The 2026-11-15 YZ Index Smoke quick test covered 13 models, with Claude Sonnet 4.6 and GPT-6 Luna tied at 95.19 points for the daily lead. Smoke is a daily 10-question check intended for short-term signals, not for replacing Full weekly leaderboard conclusions.
Some governments moved to confirm chip supply commitments, while others are monitoring the impact on local projects or addressing ongoing investigations.
From Apple's repeatedly delayed Siri AI to OpenAI's messy "super app" launch, here's a look at the AI projects that shut down or missed expectations.
The FBI has issued FLASH-20261113-001, warning that the open-source LLM-driven ARTEX attack framework has been used to breach organizations across nine sectors, with more than 1,600 publicly exposed consoles found worldwide and some intrusions reaching privileged access in under three hours. The framework lowers the barrier to end-to-end automated enterprise intrusion, with supply chains and third-party integrations particularly exposed.
Fully automated ransomware hasn’t arrived—yet. But seemingly unsophisticated attackers are using AI tools to wreak havoc.
OpenAI has launched the beta of Businesses in ChatGPT, adding sponsored placements, merchant profiles, product data feeds, and AI customer service agents to the ChatGPT interface. The move creates an advertising revenue track alongside subscriptions and brings ChatGPT into closer competition with Google.
A Wall Street Journal investigation alleges that NVIDIA systematically funded AI startups on condition that they use the proceeds to buy NVIDIA GPUs and build data centers, while Bloomberg later outlined commitments exceeding $750 billion. The report also describes CFO objections to circular financing and regulatory warnings about the structure.
Normally, aviation is slow to embrace new technology—but AI has come for the skies.
In an updated S-1/A, Anthropic says it has crossed a key threshold in its Responsible Scaling Policy: Claude's contribution to internal R&D is enough to double the pace of model progress. The disclosure places a self-imposed safety trigger within IPO risk factors, sharpening the tension between AI acceleration and commercial expansion.
The White House AI Working Group recommends that federal agencies grant companies at least 12 months of liability relief for good-faith voluntary disclosure of AI-driven cybersecurity vulnerabilities and incidents. It frames regulatory tolerance as a trade-off for greater private-sector reporting and collective transparency, while key questions remain over agency adoption, CISA reporting rules, and congressional reauthorization of the 2015 cybersecurity information-sharing law.
Airlines praised the AI tool, but the air traffic controllers’ union wasn’t involved.
Nvidia poured tens of billions into AI startups under Huang. Without him at the helm, founders and VCs are bracing for smaller checks and lower valuations.
Google has released Gemini 4 Hydrogen, a lower-priced model in the Gemini 4 family that tops Argon on the Artificial Analysis Intelligence Index and is positioned for coding and agentic execution tasks. The launch also introduces Orchestrator, which automatically routes work between Carbon and Hydrogen within Google Antigravity.
OpenAI has disbanded its standalone robotics team and folded robotics researchers into its core research organization, betting that general-purpose frontier models can directly control robots. The restructuring contrasts with Google DeepMind’s specialized robotics model strategy and leaves unresolved who will handle real-world engineering challenges.
Researcher traces a path from AI reward hacking to potential human extinction.
New research shows the most advanced AI models from China were trained on banned Nvidia chips.
The company says a bad update—not a cyberattack or AI—caused its servers around the world to crash.
LangChain now claims 7,000 customers of its model and platform agnostic agent management tool LangSmith.
In today's Smoke evaluation, Claude Sonnet 4.6's main leaderboard score dropped from 82.31 to 70.61, a decline of 11.7 points. The material constraint dimension accounted for the bulk of the drop, falling 19.9 points in a single session.
Claude Opus 4.7's main leaderboard score fell from 94.91 to 83.48 in today's Smoke evaluation, a drop of 11.4 points, while the material constraint dimension fell from 89.30 to 70.00. The decline is most likely due to a higher share of difficult constraint questions in a single day's random draw rather than an overall degradation in model capability.
