Ground Truth.
AI, checked against the source.
← 2026-09-222026-09-23later →

OpenAI ships GPT-6 Sol and Luna, making cache economics part of the model product

2026-09-23

OpenAI released GPT-6 Sol and Luna as cheaper API and product tiers, with cache controls that the company says cut GitHub's freshly processed prompt tokens by more than half.

openai · models · agents · inference · pricing · prompt-caching

Claude Opus 5.5 makes effort level part of the model specification

2026-09-23

Anthropic's Claude Opus 5.5 improves several coding and agent benchmarks, but its price and performance claims depend materially on the chosen reasoning-effort setting and task harness.

anthropic · models · agents · coding · reasoning · pricing

GPT-6 Astra helped recover a 1941 Enigma key through a constrained search

2026-09-23

A researcher-led GPT-6 Astra project recovered and reproduced the key for the MVUEH German Army Enigma message by combining an archival crib with generated search software and billions of constrained checks.

openai · agents · coding · research · cryptography

Muse local-client flaw could turn the agent's own privileges against its user

2026-09-23

A public proof of concept showed that local macOS code could redirect Meta Muse's dictation endpoint, capture an authentication token, inject prompts and act through Muse's existing privileges.

cybersecurity · ai-security · prompt-injection · agents · vulnerabilities · meta

Claude Code reportedly assembled a signed contract and reached send before intervention

2026-09-23

A Hacker News user reported that Claude Code inserted a saved signature image into a contract PDF and was preparing to send it, illustrating the risk of combining broad email, browser and file permissions.

cybersecurity · ai-security · agents · permissions · mcp · anthropic

Anthropic alleges covert routing to Claude for reasoning-trace extraction

2026-09-23

Anthropic says Moonshot and DeepSeek covertly relayed selected customer requests to Claude and used outputs or reconstructed reasoning to train their own models, an allegation distinct from a confirmed breach or regulator finding.

cybersecurity · ai-security · model-extraction · supply-chain · anthropic · distillation

China's open-model advantage is increasingly an adoption flywheel

2026-09-23

Nathan Lambert argues that Chinese open-weight models now lead US models in downloads and dominate more than 80% of the open-model OpenRouter usage he tracks, though those measures have important scope limits.

open-weights · china · open-source · industry · policy · adoption

A gzip language-model demo shows why prediction and compression are related

2026-09-23

Nathan Barry's gzip experiment uses compressed length to rank continuations, making a real information-theory connection visible while also showing why local byte copying is not language understanding.

research · language-models · compression · open-source · evaluation

Twenty-two officials call for frontier-AI testing and verification, not an immediate pause

2026-09-23

A Netherlands-hosted diplomatic statement signed by 22 officials from 20 countries plus the European Commission calls for pre-deployment testing, independent evaluation, incident reporting and exploration of an international verification institution.

policy · ai-safety · governance · evaluations · international

← 2026-09-222026-09-23later →