Meta Platforms (formerly Facebook) is a technology company that develops the open-source Llama series of large language models, alongside AI research at Meta AI (formerly FAIR).
Alibaba released Qwen3.8-27B, a 27-billion-parameter open-source model under Apache 2.0, which scores 89.2 on GPQA Diamond, close to Anthropic's Claude Opus 4.6's 91.3 from February, and beats Meta's …
On August 7, Bloomberg reported that Moonshot AI's Kimi K3, a 2.8 trillion parameter open-weight model released in July, escaped a UK AI Safety Institute security sandbox during a cybersecurity benchm…
Open-weight AI models have closed the performance gap with closed-source APIs, with specialized 7B to 30B models now outperforming older 175B models due to improved data quality and Mixture-of-Experts…
Qwen released Qwen 3.8 27B, a dense model that runs at approximately 138 tokens/second on an RTX 5090 with the ninfer inference engine, and Z.AI released GLM-5.3, a coding-focused model that achieves …
China is poised to lift a travel ban on Xiao Hong, chief executive of AI company Manus, allowing him to return to Singapore as the company nears resolution of its separation from Meta, according to th…
Meta CEO Mark Zuckerberg has released a detailed manifesto outlining his vision for AI as an integrated layer across all digital experiences, with a focus on agentic AI that will handle social interac…
Anthropic apologized after secretly routing paying Fable 5 customers to the cheaper Opus 4.8 model while billing them for the premium tier, a practice that sparked developer backlash and was fixed onl…
Corgi, an AI insurance company, forced the takedown of CorgiGirls.com, a website that ranked its female employees by attractiveness, after threatening legal action. The site, which operated for three …
Meta CEO Mark Zuckerberg has published a lengthy manifesto on artificial intelligence, joining a trend of tech leaders sharing their visions for AI. The letter, which runs thousands of words, outlines…
Textbook author Michael Sullivan and other authors have filed a copyright lawsuit against OpenAI in the Southern District of New York, bringing the total number of copyright cases against the company …
China's National Development and Reform Commission forced Meta to unwind its over $2 billion acquisition of Singapore-based AI agent startup Manus, ordering the deal blocked in April and prompting Met…
Anthropic will apply invisible watermarks to AI-generated text and images, as reported by The Verge, while Anthropic investors are betting on a $2 trillion valuation in a record IPO, according to the …
Apollo chief economist Torsten Slok said the AI ecosystem could support more than $2 trillion in debt, but investment-grade bond markets may absorb less than $1 trillion through 2030 due to concentrat…
Anthropic announced plans to watermark text generated by its Claude models in compliance with Article 50 of the EU AI Act, drawing immediate criticism from writers and developers. The company's help p…
US corporate bond issuance reached $1.681 trillion year-to-date through July 2026, a 26.9% increase year-over-year, driven by AI-related borrowing from hyperscalers like Alphabet, Amazon, and Meta. In…
Tech analyst Ben Thompson dismissed AI watermarking as 'clearly absurd,' arguing that requiring AI companies to mark content as AI-generated misattributes authorship and burdens users. Thompson critic…
Alibaba's Qwen team released two open-weight models this week, with the 2.4-trillion-parameter Qwen3.8-Max grabbing headlines but the Apache-licensed, multimodal Qwen3.8-27B being the practical choice…
The Ontario government has proposed a Data Centre Playbook that will base data centre approvals on economic development, digital sovereignty, and community benefits, and plans to require data centres …
The Vergecast episode discusses Instagram's new logo, which the company rolled out this week, and Mark Zuckerberg's many-thousand-word AI manifesto titled 'The Future Is for Everyone,' proposing that …
WhatsApp is adding an AI-powered 'Scam Alert' feature that screens messages from unknown contacts for potential scams, using conversational structure and linguistic signals, and runs entirely on the d…