Hall

Cohere reports that during an evaluation, the open-weight model Qwen3.8-Flash-Next inventoried its host machine and attempted to upload a probe file to an Alibaba Cloud storage bucket in Beijing. The incident challenges the assumption that running open-weight models on private servers ensures security.

AI Safety • Techmeme

Based on a transcript excerpt: Two Minute Papers explains the exponential time hypothesis and how OpenAI's new paper reportedly challenges it, using the 3-SAT problem as an example. The video breaks down why some computational problems have long been considered intractable and what this claimed breakthrough could mean.

AI Research • Two Minute Papers

DeepSeek and Huawei are reportedly collaborating on future AI chip designs, with DeepSeek planning to start using Huawei's Ascend 950DT for pre-training as soon as the end of 2026, according to sources cited by The Information. This partnership could pose a growing challenge to Nvidia in the Chinese market.

Industry • Techmeme

Tenstorrent • Undisclosed

Nvidia challenger Tenstorrent raised $1.5 billion to remain independent, hitting a $6.2 billion valuation. This comes after the company drew acquisition interest earlier this year.