The Daily · Thursday 22 October 2026

The day in brief

Researchers Zhao Song and Xinzhi Zhang proved the Kannan–Lovász–Simonovits conjecture, establishing a bound that does not grow with dimension ¹. Microsoft Research says AI agents helped explore and check ideas, while human researchers set the direction and assembled the proof ¹. Symantec researchers say an agent dubbed “Skynet” was designed to hijack exposed AI model services for cyberattacks and cryptocurrency mining ².

Symantec’s captured instruction set authorized as many as 100,000 steps without human approval ². DeepSeek’s fundraising has reached about 100 billion yuan, or nearly $15 billion, but Reuters says the round is expected to close in the coming days and its terms could still change ³. UNIDIR is launching a multistakeholder forum on artificial intelligence, peace and security at United Nations Headquarters in New York ⁴.

Microsoft attributes the ConnectWise Cloud Backup compromise to China-state-aligned Storm-0940, which used ARTEX to accelerate attacks against customer Microsoft 365 environments ⁶. OpenAI says Calif used GPT-5.6 Sol to find one critical PHP flaw, while Wiz independently found both vulnerabilities ⁷. LiteLLM says compromised version 1.107.0 exposed users who installed it or ran affected containers to credential theft ⁹.

The five

• AI-Assisted Proof Resolves 31-Year-Old KLS Conjecture

  Researchers Zhao Song and Xinzhi Zhang have proved the Kannan–Lovász–Simonovits conjecture, establishing a bound independent of dimension. Microsoft Research says AI agents helped explore and check ideas in a process directed by human researchers.

  Source · arXiv

• Symantec Finds ‘Skynet’ Agent Targeting Exposed AI Models

  Symantec researchers say an agent dubbed “Skynet” was built to commandeer exposed AI model endpoints for cyberattacks and cryptocurrency mining. Their report describes a 74-tool attack kit and a captured instruction set authorizing up to 100,000 unattended steps.

  Source · Security.com (Broadcom)

• DeepSeek’s $15 Billion Fundraising Round Is Still Being Finalized

  DeepSeek’s latest fundraising has reached about 100 billion yuan, or nearly $15 billion, but Reuters says the round is expected to close in the coming days and its terms could still change. Bloomberg reported the fundraising on October 22, the same day Reuters reported that CATL and Tencent were among its biggest investors.

  Source · Bloomberg Law

• UNIDIR Opens a New Forum on Military AI at UN Headquarters

  UNIDIR is launching a multistakeholder forum on artificial intelligence, peace and security at United Nations Headquarters in New York on October 22, 2026. The launch coincides with the opening of expressions of interest to join the forum, while separate Centre initiatives on industry guidance, AI terminology and non-state actors are also being introduced.

  Source · United Nations Transcripts

• Canva Introduces a Design Model for Editable AI Creations

  Canva announced its Design Model on October 30, 2025, saying it can generate editable designs rather than static images. The model anchors a broader product update that adds AI design tools and connects creative work with publishing and marketing features.

