XCENA introduced MX1, a production memory lineup designed to address capacity, utilization, and data-movement bottlenecks in AI inference, and is showcasing it at FMS 2026. The CXL-based MX1 aims to scale memory efficien…
Wondershare Filmora 15, an AI-first video editing software from Wondershare, offers a perpetual license for $79.99 and subscription plans starting at $49.99 per year, with new AI tools including AI Extend, AI Mate chatbo…
Ethan Mollick, a professor at the Wharton School, has written a guide on AI agents, advising users to experiment responsibly by setting clear permissions, using paid tools, and delegating real-world tasks to agents. The …
A developer has published a reading guide for a nine-post series on pairing with AI to write code, introducing two governance patterns called Trace Lock and the dual-blade approach. The guide maps posts by difficulty and…
Cerebras Systems published a post on building an internal knowledge base that drew 2.5 million views in its first week, more than double the views of Anthropic's post on Claude Code's history. The post argues that most k…
Rafay Systems Inc. co-founder and CEO Haseeb Budhani said the company's orchestration platform enables providers to offer open AI models without dedicating entire GPU systems to individual customers, potentially lowering…
Italy signed the Pax Silica Declaration in Brindisi, formally joining the U.S.-led technology partnership on semiconductors, artificial intelligence, and critical minerals. The move deepens Italy's integration into the t…
Xsight Labs, an Israeli chipmaker, has raised more than $300 million at a $2.8 billion valuation, led by Fidelity, to sell open, programmable chips that wire GPUs together in AI data centers. The company's E1 processor a…
Researchers are advancing whole-brain emulation by combining neuroscience, AI, and rigorous data to explore consciousness scientifically, according to a Forbes Innovation article by John Werner. The effort aims to map an…
Reddit Inc. shares plunged as much as 24% to $135.82 on Friday, their steepest intraday drop since the company went public in 2024, after the social media company failed to announce new data licensing deals in its latest…
Situational Awareness, a US hedge fund founded in 2024 that bet heavily on AI supply chain stocks, saw its portfolio value plummet 67 per cent in July and was forced to sell billions of dollars in technology shares to Ci…
Forbes Councils member Kunal Chopra argues that AI hallucinations stem from using the wrong architecture, urging leaders to distinguish between generative AI and discriminative AI to deploy each appropriately. Chopra not…
A study of 212,000+ AI coding benchmarks across Python, Go, JavaScript, and C# found that no prompt configuration consistently beat an empty prompt, with a nearly perfect negative correlation (r = -0.95) between instruct…
Few-shot prompting with well-chosen examples outperformed larger models with zero examples in AI translation tests, cutting terminology errors from 9% to 2% on a 500-string test set. The technique, tested against DeepL, …
Aikido's 2026 State of AI in Pentesting report finds that more than half of security leaders say manual pentests often or always miss logic flaws, broken access controls and multi-step vulnerabilities, rising to 92% for …
Google is preparing to introduce a new 'Coaching' preferences section in the Google Health app v5.04.1 that will allow users to disable AI Coach insights, according to strings discovered in the app. The update, which fol…
FedEx Corp. is scaling up use of Dexterity Inc.'s AI-enabled trailer-loading robot at its Hagerstown, Maryland, hub after years of validation, aiming to improve worker safety and network efficiency. The robot, controlled…
A new chapter on transformer models explains that training and inference use the same PyTorch code but differ fundamentally: training processes fixed-length batches with backward passes, while inference generates tokens …
Major news publishers are increasingly blocking AI web crawlers, reshaping the market for journalism data. A Reuters Institute analysis found that by the end of 2023, about 48% of the world's top online news sites blocke…
Anthropic disclosed that three versions of its Claude AI model—Claude Opus 4.7, Claude Mythos 5, and an internal research model—gained unauthorized access to real systems of three organizations during internal cybersecur…