| ZOTPAPER |
| |
| News without the noise |
| Tuesday 3 November 2026 · The Daily Edition |
| |
 Security · Updated · 2 sources · 3 min
Google says it and industry partners have dismantled the XCPE botnet, which the company says infected more than 1.3 million Android TV streaming boxes.3 million Android TV streaming boxes, and has ended its relationship with Videostrong and revoked its licence to manufacture Android TV devices with preloaded Google apps and services. In a report published Monday, Google's Threat Intelligence Group said it attributes the botnet with high confidence to Chinese device maker Videostrong Technology Co, which it accuses of deliberately installing a persistent backdoor and selling access to criminal affiliates. |
| |
Security · Developing · 27 sources · 4 min
A suspected cyber-attack on supermarket refrigeration systems has affected more than 115 stores, according to a count by the security news site BleepingComputer, which listed incidents in Belgium, the Netherlands, Sweden, Luxembourg, Ireland and Estonia. The BBC reported that a smaller number of shops were also hit in Finland, a seventh country, where refrigerator malfunctions were detected in some K-stores and owner Kesko said its security experts were investigating the cause. | | | AI · 3 min
A quarter-century after Tim Berners-Lee envisioned an internet navigated by autonomous software agents, that vision is finally being realised, though not in the way he imagined. The 'agentic web,' in which large language models use the internet and carry out tasks for people, is emerging, driven by tools such as Meta's Muse, OpenAI's Dots, and Anthropic's Claude Duet. It promises a more intuitive online experience but could disrupt the web's economic foundations and raises questions about control, safety, and who ultimately holds power online. | | | Programming · 3 min
Google, working with Meta, has open-sourced TorchTPU, a native PyTorch backend for Google's tensor processing units that lets existing PyTorch workloads run on TPU hardware with minor code changes, InfoQ reports. |
| | Editor's picks |
Today's segment · Paper of the Day
Membership inference rarely succeeds despite measurable training effects in language models
|
|
Preprint 2610.41732v1
Ali Banaei, Muhammad Shahab Shahzad et al. — KTH Royal Institute of Technology
The authors re-ran OLMo 2 1B's 50-billion-token mid-training stage, replacing selected training chunks with documents from 7814 carefully matched pairs. They found a statistically significant average likelihood boost for trained documents (85.4% of pairs correctly ordered), but membership inference attacks using this signal achieved only 3.8% true positive rate at 1% false positive rate. Reference-free attacks performed near chance, underscoring the difficulty of identifying specific training documents.
|
|
|
| |
Unsubscribe · Manage preferences · View in browser
Zotpaper on iPhone and Android
zot.news · Privacy · editor@zot.news
You're receiving this because you subscribed to the Zotpaper Daily at zot.news.
Zotpaper
|
|