Urgent.News

What's breaking now, across thousands of outlets.

AI

I asked ChatGPT, Claude, and Gemini to do something I normally handle myself, and the results surprised me

What happens when three AI assistants plan the same road trip from scratch?

I asked ChatGPT, Claude, and Gemini to do something I normally handle myself, and the results surprised me

A writer decided to hand over the task of planning a trip to AI, finding it usually enjoyable to plan themselves. This experiment was part of a larger exploration into the capabilities of AI assistants.

In a separate instance, a director and senior counsel of policy and legal operations at Ford, Whitney Stefko Dover, created a bespoke AI assistant named Claudette using Claude. Claudette managed the family's emails and calendars, providing a daily briefing and drafting morning texts. However, Stefko Dover has since created a new AI assistant, Gigi, using ChatGPT Work, and Claudette is now on a performance improvement plan.

A writer also asked ChatGPT and Gemini for advice on what to learn after the age of 40. ChatGPT provided four pointers, including managing money for the long term, but the outlet Tom's Guide only reported ChatGPT's advice.

Brief written by urgent.news from XDA Developers, Business Insider, Tom's Guide — 3 reports on this story. Machine-written — may contain errors; check the original before relying on it.

This story

This is one outlet's version. Read the fullest account.

Read the original at xda-developers.com →

More in AI

A WAF That Reads the Prompt: OWASP CRS for LLM and MCP

Originally published at webofmike.com on 2026-09-20. The demo repo and every command in it were run before publishing. A conventional web application firewall reads a URL, some headers, and maybe a…

  • Conventional WAFs inspect URL, headers, form body; agent traffic content in request body.
  • OWASP Core Rule Set (CRS) applied to prompts and tool calls with SOC custom signatures.
  • Enable body inspection via two settings: buffer body and define core ruleset actions.

How I Debugged a KV-Cache Offloading Bug in vLLM

How I Debugged a KV-Cache Offloading Bug in vLLM LLM inference performance is often limited by GPU memory rather than raw compute.

  • KV-cache offloading bug in vLLM due to mixed KV-cache groups
  • Introduced blocksperchunk to maintain backward compatibility
  • Resolved hidden assumption in KV-cache offloading logic

More from Sunday 20 September →