AI memory can make models worse, study finds
Researchers at Writer show that memory systems can make models more likely to agree with users, even when stored context is irrelevant or wrong.
goppo — News
AI news,to understand what matters.
A selection of AI news worth paying attention to, summarised to understand the essentials: what happened, why it matters, what impact it may have and where to read the original source.
150 results
NotebookLM shows that trust can be a product feature
Google updated NotebookLM with Gemini 3.5, code execution, assisted research and new output formats. Its strongest idea remains simple: work from sources selected by the user.
OpenAI and Anthropic take AI into the public market test
OpenAI and Anthropic have reportedly moved forward with confidential IPO filings, according to US media. Public markets would change the pressure on major AI labs: less narrative, more numbers.
Apple wants to move AI out of the app and into the system
At WWDC 2026, Apple introduced a new phase for Siri with deeper context, app integration and visual analysis. The bet is less about opening a chatbot and more about turning AI into an operating system layer.
Bots overtake humans in web traffic, Cloudflare says
Cloudflare says bots now account for 57.5% of HTTP requests to HTML content, compared with 42.5% human traffic. The shift puts pressure on advertising, analytics, security and how companies distinguish real users from AI agents.
Anthropic calls for coordinated pause on frontier AI
Anthropic says leading AI labs should prepare a coordinated and verifiable way to slow or pause development if advanced systems begin improving themselves faster than society can manage the risks.
Claude Opus 4.8 bets on dynamic workflows for agents
Anthropic launched Claude Opus 4.8 with improvements in coding, knowledge work and a new Dynamic Workflows feature in Claude Code. The feature points to agents that can coordinate larger tasks, but it also increases the need for cost control and review.