AI agent development hasn’t accelerated as expected, Zuckerberg says

Meta has quietly released a new app called Pocket, an AI-powered platform that allows users to generate small interactive games and experiences using text prompts, alongside a scrollable discovery feed of content created by others. The app, first spotted by reverse engineer Alessandro Paluzzi and confirmed live on both the App Store and Google Play […]

With the cost of new RAM soaring, Meta has found a thrifty way to reuse older memory in newer servers. The performance of about 40% of Meta’s millions of servers is limited by a lack of memory, the company said — but it has a surplus of older DIMMs from decommissioned servers, because RAM chips can last about twice as long as the rest of the machine. To profit from this imbalance, it developed a custom Computer Express Link (CXL) chip it calls Vistara, and associated software, to decouple older memory from server memory channels, enabling its reuse in new machines alongside their native memory. Using the older RAM with the CXL interface doesn’t significantly affect performance — although it would have done if the older DIMMs were plugged straight into newer servers. Kudos to tech site The Register for noticing the development, which Meta described in a technical paper: Vistara: Making CXL Real — Full Path from ASIC Design and OS Support to Hyperscale Deployment,” setting out how the ne

Meta is set to launch a new Muse Spark model with stronger coding and agentic capabilities, as Chief AI Officer Alexandr Wang touted the update as a step toward closing the gap with rival AI platforms and expanding the company’s enterprise AI ambitions. “..Our next Muse Spark update is coming soon. Big improvements in coding and agentic capabilities to be more competitive with other leading models,” Wang wrote in a post on X in an attempt to clarify CEO Mark Zuckerberg’s comments about the slow progress in AI agent development made during a company townhall. In the same townhall, Wang said that the next Muse Spark update, codenamed Watermelon, which uses far more compute than its predecessor, has already caught up with OpenAI’s flagship GPT 5.5 model, according to a Business Insider report that cited anonymous sources. What the update means for enterprises The stronger coding and agentic capabilities in Watermelon, according to Pareekh Jain, principal analyst at Pareekh Consulting, cou

Meta is set to launch a new Muse Spark model with stronger coding and agentic capabilities, as Chief AI Officer Alexandr Wang touted the update as a step toward closing the gap with rival AI platforms and expanding the company’s enterprise AI ambitions. “..Our next Muse Spark update is coming soon. Big improvements in coding and agentic capabilities to be more competitive with other leading models,” Wang wrote in a post on X in an attempt to clarify CEO Mark Zuckerberg’s comments about the slow progress in AI agent development made during a company townhall. In the same townhall, Wang said that the next Muse Spark update, codenamed Watermelon, which uses far more compute than its predecessor, has already caught up with OpenAI’s flagship GPT 5.5 model, according to a Business Insider report that cited anonymous sources. What the update means for enterprises The stronger coding and agentic capabilities in Watermelon, according to Pareekh Jain, principal analyst at Pareekh Consulting, cou

Meta's halted Samsung AI chip project highlights the competitive and dynamic nature of the AI chip market, with new players like Anthropic seizing opportunities. The post Meta’s reported Samsung AI chip deal hits a wall as Anthropic steps into the conversation appeared first on Crypto Briefing.

The Indian government has reportedly issued notices to Telegram and Signal over username features, warning they could enable impersonation and fraud, after taking similar action against WhatsApp. Regulatory Scrutiny Expands Over Username Features on Messaging Platforms The Indian government has reportedly issued notices to Telegram and Signal under the Information Technology Act, 2000, and the […]

Meta's increased compute investment in AI models could redefine industry standards, impacting infrastructure value distribution and AI token markets. The post Meta’s Watermelon AI model uses ten times more compute than Avocado, signaling a massive bet on proprietary AI appeared first on Crypto Briefing.

At an internal meeting, the Meta CEO reportedly said that AI development efforts were not moving as quickly as anticipated.