How the monitored publications were assessed

Discourse valence across every scored item: what the sampled sources published and how independent models read it. It is not a measure of public opinion, not a measure of the state of the world, and article volume is not agreement.

42 items scored in this window

• AI in your oven? Anthropic warns models ‘months away’ from self-replication on smart devices
  -81 The Guardian: Artificial Intelligence 2026-10-13 moderate agreement · ±45
• Nuclear war and malicious use of AI threaten end of humanity by 2100, says Lancet assessment
  -76 The Guardian: Artificial Intelligence 2026-10-10 low agreement · ±55
• Will it be AI or the climate crisis that ends humanity? We can make it neither | Adam McKay and Philip Eubanks
  -65 The Guardian: Artificial Intelligence 11d ago low agreement · ±55
• Progressive Misalignment Can Emerge from Benign Recursive Self-Distillation
  -63 arXiv cs.LG (Machine Learning) 2026-10-19 low agreement · ±65
• Anthropic IPO warns its AI could blackmail people, become sentient, end humanity
  -61 Ars Technica: AI 12d ago low agreement · ±50
• AI’s mathematics blitz pushes science into strange new territory
  -57 The Economist: Science and Technology 2026-10-12 low agreement · ±100
• Adapting the Process of Adaptation: Darwinian, Lamarckian, and Agent-Directed Evolution in AI
  -54 arXiv cs.AI 2d ago low agreement · ±70
• Will the real Anthropic please stand up
  -50 MIT Technology Review: AI 13d ago single model
• Harness^2: Recursive Agent Harnessing for an Open World
  -50 arXiv cs.AI 2026-10-15 low agreement · ±150
• Stronger artificial intelligence breakthroughs hasten their own exhaustion
  -48 arXiv cs.CY (Computers and Society) 2026-10-21 low agreement · ±100
• Visualizing the World Model in an Agent's Value Function
  -46 arXiv cs.AI 2026-10-15 low agreement · ±165
• When Reality Outpaces Cognition: A New Scaling Principle for Reliable Intelligence
  -40 arXiv cs.AI 2026-10-16 low agreement · ±150
• AI will make it possible to design and build organisms from scratch
  -38 The Economist: Science and Technology 2026-10-14 low agreement · ±100
• Elon Musk Says SpaceX and Tesla Are Exploring a Merger
  -33 The New York Times: Technology 2026-10-21 low agreement · ±105
• Open-world AI: a Tale of Three Requirements
  -32 arXiv cs.AI 2026-10-16 low agreement · ±105
• Artificiality's First Exam: An LLM-Referenced C-test for Supraintelligence
  -31 arXiv cs.AI 2026-10-15 low agreement · ±175
• AI companies want to become frontier scientific researchers
  -30 The Economist: Science and Technology 2026-10-14 low agreement · ±110
• Is Universal Basic Income the Answer to A.I.-Fueled Job Losses?
  -26 The New York Times: Technology 2026-10-12 low agreement · ±110
• Can a System-One LLM Perform Knowledge Tracing When Few or No Learners Are Logged?
  -26 arXiv cs.CY (Computers and Society) 2026-10-09 low agreement · ±175
• Risk Ceilings and Development Deadlines: Pacing AI under Uncertain Safety Productivity
  -26 arXiv cs.CY (Computers and Society) 2026-10-09 low agreement · ±165
• Zuckerberg says Meta has reached AI ‘frontier’ and rejects calls for slowdown
  -26 Financial Times: Artificial Intelligence 2026-10-12 low agreement · ±105
• Anthropic’s IPO S-1 Spells Out a Future of Promise and Peril
  -20 The New York Times: Technology 13d ago low agreement · ±100
• Stochastic Dynamics of Generative AI: From Autoregressive Models to Interacting LLM Networks and Neural SDEs
  -16 arXiv cs.LG (Machine Learning) 2d ago low agreement · ±110
• The World Won't Wait: Evaluating and Teaching Agents Under a Live Clock
  -14 arXiv cs.AI 2026-10-16 low agreement · ±150
• AI’s transformation of mathematics is a turning point for science
  -12 The Economist: Science and Technology 2026-10-12 low agreement · ±150
• Navalia: Advancing Cowork Agents through Verifiable Task Synthesis and Long-Horizon Post-Training
  -9 arXiv cs.AI 2026-10-21 low agreement · ±150
• Gemini 4 Argon: Built to solve your hardest problems
  -6 Google AI Blog 2026-10-13 low agreement · ±170
• SlackBench: Benchmarking Agents on Collaborative Projects Grounded in Real Code Repositories
  -5 arXiv cs.AI 3d ago low agreement · ±150
• Gemini 4 Carbon: Our next frontier model
  -5 Google DeepMind 11d ago low agreement · ±175
• The Trinity:Capability-Driven Recursive Self-Improvement for Multimodal Agents
  -4 arXiv cs.AI 11d ago low agreement · ±150
• Genie 4: A new generation of world simulation
  -3 Google DeepMind 12d ago low agreement · ±165
• AI researchers stranded Claude on a virtual desert island. It started to “play.”
  -3 MIT Technology Review: AI 2026-10-19 low agreement · ±150
• “We were looking at it wrong”: The explosive claim that’s upended the AI world
  0 MIT Technology Review: AI 2026-10-19 low agreement · ±150
• Becoming real
  0 Nature 2026-10-14 high agreement · ±0
• Anthropic warns of 'catastrophic risk' to humankind ahead of blockbuster share listing
  +2 BBC News: Technology 13d ago high agreement · ±5
• How we detect frontier model misalignment
  +7 OpenAI 10d ago low agreement · ±165
• EvoBelief: Falsifiable World Models for Self-Improving LLM Agents
  +20 arXiv cs.AI 2026-10-12 single model
• TACS-RF: Tool-driven Agentic Circuit Synthesis of RF and mmWave Modules
  +27 arXiv cs.AI 13d ago moderate agreement · ±38
• DeepMind unveils Genie 4 world model promising more realistic physics and humans
  +50 Ars Technica: AI 12d ago single model
• Solving a 31-year-old grand challenge in mathematics – the KLS conjecture
  +50 Microsoft Research 2026-10-21 single model
• EvolvingBrain: An Adaptive Skill System for Multi-modal Agents Self-Evolution in Open Worlds
  +50 arXiv cs.AI 2026-10-16 single model
• Transformer-Based Masked Diffusion Models Provably Learn Distributions with Latent Structure
  +50 arXiv cs.LG (Machine Learning) 2026-10-12 single model

Across the whole window

The two tables below describe every item in the last 30 days under agi-timelines. They are deliberately not narrowed by the reading filter, because a mean taken only over items already selected for their score would just restate the filter.

By topic

By source

A source's mean says how the items it published were read. It is not a rating of the source.