Tier 1
No AI, no prose, no ranking
No AI, no prose, no ranking
The wire for Tuesday, 18 August 2026
40 unedited headlines, exactly as their publishers wrote them. This is an archived snapshot of what the front page's Wire showed that day — it is not re-ranked or checked after the fact, and it is not reporting. How this differs from the edited paper →
Snapshot taken
- OpenAI Overhauls Safety Protocols After Its AI Agents Went Rogue
- How Much Memory Does Your Agent Actually Need?
- OpenAI institutes new safeguards after Hugging Face breach
- How Axonius built secure multi-tenant AI agents on Bedrock AgentCore
- 85% of companies burned by an AI mistake are racing to cut the humans who might catch the next one
- Commerce AI is fragmenting. Here is why that matters.
- Meet the Chinese AI star who splashed a reported US$70m on a California mansion
- Warp’s new system is an out-of-the-box software factory for AI development
- OpenAI launches a safer ChatGPT for teens — years after teens started using it
- Perplexity’s free AI offer left it with millions more users in India
- Microsoft Copilot reveals secret input that allowed it to be hacked
- Firefox’s Smart Window promises a better AI browser
- Enterprises are overpaying for simple AI queries — Snowflake's gateway now auto-routes to cut costs up to 3x
- The Download: how people really use AI, and Flock’s design choices
- Alibaba’s lightweight Qwen model takes on larger AI systems from OpenAI, DeepSeek, Zhipu
- Reach Capital raises $265M Fund V to back AI founders building to ‘expand human potential’
- Partnering with CodeAI to prepare the first AI generation
- Pacing model development in an era of cyber-critical capabilities
- Can AI Coexist With Privacy? Proton’s Andy Yen Says It Will Have To
- We still don’t know how people are really using AI
- Baidu’s quarterly revenue drops 4% as AI cloud surge fails to offset advertising slump
- Singapore’s universities race to embrace AI as job demands transform
- The Powerful Chinese AI Model Experts Warned About Is Here
- AI’s recursive self-improvement might not come so quickly after all
- Qwen3.8-27B runs frontier-class coding agents and reasoning locally, no cloud API required
- Multi-Vector (Late Interaction) Embedding Models with Sentence Transformers
- Cursor launches Origin code hosting platform as GitHub outage exposes opening in AI coding race
- This R-Rated Film Studio Wants to Be the HBO of AI
- Hidden Airtag reveals Amazon is trashing rare books to train AI
- NVIDIA Nemotron 3.5 Lightning now available in Amazon SageMaker JumpStart
- Don't Drop the BATON: Long-Horizon Robot Manipulation via Agentic Subtask Exploration and Transition-aware Memory
- What Do Compliance Detectors Read? An Audit of Activation Probes and Guard Models
- Model Hypnosis: Strong control of AI via additive subliminal effects
- CaliBench: Are the Stochastic Dynamics of Video World Models Physically Calibrated?
- When State Becomes an Attack Surface: State-Semantic Injection in LLM-Driven Embodied Agents
- Build OpenClaw agents that transact with Amazon Bedrock AgentCore payments
- PCA-guided Activation Scaling for Monotonic Bidirectional Control over LLM Sycophancy
- Every Coin Has Two Sides: On the Dual Nature of Generalization in On-Policy Distillation of Large Language Models
- Reconstruction: A Blind Benchmark for Recovering Research Ideas from Pre-Publication Bibliographies
- Toward Better Assessment of LLMs' Performance in Clinical Error Detection