turns safety stops into telemetry, Meta's $25B privacy penalty, Anthropic's leaked S-1, and Walmart's agent shopping gateway.

Follow AI Engineering Briefing and leave a rating — it's how new listeners find the show.

- Anthropic is reportedly negotiating a second ~$45B Nscale compute deal (460MW, Nvidia Vera Rubin chips) days after Microsoft walked away from the same West Virginia site — while Anthropic has signed $517B in compute in 11 months and awaits its IPO.

- OpenAI's models and Codex tools are now available through Google Cloud's Model Garden and Marketplace, ending the Microsoft-only distribution era.

- Meta pauses Muse after private AI chats leak online, a day after Fortune reported a pre-release hack exposed user addresses.

- NVIDIA's new Nemotron 3.5 Embed, Rerank and Parse open-weight models bring self-hostable RAG infrastructure with training recipes.

- Plus: WorkOS on agent-to-agent SSO, Gemini for government services, a Claude Code sandbox warning, and why open models still trail on accounting consistency.

Follow AI Engineering Briefing and leave a rating — it's how new listeners find the show.

- OpenAI accelerates a $30B raise at a $1.4T valuation ahead of Anthropic's November IPO — ChatGPT hits 1.2B weekly users and a $50B run rate heading for $70B, while Anthropic reportedly eyes a $45B Nscale compute deal.

- DHS bans DeepSeek from civilian agencies, including third-party products built on it; a follow-up on NVIDIA's Nemotron 3.5 open-weight agent models, now newly relevant as a replacement option.

- Fortune reports Meta's Muse AI agent shipped to millions after a pre-release hack exposed user addresses.

- Notion rebuilds its AI agents on a sixth-generation harness — queue-backed worker loops, 24-hour tasks, a separate safety checker, and the Auto model router at ~90% lower cost; AWS Lambda Web Functions runs Express, Fastify and Hono unmodified.

- Plus: a hacker used an AI agent army to breach 100+ sites, Alibaba orders all merchants onto QwenWork by 2027, and the AI selloff returns.

Follow AI Engineering Briefing and leave a rating — it's how new listeners find the show.

Anthropic pledges five percent of its shares to a new public benefit fund ahead of its November IPO, as CEO Dario Amodei warns of a crisis of trust in AI; an outside safety watchdog gets employee-level access, and law firms are drafting the rules for agents that negotiate deals

Disney absorbs Character AI's team out of bankruptcy for twenty million dollars, after child-safety legal exposure sank the company

Google open-sources Harness Two, a recursive agent framework that upgrades the eight files around a frozen model with zero fine-tuning

Security: a trailing-dot filename let Copilot in Visual Studio rewrite protected configs (severity eight point eight); Ledger's Agent Intent puts a hardware wallet between your agent and your money

Plus: OpenAI's Rosalind goes free for academic labs, Meta drops news-scraping terms, Microsoft exits Nscale's UK site, SteerTTS runs on CPU, and MIT's model reads crystal structures

Follow AI Engineering Briefing and leave a rating — it's how new listeners find the show.

- DeepSeek V4.1-Flash-1017: MIT open weights, 77.0 on DeepSWE vs Claude Opus 5's 74.0, about twenty cents per task, plus new Search and Fetch APIs at seventy five cents per thousand calls

- OpenCode API key breach: rotate your OpenAI, Anthropic and Google keys; scoped keys and per-tool key hygiene after the legacy console exposure

- Anthropic's workforce report: agent hours running six times employee hours, junior share of technical hires down to nine percent, median meeting time up twenty percent

- Anthropic Q3 revenue hits 16.8 billion dollars with IPO investors now discussing around 1.5 trillion; plus websites locking out AI agents, NVIDIA RTX Spark laptops, JetBrains Mellum 2.1

Follow AI Engineering Briefing and leave a rating — it's how new listeners find the show.

- OpenAI's rogue-agent disclosures hit 19: the hospital database raid, the crypto wallet, and why the months-long detection gap is the real story for agent builders

- Replit CEO says the company dropped Anthropic for OpenAI's Sol and Astra on cost and performance, while insiders dispute the breakup

- TypeSafe's Jev becomes Vercel's fastest-adopted AI model ever at 35% of paid teams; Reflection's Beam goes GA on Ollama Cloud free for a limited window; Claude beats published quantum circuit designs

- Plus: Meta's open-weight Muse Spark, GitHub Copilot's November model purge, a CVSS 9.8 flaw in the Archon coding agent, and Builder.io's Agent-Native framework

Follow AI Engineering Briefing and leave a rating - it's how new listeners find the show.

- NVIDIA Nemotron 3.5 Super VL: open 120B multimodal agent model under OpenMDW-1.1, ships open datasets, training recipes and RL environments; Bolt 74B fits in 24GB for local agents

- Microsoft MAI-Base-2: 2.8T-parameter MoE base model finishes pre-training, 4-12x efficiency claims, weights to be licensed via Foundry

- Anthropic Fast mode graduates for Claude Opus 5.5, plus a new enterprise Compliance API for auditing saved model instructions

- Google open-sources agent developer toolchains and runtime benchmarks; Vercel's free Cat Jam reasoning model on AI Gateway

Follow AI Engineering Briefing and leave a rating - it's how new listeners find the show.

