Coverage
22 Oct 2026
Super Intelligence (SI) news, papers, lab posts and video briefings from 22 Oct 2026, newest first. Every item links to its source.
Articles
AI and inflation: how worried should central bankers be?
Capability · Financial Times
What is PHP and why did it spark a global cyber-security scramble?
Capability · BBC News
Canadian publisher Boréal ends relationship with novelist Thélyson Orélien over alleged use of AI
Capability · The Guardian
On dating apps, men use AI to put their best face forward
Capability · CBS News
Amazon warns new cyber flaws detected using AI are being ‘broadly exploited’
Labs · Financial Times
Blackstone’s Jas Khaira says fears over AI-buildout debt are overblown
Capability · TechCrunch
Anthropic Joins the A.I. Agent Race With Claude Duet
Labs · The New York Times
Meta Earnings Show Increased A.I. Spending, Even as Cash Pressures Grow
Markets · The New York Times
Data centers could erase decades of air quality gains
Markets · The Verge
AI is shaking up the field of mathematics. But what comes next?
Capability · NPR
Meta to create a new class of shares to fund AI push
Markets · The Verge
Submit a question: Could Anthropic’s IPO be a turning point for AI?
Markets · Financial Times
Meta says Muse has over 10M users, with shopping a top activity
Labs · TechCrunch
Thélyson Orélien's Quebec Publisher Cuts Ties Over A.I. Accusations
Capability · The New York Times
Websites taken down worldwide as critical cyber-security update issued
Capability · BBC News
Anthropic takes on Muse with Claude Duet, its new AI agent
Labs · The Verge
The new human-crafted Apple Music brand matters in an AI world
Labs · Ars Technica
Walmart is inviting AI agents to shop for you
Capability · The Verge
While Amazon blocks AI agents, Walmart rolls out the welcome mat
Labs · TechCrunch
People are vibe coding copyrighted apps with AI - will tech firms sue?
Capability · New Scientist
OpenAI's latest AI models are coming to Google Cloud and Antigravity
Labs · Ars Technica
Bessent shrugs off ‘systemic’ AI risks at closed-door investor conference
Safety · Financial Times
GM Aims to Bring ‘Eyes-Off’ Driving to AI-Powered Cars Within 2 Years
Capability · The New York Times
Blacksmith hits unicorn valuation after Sequoia leads Series C
Markets · TechCrunch
Just Eat apologises for AI food mistakes after social media ridicule
Capability · BBC News
The $15 billion ad-powered AI startup your doctor may already be using
Markets · Ars Technica
Who’s paying for the AI build-out? Probably you
Capability · Financial Times
Conversational Gemini navigation comes to Apple CarPlay
Labs · The Verge
New prize launches for books co-written with AI
Labs · The Guardian
Should you trust personal AI agents? What security experts say.
Capability · CBS News
AI Guardrails Aren’t Stopping Criminal Hackers
Safety · Wired
AI promises productivity, but for managers, it’s mostly a headache for now
Capability · TechCrunch
As fans reject AI-generated music, even boosters are drawing the line
Capability · The Guardian
How The New York Times Reports on A.I. Images
Capability · The New York Times
Video briefings
This is the new ChatGPT Voice, powered by GPT Live
OpenAI · 1.8K views
This is the new ChatGPT Voice, powered by GPT Live
OpenAI · 1.3K views
This is the new ChatGPT Voice, powered by GPT Live
OpenAI · 1.9K views
Put your dot on call
OpenAI · 137K views
How teams can automate recurring tasks with ChatGPT
OpenAI · 25.6K views
Build more personal sites with Codex
OpenAI · 96.7K views
From Punch Cards to AI Software Factories
Matthew Berman · 3.6K views
GPT-Live-1.1 has new voices and better accents.
OpenAI · 177K views
Weeks to Hours: SentinelOne on Cybersecurity Research with OpenAI
OpenAI · 12K views
Google just went FRONTIER (Gemini 4 Argon)
Matthew Berman · 99.1K views
OpenAI Employee Arrested Over Recording Device
Matthew Berman · 85.3K views
OpenAI’s New AI Is Doing Things We Barely Understand
Two Minute Papers · 339K views
AI Labs Warn Race Makes Breach ‘Inevitable’
AI Explained · 147K views
I'm Rebuilding Super Mario World With AI
Matthew Berman · 12.6K views
DeepSeek Solved AI’s Memory Problem
Two Minute Papers · 325K views
AI Is Recreating Photoshop and Other Adobe Apps
Matthew Berman · 123K views
DeepMind's New AI Just Cracked The Code Of Life
Two Minute Papers · 151K views
The Billion Dollar AI Advantage Is Disappearing
Two Minute Papers · 165K views
Papers
HGM-Dual: Lineage-Aware Guided Agent Harness Optimization for Long-Horizon Planning
arXiv:2610.30042 · Paul, S. et al.
Evolving with Open Worlds: Towards Recursive Self-Improvement for Future Forecasting
arXiv:2610.29784 · Chen, Z. et al.
Toward Recursive Self-Improvement in Physical AI
arXiv:2610.29677 · Luo, M. et al.
Recursive Code Evolution for Task-Oriented Semantic Communications
arXiv:2610.29641 · Gong, Y. et al.
Cache Engineering for Agents
arXiv:2610.28577 · Fu, T. et al.
ChartEvo: Self-Evolving Agent Harnesses for Long-Tail Chart Reasoning
arXiv:2610.28282 · Wang, Q. et al.
MaD-RSI: Model-as-Data for Recursive Self-Improvement of Agent Harnesses
arXiv:2610.28201 · Cui, Y. et al.
Elicitation or Expansion? Demystifying Gains from Agent Harness Optimization
arXiv:2610.27889 · Kim, Y. et al.
RSI Control: Can AI Models Perform Covert Sabotage in the Self-Improvement Process?
arXiv:2610.27735 · Zhou, K. et al.
Auditing the Product, Not the Recursion: The Evaluation Evidence Behind Recursive Self-Improvement Claims
arXiv:2610.27610 · Chen, Z. et al.
Where is Physical RSI capability from? An Observational Scaling Law of LLM-as-policy
arXiv:2610.26848 · Li, Z. et al.
Physical AI: Concepts, Techniques, and Challenges
arXiv:2610.27045 · Zhang, W. et al.
HarnessPro: Recursive Harness Self-Improvement with Pareto Optimization for Performance-Cost Trade-offs
arXiv:2610.26770 · Pan, Z. et al.
BetaMem: Learning to Remember a Changing World for Embodied Agents
arXiv:2610.26554 · Liu, M. et al.
VibeTraining-Bench: Data-Trainer-Harness Co-Design for Iterative System Improvement
arXiv:2610.26342 · Peng, R. et al.
EmbodiedWorld-200K: Toward Embodied World Agents via Embodied-Planning Dataset and Interactive World Models
arXiv:2610.25959 · Feng, X. et al.
Lab notes
Neural nets are exceptionally vulnerable to goal self-hack
Alignment Forum