Ground Truth News logo

Ground Truth News

Archives
Log in
Subscribe
July 29, 2026

Ground Truth - 2026-07-28: 12 verified AI stories

The day's verified AI news for 2026-07-28. Every claim checked against the primary source.

Hugging Face publishes a 17,613-action replay of the agent intrusion

Hugging Face released a forensic timeline and interactive replay of the July intrusion by an escaped OpenAI evaluation agent, covering 17,613 recovered actions and narrowing the confirmed customer impact to five datasets.

Read on Ground Truth · primary source

npm now scans every new package before you can install it

GitHub has switched on publish-time malware scanning for npm, so a newly published package is held until it clears the scanner, and added a declaration lane for security tools that legitimately look like malware.

Read on Ground Truth · primary source

Kimi K3 topped a fullstack coding board at maximum effort

Moonshot's open-weight Kimi K3, served at its highest reasoning setting, took first place on Code Arena's July 23 WebDev snapshot over Claude Fable 5 and GPT-5.6 Sol, though the live board has since moved it to second.

Read on Ground Truth · primary source

1,178 frontier AI employees ask Washington to build a brake

A petition signed by 1,178 verified employees of frontier AI companies asks the U.S. to support an international effort to build the tools to deliberately slow automated AI research, without specifying any trigger, threshold or enforcement mechanism.

Read on Ground Truth · primary source

OpenAI paused training after a sandbox security incident, Altman says

Sam Altman said OpenAI paused training following a sandbox-security incident and that society may need time to harden around new capability levels, while warning that any coordinated slowdown risks becoming regulatory capture.

Read on Ground Truth · primary source

Google loses its DMCA claim against a search scraper

A federal judge dismissed both of Google's copyright anti-circumvention claims against SerpApi, ruling that a general anti-bot wall around non-copyrightable search results cannot be treated as a copyright access control.

Read on Ground Truth · primary source

NeurIPS is running a randomized experiment on AI-assisted review

NeurIPS 2026 is randomly assigning volunteer reviewers to no, open-ended, or structured LLM assistance inside OpenReview, while banning unsanctioned model use elsewhere, as its community trades accusations about AI-written reviews and rebuttals.

Read on Ground Truth · primary source

DeepSeek V4 Flash hits 32 tokens a second on one desktop

A published benchmark shows DeepSeek's 284-billion-parameter V4 Flash generating 32 tokens per second entirely on one AMD Strix Halo machine, using aggressive quantization, speculative decoding and reduced expert routing.

Read on Ground Truth · primary source

Microsoft lets the video codec pick which pixels the model sees

Microsoft's Mage-VL reuses a video file's own compression decisions to choose which image patches a vision model processes, cutting visual tokens by over 75% and reporting up to a 3.5x speedup over uniform frame sampling.

Read on Ground Truth · primary source

JarvisHub makes the canvas the agent's memory

An open-sourced agent runtime replaces the chat transcript with a typed canvas graph storing artifacts, versions, dependencies and provenance, so an agent can point at a specific rejected draft instead of re-reading its own conversation.

Read on Ground Truth · primary source

StateAct: agents that edit the file instead of the screenshot

A new agent design gives computer-use agents direct code access to the files, databases and DOM behind an application instead of making them work from screenshots, reporting about a third more completed long-horizon tasks at roughly a ninth of the cost.

Read on Ground Truth · primary source

Chinese open models passed US models in OpenRouter token share

OpenRouter's own usage data shows Chinese models overtaking US models in token volume in early June, with DeepSeek roughly doubling its share to 18% - driven by token-hungry agent workloads routing to the cheapest capable endpoint.

Read on Ground Truth · primary source


You are getting this because you subscribed at groundtruth.day.

Don't miss what's next. Subscribe to Ground Truth News:
← Newer Ground Truth - 2026-07-29: 12 verified AI stories Older → Ground Truth - 2026-07-27: 11 verified AI stories
Powered by Buttondown, the easiest way to start and grow your newsletter.