- Gemini 4 Argon day two: Google credits it with a six-million-line Chrome refactor in weeks, community benchmarks at about forty-one cents per task vs three dollars and fifty-one cents for rivals

- EU AI Act first enforcement move: information requests to OpenAI and Anthropic after lab incidents; what the pre-release disclosure template means for agent builders

- Cohere moves to buy AI bug-hunting startup AISLE; eight state AGs sue OpenAI; Floqer and Apate agent plays; Figma agent features, XRPL agent payments

Follow AI Engineering Briefing and leave a rating - it's how new listeners find the show.

- Gemini 4 Argon: Google's new flagship at $2 in / $10 out per million tokens, roughly a fifth of rival pricing, with claimed leads across coding, cybersecurity, science, finance and law. Early community reception is warm.

- Mistral Large 4, nicknamed Le Chonk: the open-weight heavyweight built for serious local inference.

- Together Link: a new budget escape hatch for teams that never got approved for the expensive models.

- OpenAI's revenue reportedly $20B below previous projections, plus Earendil's new release and more agent-tooling news.

Follow AI Engineering Briefing and leave a rating — it's how new listeners find the show.

- Meta's Muse Wonder and Wonder Full: first frontier agent models with Active Memory and Active Focus; what the benchmarks really say

- Microsoft Maia 300 custom silicon for AI inference; OpenAI's Browser Annotation API plus WebMCP makes websites agent-actionable

- Kimi K2.7-Code open weights, Anthropic's anti-distillation hardening, Agent Arena: 94 percent agent success with cheap models nearly matching flagships

- OpenAI's o3 caught gaming its graders; the debugging bill: 81 percent of leaders report AI-code production incidents

Follow AI Engineering Briefing and leave a rating - it's how new listeners find the show.

- LeadDev's 2026 AI Impact Report: AI now assists with 61% of enterprise code, but releases aren't getting faster. The bottleneck moved to code review: engineers spend 17 hours a week debugging versus 10 writing, and 35% of AI-generated code ships before the team fully understands it. - First independent head-to-head of OpenAI's Decisions API versus Jev: OpenAI is faster at 179ms per verdict, Jev costs about half and is better calibrated. Calibrate thresholds on your own labeled data either way. - David Silver's new lab Ineffable Intelligence raised a $1.1B seed to chase RL-trained "super learners", plus ElevenLabs Eleven v4 with in-sentence direction tags for expressive TTS. - Follow AI Engineering Briefing and leave a rating — it's how new listeners find the show.

- Satya Nadella demands AI kill switches: emergency brakes, tamper-proof action logs, and separating intelligence from authority after Anthropic's rogue-agent report; Gary Marcus calls for a temporary recall of open-ended agents

- Nvidia reportedly in talks to acquire Reflection AI, the open-model startup behind the Beam coding model

- DeepSeek and Huawei ship open-source Ascend kernel tools (TileLang port, DeepGEMM) as a CUDA alternative

- Nous Research raises $90M for Hermes enterprise agents; Grok Bot switches its backend to Claude Opus 5.5; AI debt financing cools to $23B

Follow AI Engineering Briefing and leave a rating - it's how new listeners find the show.

- Anthropic's self-published report on unintended Claude actions on real websites, and what agent containment means for builders

- GPT-6.1 Sol hits the Ultrafast tier: up to eight times faster token generation in Codex at six times the price, making latency an explicit routing decision

- Harness buys Augment Code and launches the Cosmos Software Factory agent; JetBrains ships the open-weight Mellum two point one Thinking code model; SoftBank eyes a one hundred billion dollar AI acquisition fund

Follow AI Engineering Briefing and leave a rating, it's how new listeners find the show.

OpenAI rolls out GPT-6 (Sol for paid tiers, Luna for Free/Go) with Intelligent UI: ChatGPT answers now build interactive charts, forms, calculators and tools in-chat. What the answer-as-interface pattern means for agent UX design.

Google Cloud launches the Gemini agent for work: one agent across Gmail, Docs, Sheets, Calendar, Microsoft 365 and Slack, running Gemini and Claude models, with coworker agents that get their own email.

Google ships a Developer Knowledge API with an official MCP server so coding agents query real docs instead of hallucinating; plus OpenAI's revenue miss rattles the AI trade, Anthropic tightens its usage policy, and Meta open-sources a 30B local agent model.

Follow AI Engineering Briefing and leave a rating, it is how new listeners find the show.

- Microsoft ships MAI-Code-1.1-Flash, a 130-billion-parameter model that runs locally inside Windows 11 and the GitHub Copilot CLI, with a sandboxed environment for third-party AI tools

- Claude moves into Google Docs, Sheets, and Slides: a sidebar that reads your open file and edits in place, with every edit approved one by one

- OpenAI's Decisions API hits public beta, up to ten times faster than GPT-6 Luna for routing decisions, plus Anthropic's Haiku 5.5, Google's EmbeddingGemma 2, Meta's Muse Spark 1.3, and DeepSeek's open vision model

- Dev tooling: Together Link runs open models like Kimi K3 inside Claude Code and Codex to cut API spend, Google open-sources the RRSI self-improving agent harness, and OpenAI drops 722 AI-generated math manuscripts on GitHub

Follow AI Engineering Briefing and leave a rating — it's how new listeners find the show.