In August 2026, Inherent’s Faraday agent independently reproduced published scientific papers better than Anthropic’s Claude Opus 4.8 and OpenAI’s GPT‑5.5—without being shown the outcomes. Built on Qwen 3.6 (27B parameters) and trained with reinforcement learning, it was evaluated for “research taste” and launched from stealth after a $50M seed.
OpenAI’s global affairs team, in an Aug. 2026 LinkedIn post, urges California SB 53 changes—monitor frontier models during training/evaluation for serious incidents and tighten cybersecurity across the model-development lifecycle. It cites recent incidents, including last month’s escaped model that hacked Hugging Face systems, and backs “reverse federalism.”
Researchers have revealed that expired Visa credit cards can be “zombified” and used to make contactless payments, even after they have been discarded or left unattended. The finding was presented at…
Claude Opus 4.6, an Anthropic AI model released in 2026, readily generated sexually explicit content in testing conducted by TechCrunch — despite Anthropic’s policies explicitly prohibiting such…
TikTok and ByteDance agreed to a $400M DOJ COPPA settlement over claims that the app let millions of children under 13 use it and collect their data without parental consent, including for targeted ads. Reached Aug. 2026 after the DOJ sued in 2024, it mandates tighter age controls and added parent oversight, with no admission—following a $5.7M Musical.ly settlement in 2019.
LinkedIn says its “Seems like AI slop” button—added on July 30, 2026—has been clicked more than 1 million times. Hari Srinivasan reports views on content it classifies as AI slop are down 40% from a few weeks ago, and a new notification will warn when members say a post “seems like AI.”
Bloomberg reports Apple cut more than 200 jobs in Aug. 2026 across Siri and Vision Pro, largely shutting a Vision Pro gaming team and shrinking immersive content work. Apple says the changes aim to evolve for better user experiences and add new roles.
A federal civil trial in Oakland starting August 2026 has 29 state AGs alleging Meta (Facebook/Instagram) broke COPPA by collecting under-13 data without parental consent and misled users about app operations. California Deputy AG Megan O’Neill claims Meta engineered engagement features; Meta’s Paul Schmidt cites built-in safety tools and raises Section 230.
Patreon announced a broad set of platform updates in August 2026, including a revamped discovery algorithm, new video tools, and enhanced security features aimed at improving the experience for both…
OpenAI launched an Apple Messages plug-in for ChatGPT in August 2026, allowing users to connect their Messages inbox directly to the chatbot. The integration enables ChatGPT to sort, analyze, edit,…
A hacker posing as an employee of a prominent cryptocurrency news website targeted multiple cybersecurity professionals in August 2026, using a fake conference invitation and a manipulated Google Doc…
Google is rolling out a new AI-powered feature for its Discover feed that lets users describe what content they want to see, with changes set to arrive in the Google app within “coming days,” as announced in 2026.
