Ground Truth.
AI, checked against the source.
← 2026-10-302026-10-31later →

Malicious Claude Code copies in Cursor’s marketplace tried to plant an SSH backdoor

2026-10-31

Tampered Claude Code extensions distributed through Cursor’s marketplace attempted persistent remote access; upstream stores were not affected.

cybersecurity · ai-security · supply-chain · coding-agents · cursor · endpoint-security

Amazon’s AI build-out and Google’s search economics expose different return pressures

2026-10-31

Amazon disclosed $75.9 billion in quarterly cash capital spending while Google said AI Mode queries currently earn less than conventional search.

ai-economics · infrastructure · amazon · google · earnings

House Democrats introduce a resolution opposing special AI-industry bailouts

2026-10-31

Greg Casar and Maxine Waters introduced a nonbinding House resolution opposing special federal rescues for insolvent AI firms and their financial backers.

ai-policy · ai-economics · congress · infrastructure · accountability

Claude Ultra removes the five-hour cap for verified accounts, while weekly limits remain

2026-10-31

Anthropic’s $200 Ultra plan removes the five-hour usage cap for verified accounts but retains a weekly allowance of ten times Pro.

anthropic · claude · subscriptions · coding-agents · ai-economics

OpenAI plans to honor a shared website signal for AI training preferences

2026-10-31

OpenAI says it will honor a shared robots.txt training-preference signal in the coming months, while existing crawler-specific controls continue.

openai · training-data · web · standards · publishers

OpenAI announces tighter violence safeguards and a first community safety report

2026-10-31

OpenAI plans new violence-related limits, stronger teen protections, and a first Community Safety Report in early 2027.

openai · ai-safety · teen-safety · transparency · product-policy

OpenAI makes its fair-use case through supportive briefs in the New York Times lawsuit

2026-10-31

OpenAI’s October 30 fair-use post highlights supportive court briefs in the New York Times litigation; it is advocacy rather than a new judicial ruling.

openai · copyright · training-data · publishers · ai-policy

Nesformer compiles a Nintendo processor into Qwen weights instead of training it on games

2026-10-31

Drew Breunig’s nesformer executes a compiled Nintendo CPU inside Qwen model weights, with a separate host drawing frames and major speed and compatibility limits.

open-source · neural-computation · qwen · program-synthesis · agents · games

← 2026-10-302026-10-31later →