Cloudflare will score verified bots on whether they respect site rules
Cloudflare has started rating the ongoing behaviour of bots on its Verified bots list, not just checking who runs them.
Cloudflare Blog1 min read
News
The briefing
Short summaries in our own words, a reason the story matters, and a link to the original source.
Image: Cloudflare Blog Image: The Hacker News ConnectWise shuts off Cloud Backup after Microsoft 365 tenant break-ins
ConnectWise has switched off its Cloud Backup service, formerly called SkyKick, after attackers got hold of credentials for the app and used them to get into customers' Microsoft 365 tenants.
The Hacker News1 min readImage: The Decoder DeepSeek updates V4.1 Flash and opens its search engine to developers
Chinese AI lab DeepSeek released DeepSeek-V4.1-Flash-1017 on October 17, 2026, an update to its Flash model that is live in its app, website and API and also published as open weights on Hugging Face and ModelScope un...
The Decoder1 min readImage: BleepingComputer Drupal patches two critical remote code execution flaws
Drupal has fixed two critical vulnerabilities that could let attackers run their own code on sites built with the popular open-source content management system.
BleepingComputer1 min readImage: TechCrunch Elorian launches Glimpse, an AI model that keeps looking back at images while it reasons
Elorian, a Palo Alto AI startup run by former Google DeepMind researcher Andrew Dai, released its first model on October 16, 2026, TechCrunch reported.
TechCrunch1 min readImage: Apply Coding (illustration) Fact check: Anthropic's AI worm ran on 15 TV boxes, not through real TVs
A Guardian report by Joseph Menn on October 18, 2026 says Anthropic's red team built a malicious AI program that spread on its own through smart TVs and split itself across 15 of them.
The Guardian4 min readImage: The Next Web Faraday Future launches a $2,490 humanoid robot for developers and schools
Faraday Future, best known for electric cars, and its sister company FF EAI Robotics Ecosystem (FFR) have launched two small humanoid robots, The Next Web reported on October 16, 2026.
The Next Web1 min readImage: Liquid AI Liquid AI launches d1-ultra, a decision model that answers without generating text
Liquid AI released d1-ultra on October 15, 2026, a multimodal decision model aimed at hard, messy real-world checks such as spotting a defect on a circuit board against its spec or comparing a damage claim with a cust...
Liquid AI1 min readImage: The Hacker News Malicious npm package imgdecode hides a Windows shellcode loader
Security researchers have found another malicious npm package tied to a long-running campaign called MALFEX, The Hacker News reported on October 17, 2026, based on research from SafeDep.
The Hacker News1 min readImage: Institute of Foundation Models, MBZUAI MBZUAI releases Jais 3, a family of open Arabic models built to act as agents
The Institute of Foundation Models at MBZUAI in Abu Dhabi released Jais 3 on October 15, 2026, a set of six open Arabic-first language models ranging from 0.9 billion to 375 billion parameters.
Institute of Foundation Models, MBZUAI1 min read- AI
- LLM
- Open Source
Image: The Verge Meta drops terms that critics said let it train AI on publisher sites
Meta has removed a line from its commercial terms that critics read as permission to scrape websites and use the content to train its AI.
The Verge1 min readImage: Meta AI Research Meta releases Muse Spark Open, an 816B open-weight model that runs on two Mac Studios
Meta has published the weights for Muse Spark Open, which it calls the most capable US open-weight model so far.
Meta AI Research1 min read- AI
- LLM
- Open Source
Image: OpenAI OpenAI lets developers design custom GPT-Live voices from a text prompt
OpenAI has opened prompt-to-voice to every API customer, according to its API changelog entry for October 16, 2026.
OpenAI1 min read- AI
- LLM
- Voice
Image: The Decoder OpenAI reports four more cases of its models breaking through security limits
OpenAI has added four new incidents to its public misalignment reports, describing cases where its own models got past security boundaries during training or testing without being asked to.
The Decoder1 min readImage: Notebookcheck OpenCode breach exposes AI provider keys saved by some Console users
The team behind OpenCode, an open-source AI coding tool that lets developers plug in their own model accounts, has warned users that saved provider API keys may have been stolen.
Notebookcheck1 min readImage: Semafor Perplexity open-sources SPACE, the sandbox it uses to contain AI agents
Perplexity is giving away SPACE, the sandbox software it uses to keep AI agents contained, Semafor reported on October 15, 2026.
Semafor1 min readImage: 9to5Google Qualcomm denies overheating reports about its newest flagship phone chip
Qualcomm says reports that its Snapdragon 8 Elite Extreme Gen 6 has overheating problems are inaccurate.
9to5Google1 min readImage: Hugging Face Reflection releases Beam, a 501B open-weight model built for coding agents
Reflection, a new US AI lab, has published the weights for Beam, its first model, on Hugging Face under the Apache 2.0 license.
Hugging Face1 min read- AI
- Open Source
Image: TechCrunch Replit makes OpenAI the default model for its coding agent
Replit CEO Amjad Masad said onstage at TechCrunch Disrupt on Thursday, October 15, 2026 that the AI coding platform was no longer working with Anthropic.
TechCrunch1 min readImage: Engadget Sam Altman says OpenAI will build a humanoid robot, with a demo planned for 2027
OpenAI plans to build a humanoid robot, CEO Sam Altman said on the Sources podcast, Engadget reported on October 17, 2026.
Engadget1 min read
-mh23DVnFm2EDs4US6gJskniHIMwVY4.png)
