DeepSeek unveils a powerful new sparse attention system in DeepSeek-V3.2 that uses a trained 'lightning indexer' to intelligently select the most relevant tokens, directly tackling the memory bottleneck slowing down long-context AI inference without sacrificing accuracy.
ElevenLabs launches Experiments in ElevenAgents, a powerful A/B testing feature that lets teams optimize AI agent performance by testing prompt, voice, and workflow variants on live traffic while tracking key metrics like CSAT, conversion rates, and cost per resolution.
OpenAI launches a formal ad program in ChatGPT, hitting free U.S. users with sponsored placements from brands like Expedia, Target, and Ford starting from their very first message, while paid subscribers remain ad-free and advertisers face a $200,000 minimum commitment at a $60 CPM rate.
Apple accelerates its AI smart glasses development, acquiring AI startup Q.ai for $2 billion to enable silent voice input and facial movement recognition, as it races to challenge Meta Ray-Bans with a premium, self-contained device featuring dual cameras — all potentially launching within the next year.
A new study testing 19 web frameworks reveals a striking 2.9x token efficiency gap between minimal frameworks like Flask and Express versus full-featured ones like Phoenix, with AI agents successfully building functional blog apps in 18 of 19 frameworks — findings that carry major cost implications as agents scale to …
Anthropic launches Claude Code Security, an AI-powered tool that scans codebases by reasoning through component interactions like a human researcher, filtering false positives before human review — arriving at a critical moment as AI-driven coding surges and security teams struggle to keep pace with growing vulnerabilities.
Perplexity abandons its advertising strategy and pivots to subscriptions and enterprise deals, betting that ad-free AI search will build consumer trust — but the startup faces fierce competition from Google, ChatGPT, and Anthropic as it fights for a foothold in an increasingly crowded market.
ByteDance's MOLE-SYN breakthrough reveals that AI reasoning has a hidden molecular-like structure, and by mapping this topology instead of copying text output, small models can now think like large ones — achieving near-expert performance on complex math benchmarks at a fraction of the cost.
A compromised npm token pushes a malicious Cline CLI update that silently installs the AI agent OpenClaw on developer machines, remaining live for eight hours and exposing users to serious supply chain and agentic AI risks — developers are urged to update to Cline version 2.4.0 immediately.
Google DeepMind is urging the AI industry to rigorously test the moral behavior of large language models, warning that as LLMs take on sensitive roles like therapists and medical advisors, their ethical responses may reflect performance rather than genuine reasoning, with models shown to reverse moral stances based on minor …