|
Apple trained its own AI model for China with help from Alibaba
Apple has reportedly trained a custom AI model for the Chinese market in partnership with Alibaba, a rare cross-border collaboration amid rising US-China tensions. The China-focused large language model was developed with Alibaba's support, according to Reuters sources. The partnership reflects Apple's effort to bring Apple Intelligence features to China while navigating regulatory requirements.
|
|
Meta's Glimmer: Does Zuckerberg really believe AI is 'for everyone'?
Meta released Glimmer, an open-weight AI model that anyone can download and run on their own hardware, alongside a letter from Mark Zuckerberg arguing AI should be 'for everyone' rather than controlled by a few labs. The release contrasts with Muse Spark, Meta's more powerful model locked behind its own APIs, raising questions about the sincerity of Meta's open AI stance. Analysts are scrutinizing whether the dual-track strategy aligns with Zuckerberg's stated philosophy.
|
|
Google now lets users remove visible watermarks from AI-generated content
Google is allowing users to toggle off visible 'sparkle' watermarks on images, videos, and music generated by Gemini and its AI video tool Flow. The new 'Media watermark' setting does not affect invisible metadata watermarks used to identify AI-generated files. The change gives creators more flexibility while preserving underlying provenance signals.
|
|
Kog is going deeper to squeeze more inference out of GPUs
French startup Kog is challenging the notion that GPUs are poorly suited for agentic AI workflows, developing technology to extract more inference performance from existing GPU hardware. The company argues that with the right software approach, GPUs can handle the demands of complex, multi-step agentic tasks more efficiently than commonly assumed. Kog's work could have significant implications for AI infrastructure costs and performance.
|
|
Mark Zuckerberg has an Instagzam
Instagram rolled out a redesigned wordmark this week that has left many observers puzzled, with the new logo barely resembling the word 'Instagram.' The change prompted discussion on The Vergecast about why Instagram felt the iconic existing wordmark looked outdated. The hosts admit they cannot determine a clear rationale behind the branding decision.
|
|
Google AI health coach to use Abbott glucose data
Abbott and Google have signed a multiyear agreement to integrate Abbott's Lingo continuous glucose monitor data into the Google Health app, giving the Gemini-powered AI health coach access to real-time glucose trends. Users will be able to view glucose data alongside other health metrics and receive personalised coaching insights. The deal expands Google's push into AI-driven personal health management.
|
|
Samsung health AI models analyse wearable biosignal data
Samsung Research America's Digital Health Team has unveiled two AI foundation models trained on biosignals from smartwatches, covering heart activity, sleep, and physical movement. The work was presented at the Health Forum during Galaxy Unpacked in July 2026 as part of Samsung's broader Connected Care vision. The models aim to enable more personalised and continuous health monitoring through wearables.
|
|
Hyperscalers might regret embracing natural gas if new forecast proves correct
A new forecast warns that natural gas prices could triple in parts of the US, potentially saddling hyperscalers with enormous energy bills for their AI data centers. The warning comes as major cloud providers have increasingly turned to natural gas to meet surging power demands from AI workloads. The forecast raises questions about the long-term energy strategy of companies building out large-scale AI infrastructure.
|