|
Meta launches Muse Spark Open: an open-weight AI model optimized for Macs
Meta announced Muse Spark Open, claiming it is the 'most capable US open-weight model to date.' The model was developed in collaboration with Apple and runs locally on Apple silicon, including two M5 Ultra Mac Studios using the MLX framework. Meta describes it as the first model in its class capable of running on desktop hardware.
|
|
Elorian's Glimpse model claims to outperform GPT-6 Astra on visual reasoning tasks
Elorian has released its first AI model, Glimpse, which is built to reason with images rather than merely about them. According to the company's own evaluations, Glimpse outperforms GPT-6 Astra on scientific chart reasoning and short-form video captioning. The model also posts competitive results across other visual reasoning benchmarks.
|
|
Jev becomes Vercel's fastest-adopted AI model ever
Vercel's new AI model, Jev, has achieved the fastest adoption rate of any model the company has launched, according to data from its AI gateway. Vercel's AI gateway serves as a unified API service similar to OpenRouter, giving developers streamlined access to multiple models. The rapid uptake signals strong developer demand for Vercel-native AI tooling.
|
|
Anthropic gives METR researchers employee-level access for independent safety assessments
Anthropic is granting METR, an independent AI safety research organization, unprecedented employee-level access to examine its models and internal practices. METR will publish its own assessments of Anthropic's risks and incidents, and Anthropic has waived its right to block those publications. The move is seen as a significant step toward third-party AI accountability.
|
|
Meta's new business terms may allow it to scrape news publishers' websites for AI training
Meta's updated business terms, taking effect October 30th, appear to grant the company permission to crawl and scrape websites 'associated' with businesses using its platforms — potentially including news publishers. The terms state Meta can use this outside content to 'develop, improve, research,' and 'train' its systems. The implications were flagged in a report by Nieman Lab.
|
|
Amazon's Panos Panay sees AI devices replacing the need for apps
Amazon devices chief Panos Panay spoke at TechCrunch Disrupt about a future where ambient AI devices work silently in the background, anticipating users' needs without requiring explicit app interactions. He envisions AI hardware that proactively serves users rather than waiting for commands. The comments underscore Amazon's strategic push to embed AI deeply into its device ecosystem.
|
|
Trump posts AI-edited photo to tout White House renovation work
President Donald Trump shared photos on Truth Social showcasing White House renovations, but one image comparing an old column to a 'restored' one appears to have been altered with AI. The photo contains several visual discrepancies and was flagged by Google's AI detection tool as 'made or edited' by AI. The incident adds to ongoing scrutiny over the use of AI-generated imagery in political communications.
|
|
Sega pulls Total War: Shogun 2 remaster after users complain about AI-assisted upscaling
Sega has removed Total War: Shogun 2 - Complete Edition from storefronts shortly after a troubled launch that drew user complaints about bugs and visual upscaling seemingly done by AI. The remaster of the 2011 strategy game bundled the main title and all its DLC but quickly ran into backlash over quality concerns. It is unclear when or if the game will return to sale.
|
|
Mark Wahlberg open to an AI clone carrying on his legacy
Actor and angel investor Mark Wahlberg said at TechCrunch Disrupt that he supports the use of AI in filmmaking, as long as it is used 'to enhance what you're doing.' He also expressed openness to the idea of an AI clone continuing his on-screen presence after his career ends. Wahlberg's comments reflect a growing conversation in Hollywood about talent rights and AI's role in entertainment.
|
|
Hollywood takes a harder look at Silicon Valley in a wave of new films
A notable cluster of 2026 films is critically examining the influence of tech founders and Silicon Valley on society. Aaron Sorkin's The Social Reckoning portrays Mark Zuckerberg as a destabilizing force, while other features take similarly skeptical stances. The trend suggests a cultural shift in how popular media frames the tech industry.
|
|
Bluesky's Attie AI tool expands to let users build games, apps, and websites
Bluesky's Attie AI tool, originally launched to help users build custom feeds, has expanded to allow anyone to create interactive social experiences like games, polls, quizzes, and dashboards that can be shared as standalone websites. The announcement was made by Bluesky CEO Toni Schneider at TechCrunch Disrupt. The platform is positioning itself around 'social vibe coding,' tapping into Bluesky's growing audience.
|
|
ChatGPT Chrome extension can now view your open browser tabs
OpenAI has upgraded its ChatGPT Google Chrome extension to allow the AI to view browser tabs that users already have open. Users can tag Chrome in ChatGPT's web or desktop app to activate the feature, enabling more contextual assistance. The update comes after OpenAI shut down its separate ChatGPT Atlas browser.
|
|
Nvidia RTX Spark N1X: a week of testing the new Arm-based laptop superchip
Nvidia's RTX Spark N1X processor has arrived in a range of laptops from Asus, HP, and Lenovo, and The Verge has been collectively testing them for a week. The Arm-based SoC is designed to rival offerings from Apple, Intel, AMD, and Qualcomm with high core counts and strong AI capabilities. Early impressions highlight it as a significant contender in the premium laptop market.
|
|
Cal AI's viral success was driven by a small army of influencers — and led to a MyFitnessPal acquisition
Cal AI, an AI-powered calorie-tracking app, grew rapidly by leveraging an extensive network of influencer marketing partnerships. The company generated over $30 million in revenue within two years and was subsequently acquired by MyFitnessPal for an undisclosed sum in March. The story highlights influencer-led distribution as a viable growth strategy for consumer AI products.
|
|
Veritas expands anti-counterfeit chip technology to jewelry, trophies, and fragrance bottles
Startup Veritas is extending its authentication chip technology beyond its initial use cases to include luxury goods like jewelry, trophies, and fragrance bottles. The chips are designed to combat increasingly sophisticated counterfeit products, known as 'superfakes,' that closely mimic authentic brand markings. The expansion signals growing demand for hardware-level product verification in the luxury sector.
|
|
Replit CEO envisions 'self-driving companies' powered by AI agents
Replit CEO Amjad Masad outlined his vision for the future of work at TechCrunch Disrupt, describing 'self-driving companies' where AI agents autonomously handle day-to-day operational tasks. In this model, humans would focus on higher-level decisions while AI manages execution. The concept reflects a broader industry conversation about AI's growing role in business operations.
|