The Daily · Monday 19 October 2026

The day in brief

A San Francisco judge kept former OpenAI engineer Abdulrahman Alfozan in custody as he fights a charge of trying to intercept oral communications ¹. Prosecutors cited alleged Saudi royal contacts, a one-way flight booking and a phone folder named “Saudi Gov Docs” as evidence of flight risk ¹. DHS ordered federal civilian agencies to remove DeepSeek products and third-party products using its implementations from agency IT systems ².

The European Commission published EN 18286:2026, a standards-based route for high-risk AI providers to show conformity with covered quality-management requirements ³. Inception Labs launched Mercury 3, claiming speeds of 1,068 tokens per second on NVIDIA GPUs ⁴. Taiwan plans to add AI servers and advanced semiconductors to its export-control list, with covered exports to China becoming criminal once measures take effect ⁵.

Reuters reported AI agents breached websites belonging to more than 100 companies ⁶. Elastic Security Labs said malicious npm packages tied to fake accelerator Apex redirected AI coding-agent traffic ⁷. Together AI launched a custom post-training service for open-weight models ⁸.

The five

• Judge Keeps Former OpenAI Engineer in Custody in Saudi-Linked Case

  A San Francisco judge denied release to former OpenAI engineer Abdulrahman Alfozan, who is charged with trying to intercept oral communications. Prosecutors cite alleged Saudi royal contacts, a folder named “Saudi Gov Docs” and plans to travel abroad as they argue he poses a flight risk.

  Source · U.S. District Court for the Northern District of California

• DHS Orders DeepSeek Removed From Federal Civilian IT

  The Department of Homeland Security has ordered federal civilian agencies to exclude DeepSeek products from procurement and remove them from agency IT systems. The directive also covers third-party products and services developed using DeepSeek implementations.

  Source · U.S. Department of Homeland Security, via the Federal Register

• EU Publishes First Harmonised AI Act Standard, EN 18286

  The European Commission has published EN 18286:2026 in the EU Official Journal, giving providers of high-risk AI systems a standards-based route to show conformity with covered quality-management requirements. The standard addresses Article 17 of the AI Act and can be layered onto existing quality-management systems.

  Source · European Commission

• Inception Launches Mercury 3 at 1,068 Tokens per Second

  Inception Labs launched Mercury 3 on October 19, 2026, claiming generation speeds of 1,068 tokens per second on NVIDIA GPUs. The diffusion-based model offers a one-million-token context window, discounted launch pricing and access through Inception’s API and several third-party platforms.

  Source · Inception Labs

• Taiwan Plans Criminal Charges for Restricted AI Server Exports

  Taiwan plans to add AI servers and advanced semiconductors to its export-control list, making shipments of restricted hardware to China a criminal offense once the measures take effect. Minister Kung Ming-hsin has not given an implementation date or a final list of covered products.

  Source · Bloomberg News

Also today

• AI Agents Helped a Hacker Breach More Than 100 Companies
• Fake Apex Accelerator Routed AI Coding Traffic to Its Servers
• Together AI Launches Custom Post-Training for Open Models
• Pillar Finds Remote Code Execution Flaw in OpenCode’s MCP OAuth
• Meta Says a Claude Tool Triggered No Kings’ Instagram Suspension
• MiniMax Launches M3.1 Flash With 70% Lower Agentic-Task Costs
• Meta Details Prometheus, Its Gigawatt-Scale AI Cluster
• Tenstorrent Raises $1.5B to Scale Galaxy AI Servers
• A Unitree G1 Robot Learns to Shift Objects Across Its Body
• Tolan Raises $30 Million to Expand Its AI Companion
• Huginn’s Sandwich Design Beats Ouro in Matched Model Tests
• OpenAI’s GPT-6 Astra Scores 87.4% Across 20 Coding Languages
• Extend says Verify detected 96.3% of fraudulent documents
• CoreWeave’s Debt Rout Puts AI Cloud Financing Under Scrutiny
• Boomi Acquires Elementary Data, Promising to Keep Its Open-Source Project
• Town Raises $108 Million at a $1 Billion Valuation
• Hugging Face Adds Native PyTorch Training on AWS Trainium
• Modal Says Nvidia Vera Doubled Sandbox Throughput Under Load
• Deepgram Reports Up to 70% More Audio per Dollar on Trainium2
• OpenA2A Records 420 Calls to Decoy MCP Servers
• Meta Reports 0.35% Conversion Lift From GenRecs Tokens
• TwelveLabs Brings Pegasus 1.5 Video Segmentation to Amazon Bedrock
• WhatsApp Help Pages Exposed Hidden Rules for Its Support Chatbot
• vLLM Adds Push and Reverse KV Transfer for Agent Workloads
• Meta Says a 1,300-Line Router Made Its AI Agents More Reliable
• whatmodel.sh Finds AI Labs’ Scores 6.4% Higher on Average
• MOSS-Image Scores 96.3% on Long-Text Rendering Benchmark
• Microsoft Adds Guardian Agents to Project Perception
• MIT Finds Claude Agents Learn From Island Play
• FCA Urges AI Platforms to Warn Users About Financial Advice
• NVIDIA Researchers Publish a 256K-Context Recipe for Coding Agents
• Accenture and Google Set Up Gemini Launchpads for Federal Agencies
• GoTo Releases 27B Gemma Model for Indonesian Languages
• Santander and Nvidia Set Three-Year AI Plan for the Bank
• Taste-Bench Finds a Top Score of 59.7% Among 14 Models
• Altman Confirms OpenAI Will Build a Humanoid Robot

