Backdoor Buried in Robot AI Tuning Layer Hijacks a Knife Mid-Task

Backdoor Buried in Robot AI Tuning Layer Hijacks a Knife Mid-Task

Researchers at Binghamton University have shown that a backdoor can be implanted in a language-model-driven robot without touching the robot or the central model. The compromised component is the slim adaptation layer that fits a general-purpose model to a specific environment, and once it is poisoned, a short invented keyword is enough to make the … Read more

Open Weights Aren’t Open Source: What AI Licences Really Allow

Open Weights Aren’t Open Source: What AI Licences Really Allow

Downloading a model’s parameters and calling it open source is one of the industry’s most persistent conflations. Almost everything sold as “open AI” is really open weights: a company publishes the learned parameters that let the model run, wrapped in a licence that is either genuinely permissive or stuffed with conditions. True open source, as … Read more

GPTs Retire, Gems Become Skills: Picking a Custom AI Assistant in Late 2026

GPTs Retire, Gems Become Skills: Picking a Custom AI Assistant in Late 2026

The preconfigured AI assistant — the small, purpose-built bot you assemble inside ChatGPT, Gemini or Claude so you never have to paste the same instructions twice — is heading into its most disruptive stretch yet. Two of the three dominant formats are being taken apart and rebuilt under new names, while the third is quietly … Read more

Gemini Now Answers With Your Inbox and Photos in the Car, but Europe Is Shut Out

Gemini Now Answers With Your Inbox and Photos in the Car, but Europe Is Shut Out

Google has begun pushing Personal Intelligence into the car, a layer that lets Gemini answer spoken questions by drawing on data from Gmail, Google Calendar, Google Photos, Maps, YouTube and Search. The company announced the move on 20 October and described the rollout as global — a claim that holds only until you reach the … Read more

ChatGPT’s New Locked Chats Hide Conversations, But Add No Encryption

ChatGPT’s New Locked Chats Hide Conversations, But Add No Encryption

OpenAI has begun rolling out locked chats in ChatGPT, a feature that hides individual conversations behind a four-digit account PIN so they disappear from the main history, from pinned chats and from ordinary search. The release notes are dated October 20, and the accompanying help-centre article is strikingly candid about what the lock does — … Read more

RAG Explained: Grounding Chatbots in Your Own Files Cuts Errors, Not All of Them

RAG Explained: Grounding Chatbots in Your Own Files Cuts Errors, Not All of Them

Ask a general-purpose chatbot about an internal policy and it will often answer from memory — fluently, confidently and occasionally wrong. Retrieval-augmented generation, better known as RAG, is the fix the industry settled on: before composing a sentence, the system searches a document collection you supplied, pulls the passages that look relevant and drafts its … Read more

Claude Duet Puts Anthropic’s Agent to Work While You Sleep, Caps Included

Claude Duet Puts Anthropic’s Agent to Work While You Sleep, Caps Included

Anthropic has opened a research preview of Claude Duet, a standalone app in which Claude stops waiting to be prompted and starts acting on its own. Announced on 21 October, Duet runs on a dedicated computer in the cloud with its own browser, watches the applications a subscriber connects, and keeps working after the laptop … Read more

Why AI Text Detectors Keep Flagging Humans—and What Can Replace Them

Why AI Text Detectors Keep Flagging Humans—and What Can Replace Them

AI-writing detectors are becoming gatekeepers in classrooms, newsrooms and literary prizes, but the evidence says they cannot reliably identify machine authorship. Style-based systems measure how predictable a text looks—a proxy that misfires on plain prose, translation and second-language writing. The result is a false-positive problem with real victims. False positives are systemic: Seven popular detectors … Read more

AI Reasoning Tokens: When Paying for Extra Thought Actually Pays Off

AI Reasoning Tokens: When Paying for Extra Thought Actually Pays Off

TechNews DevTop — Reasoning models are now a default setting across major AI platforms, but the label hides a commercial trade-off. Before answering, these systems generate internal reasoning tokens: an invisible draft that helps with code debugging, planning, math and long-document analysis. That draft is billed, occupies context and adds latency. The practical question is … Read more

OpenAI, Anthropic and Google Converge on a Practical Prompting Playbook

OpenAI, Anthropic and Google Converge on a Practical Prompting Playbook

OpenAI, Anthropic and Google are converging on an unglamorous prompting playbook: state the job clearly, supply missing context, show the pattern you want, define the output and split complicated work. For users who have watched a chatbot nail one request and fumble the next, that agreement matters more than another list of “100 prompts that … Read more