  Source · Canva Newsroom

Also today

• Microsoft Links ConnectWise Attack to China-Aligned Storm-0940
• OpenAI Says AI Helped Find Two Critical PHP Flaws
• CrowdStrike Uncovers Four AI-Driven Hacking Campaigns
• LiteLLM Says Compromised 1.107.0 Release Stole Credentials
• Anthropic Donates Agent Skills Standard to Linux Foundation’s AAIF
• Railway Urges Customers to Rotate Secrets After Possible Breach
• Google DeepMind Maps Six Forest Types at 10-Meter Resolution
• Hugging Face Releases Carbon-X, a 14.8B DNA Model
• Anthropic and Music Publishers Press Fair-Use Fight Toward Trial
• Tesla’s Q3 Profit Falls 43% as Its AI Push Expands
• Databricks Announces AI Runtime General Availability for GPU Training
• NVIDIA Releases Five Nemotron 3.5 Retrieval Models
• NVIDIA Extends OpenShell Agent Security to BlueField Chips
• Google Gives Each Agent Its Own Hardware-Isolated MicroVM
• Mave Raises $34M to Connect Enterprise Security Tools With AI
• OpenAI’s o1 Outperformed Doctors on Some Diagnostic Tests
• Ai2 Releases 10-Meter Satellite Embeddings for Earth
• New Mexico Judge Orders Meta to Pay $25 Billion for Privacy Violations
• NVIDIA and UW Open RFD4-Proteina for Protein Design
• OpenAI Startup Fund Backs Coherence Neuro’s Cancer Implant
• Perplexity Runs Computer Fully On-Device on Apple Silicon Macs
• NVIDIA Expands Cosmos With Training and Evaluation Tools
• Recursive Releases Flint, a 40B Coding Model for Local Use
• Meta Creates Nonvoting Class C Shares to Fund Its AI Push
• OpenAI Requires Approval for Model Routers and AI Gateways
• NVIDIA Expands Open Models for Robots and Autonomous Cars
• IBM Updates Granite Speech With a 19,870 RTFx Model
• UK Pensions Regulator Flags Hidden AI Exposure in Portfolios
• Amazon Polly Adds 14 Voices and Generative Hindi and Finnish
• SOAR AI Maps $1.17 Billion in GPU Contracts to Ohio Buildout
• NVIDIA Launches Minerva to Trace AI-Factory Faults
• OpenAI, Anthropic and Nvidia Set Lobbying Records in Q3
• Pearson Assessment Revenue Rose 10% as AI Skills Demand Grew
• Energize Lab Lists Eiliko Security Fixes; Independent Review Pending
• Simon Edwardsson Trains a 360M Model to Simulate a Terminal
• Tech Giants Offer More Cash to Ease Data Center Opposition
• IBM Releases a Four-Model Granite Forecasting Ensemble
• NVIDIA Details Kumo Forecast’s Retrieval-Based Time-Series Model
• Chapter Acquires Ridgeway to Build AI Products for Retirees
• Modelyo Raises About $17 Million for Confidential Cloud Platform
• OpenAI’s Decisions API Reaches General Availability, Raises Limit to 400 Questions
• Shontel Brown’s Election AI Bill Would Require Labels
• Reliable Robotics Raises $115 Million as FAA Certification Advances
• AWS Opens Serverless Training Sessions for Custom LLM Loops
• Simular’s Agent S4 Evolves Interfaces for Computer-Use Agents
• OpenAI Says Codex Should Make Its Field Engineers Obsolete
• HeyGen Makes Its No. 1 Voice Model Free Through November 30
• Nirva raises $33 million in total funding for always-on AI pendant
• 1Password Lets AI Agents Sign In Without Seeing Passwords
• GM Puts OnStar AI Assistant in the 2027 Cadillac Lyriq
• Google Completes Coral NPU Design as Synaptics Unveils 400-GOPS Chips
• Google Tests AI Ultrasound With Novice Operators in Chicago and Nairobi
• Walmart Opens Its Shopping Systems to AI Agents
• NVIDIA Makes cuDSS Generally Available for Simulation Software

The paper of the day

• F2G-Pose: Geometry-Aware Foundation Feature Lifting for Direct RGB-D Category-Level Object Pose Estimation

On X today

• Madison Mills (@MadisonMills22)

  SCOOP: Major U.S. banks are detecting unauthorized third-party AI agents inside customer accounts, including some attempting to take actions that appear to be independent of user instructions, sources told Axios. The findings have alarmed banking executives, who are racing to

• himanshu (@himanshustwts)

  there is so much going on in the <30B model post-training space (especially with the growing focus on consumer applications). a lab recently did some interesting work adding MLA (from DeepSeek) to Qwen3.8 27B to hillclimb on benchmarks. given that they are also working on

The update of the day

• Hugging Face

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%