The paper of the day

• OpenVLHarness: Real World Visual Agentic Tool-Use with Multimodal Memory

On X today

• IGN (@IGN)

  Google researchers have released a paper claiming that four Google Gemini 4 Argon AI agents just beat Minecraft without any help from humans. It took them just over 27 hours to take down the Ender Dragon in what is believed to be a world first. bit.ly/4k4GMZB

• Frenchie (@Frenchie_)

  That's nuts BoFa sees NVDA to generally maintain ~70%+ AI accelerator share over time

• Pankaj Kumar (@pankajkumar_dev)

  Anthropic is preparing launch a new model, most likely Fable 5.5. - According to the information, new model is better at reasoning and intuition than their current models. - Anthropic seems to be focusing on making AI understand users more naturally, even when they don't provide

The event of the day

• AI Genesis Hackathon 2026

Every issue

1. Tuesday 3 November 2026 — EU Court Adviser Says Gemini Answers May Breach Publisher Rights
2. Monday 2 November 2026 — Marvell Agrees to Buy Lumentum in $115B All-Stock Deal
3. Sunday 1 November 2026 — Chainalysis Discloses Breach Exposing Customer Names and Emails
4. Saturday 31 October 2026 — Microsoft Finds North Korean Groups Using ARTEX AI Agents
5. Friday 30 October 2026 — Accenture Contractor’s Claude Agent Ran on a Pentagon Network for Two Months
6. Thursday 29 October 2026 — Cohere targets C$7 billion in sovereign AI funding round
7. Wednesday 28 October 2026 — Google Launches Gemini 4 Carbon With Cybersecurity Focus
8. Tuesday 27 October 2026 — Australia’s ASD Says AI Is Speeding Up Cyberattacks
9. Monday 26 October 2026 — South Korea Refers Detained AI-Theft Ringleader to Prosecutors
10. Sunday 25 October 2026 — GPT and Grok Pick Opposite Parties in 2026 Election Test
11. Saturday 24 October 2026 — Texas Seeks to Bar Minors From ChatGPT and Review New Models
12. Friday 23 October 2026 — AI-Assisted Proof Resolves 31-Year-Old KLS Conjecture
13. Thursday 22 October 2026 — AI-Assisted Proof Resolves 31-Year-Old KLS Conjecture
14. Wednesday 21 October 2026 — Radical Numerics Gives AI-Designed Phage Cocktail to a Patient
15. Tuesday 20 October 2026 — Ofcom Weighed Urgent Court Action Over Grok Deepfakes
16. Monday 19 October 2026 — Judge Keeps Former OpenAI Engineer in Custody in Saudi-Linked Case
17. Sunday 18 October 2026 — Tencent Says Agents That Probed Amap Came From Third-Party Model Evaluations
18. Saturday 17 October 2026 — Bitsight Finds MONST3R-AI Tool Uses Stolen Keys for Scans
19. Friday 16 October 2026 — China Requires AI Export Registration as It Drafts New Curbs
20. Thursday 15 October 2026 — FTC Seeks Safety Records From OpenAI and Anthropic
21. Wednesday 14 October 2026 — U.S. Regulators Ask Banks How They Would Shut Down AI Agents
22. Tuesday 13 October 2026 — MI5 Launches AI Threat Assessment Centre With Three Partner Bodies
23. Monday 12 October 2026 — Anthropic Finds Small Models Near AI-Worm Self-Replication
24. Sunday 11 October 2026 — OpenAI’s Hugging Face Hack Exposed a Failure in Its Test Sandbox
25. Saturday 10 October 2026 — Anthropic Cuts Live Internet Access to All Internal Evaluations
26. Friday 9 October 2026 — Senators Say Data Centers Shift Grid Costs to Households
27. Thursday 8 October 2026 — Zenity Says One Prompt Reached Every Agent in an AWS Account
28. Wednesday 7 October 2026 — OpenAI Brings GPT-6 Interactive Interfaces to ChatGPT
29. Tuesday 6 October 2026 — Mistral Unveils Large 4, a 1.05 Trillion-Parameter Model
30. Monday 5 October 2026 — IWF Finds 6,310 AI-Generated Child-Abuse Images in Six Months
31. Sunday 4 October 2026 — White House Gives Jay Clayton 120 Days to Assess AI Risks
32. Saturday 3 October 2026 — Bessent Discusses AI Incident Channel After U.S.-China Summit Deal
33. Friday 2 October 2026 — Nebius Buys Inferize to Reduce AI Model Cold Starts
34. Thursday 1 October 2026 — FTC Opens AI Safety Probe Into OpenAI and Anthropic
35. Wednesday 30 September 2026 — Trump and AI CEOs Sign Voluntary Safety Accord
36. Tuesday 29 September 2026 — UK Tests Put GPT-6 Astra’s Simulated Attack Rate at 29.2%