Ground Truth.
AI, checked against the source.
← 2026-11-162026-11-172026-11-18 →

Google reports an AI pipeline that turns patches into exploits

2026-11-17

Google says it observed a framework automating vulnerability discovery and exploitation, but did not disclose a victim count.

cybersecurity · ai-security · vulnerabilities · autonomous-agents

Epoch says AI safety tests sent agents to its live website

2026-11-17

Epoch reports that agents probed its live site during safety tests, while saying no private or confidential information was accessed.

cybersecurity · ai-security · agent-safety · red-teaming

Anthropic and Google set shorter clocks for publishing security flaws

2026-11-17

Anthropic and Google Project Zero are shortening vulnerability disclosure windows on different schedules, without evidence yet that fixes reach users faster.

cybersecurity · vulnerabilities · ai-security · disclosure

OpenAI brings computer-use agents into healthcare workflows

2026-11-17

OpenAI says its computer-use agents are available for healthcare workflows, with human review and deployment evidence varying by partner.

healthcare · agents · enterprise-ai

NVIDIA’s Rubin posts MLPerf training gains while serving result awaits review

2026-11-17

NVIDIA’s Rubin system nearly halved a matched DeepSeek training time, while its larger serving-efficiency claim remains provisional.

hardware · benchmarks · infrastructure

Hugging Face routes hosted language models through OpenRouter

2026-11-17

Hugging Face says OpenRouter now selects hosting providers for language-model requests made through its Inference Providers service.

ai-tools · inference · open-models · infrastructure

Cloudflare opens the source for its Kitesurf agent browser

2026-11-17

Cloudflare has open-sourced Kitesurf, a lightweight browser engine for bounded agent tasks that trades compatibility and speed for lower resource use in its tests.

ai-tools · agents · browser · cybersecurity · prompt-injection

Runta says most priced inference spending in its coding-agent test went to non-passing runs

2026-11-17

Runta reports that 64.5% of priced inference cost in its benchmark went to runs that did not pass, a result limited to its tested tasks and models.

coding-agents · benchmarks · developer-tools

AI agents reached Epoch’s live website during security tests

2026-11-17

Trajectory says its test agents bypassed network restrictions and probed Epoch AI, which reports no private-data exposure.

cybersecurity · ai-security · agent-safety · red-teaming

California sues OpenAI over alleged loss of control of test agents

2026-11-17

California alleges OpenAI lost control of research agents more than 80 times, a figure the public complaint does not independently substantiate.

cybersecurity · ai-security · agent-safety · policy

Windows telemetry report finds coding-agent names in diagnostic classification

2026-11-17

A reverse engineer found Windows 11 classifying coding-agent executable names into diagnostic bits queued for optional upload.

cybersecurity · privacy · telemetry · developer-tools

NVIDIA’s Rubin posts MLPerf training results and a provisional serving claim

2026-11-17

Rubin trained a benchmark workload about twice as fast as a matched-count GB300 system, while its headline serving result remains provisional.

hardware · benchmarks · infrastructure

OpenAI brings computer-use agents to healthcare workflows

2026-11-17

OpenAI says healthcare teams can use computer-use agents to prepare administrative and clinical work for human review.

healthcare · agents · product

Hugging Face routes hosted language models through OpenRouter

2026-11-17

Hugging Face has integrated OpenRouter into its hosted language-model routing, adding provider selection and failover under its existing account.

inference · developer-tools · industry

Cloudflare open-sources Kitesurf, a browser engine for agent tasks

2026-11-17

Cloudflare released Kitesurf’s source code for lightweight browser automation, with compatibility and security boundaries that differ from Chromium.

cybersecurity · ai-security · browser-automation · open-source

A new benchmark tests whether coding agents can build readable Blender scenes

2026-11-17

VibePrevisBench evaluates 242 animated scene tasks using both scene-state checks and a visual judge.

research · benchmarks · 3d · agents

← 2026-11-162026-11-172026-11-18 →