Urgent.News

What's breaking now, across thousands of outlets.

AI

Google let me remove Gemini's watermark because the real one is baked into the pixels

Visible AI watermarks were security theater. Learn why C2PA fails, how SynthID survives Photoshop, and how to check images using Gemini

Google let me remove Gemini's watermark because the real one is baked into the pixels

Google is releasing an updated Gemini for Home voice assistant, aimed at fixing voice command issues and adding new features. However, users are expressing frustration with the transition, as the latest update only addresses some of their long-standing problems. The Gemini for Home voice assistant has been in early access, with Google announcing three improvements on August 20th: better recognition of Spotify requests, improved message broadcasting, and more precise lighting control.

Despite these enhancements, users are criticizing Google for not tackling the main issues they've been facing. Reddit users report ongoing difficulties with basic commands, like setting alarms or scheduling thermostat automations, even after resetting their devices. The slow pace of the updates is raising concerns about the future of Google's smart home system, as users are growing increasingly skeptical.

Written by urgent.news from TechRadar's reporting — not their text. 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 androidpolice.com →

More in AI

I built K-Humanizer, an open-source Agent Skill for natural Korean writing

Korean text from LLMs is often grammatically correct but still sounds translated. The word order follows English, abstract nouns hide the action, and the tone is often more formal than the situation…

  • K-Humanizer refines Korean LLM output for native readers
  • Maintains facts, numbers, names, uncertainty, domain terms
  • MIT licensed GitHub repository with synthetic examples

AI chat moderation in Roblox — by meaning, not a word list

Word-list chat filters break the moment someone types a s s a s s i n with spaces, or a slur spelled with a 0 , or a perfectly innocent word that happens to contain a banned substring.

  • Roblox replaces word-list filters with AI chat moderation system
  • AI model analyzes message meaning, catches obfuscated toxicity
  • Developers use Cortex component, specify game context for moderation

Generate endless Roblox quests with AI (and cache them so it's cheap)

Hand-written quests run out. Players finish your content faster than you can write it, and the tenth "collect 10 wolf pelts" quest feels like filler.

  • AI generates dynamic quests for Roblox players
  • Cortex library integrates AI into Roblox game
  • Cached content eliminates repetition and reduces costs

AI Agent Standards Experiment: Test Rules Before Teams Trust Them

AI agents can look reliable after one impressive demo and still fail the moment real users, messy repositories, and conflicting instructions enter the room.

  • AI agents may appear reliable but can fail in real-world scenarios.
  • Teams often change agent rules based on gut feelings, not evidence.
  • The guide provides a lightweight experiment system for AI agent standards.

More from Sunday 23 August →