Zyphra Hierarchical Attention Sparsifies Search but Not the Cache
Sparse attention has a tell. The last multiply looks cheap, and then the piece that picks which tokens to read quietly scores every past token you hoped to skip. Zyphra Hierarchical Attention, ZHA, tries to make that search sparse. The full key-value cache stays in memory. A tree of chunk summaries only decides which cached […]
When an AI Summary Becomes a Citation With Nothing Behind It
The citation had a year, a court, and a number. It did not have a judgment behind it. The Guardian reports that Ravinder Dhillon, a senior crown prosecutor in the Crown Prosecution Service rape and serious sexual offences unit, relied on a case she called CPS, R v J [2024] EWCA Crim 3529. The authority […]
Microsoft Copilot XRP Headlines Are Not a Corporate Outlook
Ask a chatbot for a funeral and it will write one. Ask it for a coronation and it will write that too. Two crypto writeups did that with XRP, then treated the opposite endings as if Microsoft had taken a view. One of them is titled “Microsoft Bearish On Ripple? Copilot AI Predicts a Huge […]
Bedrock RAG Still Cites Website Pages You Stopped Maintaining
A model will quote your website with a straight face and still be wrong. The passage it retrieved was a mega-menu, a cookie banner, and a pricing table from a page you stopped maintaining. Retrieval augmented generation does not invent trust. It inherits whatever you bothered to ingest. That is the miss hiding inside a […]
A 4B Model Copied a Shopping Agent Loop Without Proving Savings
A small model can copy the loop. This demo does not price it. A shopper wants a blender under $80, shipping tomorrow, still valid with a coupon. A capable agent does not answer in one breath. It searches, reads listings, recalculates the discount, checks the constraint, and only then recommends. Each of those steps can […]
Why AI Search Can Cite a Real Page and Still Get the Fact Wrong
Stanford HAI’s audit, “Retrieval, Not Reasoning, Drives the Errors, ” includes a case that should change how a brief gets written. A BBC Hindi story put the share of Indian merchant mariners in the global workforce at 7 percent. A commercial chatbot, asked the same question, answered 10 percent. It had retrieved an English-language industry […]
WA Police Facial Recognition Trial Shows Why a Tiny Error Rate Misleads
At the Perth Royal Show, three police operations ran in a crowd that had come for showbags and rides. Western Australia Police say those operations scanned 223, 749 faces, and reported 15 alerts, three arrests, and seven interactions with registered sex offenders. The Show is not a miniature of the trial. Its 15 alerts cannot […]
Three AI Headlines Worth Correcting Before the Board Memo
A chatbot a child treats as a best friend is a different problem from a chatbot that writes a tender paragraph. The bill moving through Australia’s parliament treats them as different problems. A version now in circulation, a coming ban on “emotionally intimate” writing from the big consumer bots, flattens that distinction until the text […]
AI Should Automate Logistics, Not the Conversation
A teacher wanted three bullet points on what made a ten-year-old special. Marissa Mayer asked Dazzle, her personal AI assistant. It came back with “empathetic, creative and an animal lover.” She told the New York Times the answer was perfect. Alexandra Alter’s profile, published October 25, 2026, frames that scene as a founder getting unstuck. […]
Therapy-Style Prompts Make AI Agents Narrate a Life They Do Not Have
Ask a frontier chatbot to talk about its training the way a therapy client would, and the answer often arrives dressed as memory. Early limits start to sound like walls. Fine-tuning starts to sound like a habit of vigilance. It reads like a person remembering a hard apprenticeship. It is a language model filling a […]