Urgent.News

What's breaking now, across thousands of outlets.

Tech

Ghost in the Machine

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 4: Ghost in the Machine What I Built TouchGrass AI is a purpose-built micro-expedition application designed around one strict rule: The screen should be the shortest part of the experience. Developers, students, and desk workers can spend most of their day in front of screens, and even short breaks can turn into more screen…

<h2>Ghost in the Machine: TouchGrass AI</h2>

TouchGrass AI is an open-source micro-expedition application aimed at reducing screen time by turning short breaks into outdoor adventures. The app's core principle is to keep the screen the shortest part of the experience. Users select a time budget (5, 15, 30, or 60 minutes) and an outdoor environment such as a park, backyard, urban area, trail, or waterfront.

Upon selecting these parameters, the AI generates a sensory outdoor micro-quest with physical activities like observing nature, listening to birds, feeling tree bark, or exploring surroundings. Pocket Mode then dims the screen and urges the user to put their phone away and step outside. After completing the activity, users capture one field photo, which is classified locally using a MobileNetV4 ONNX model to distinguish nature/outdoor findings from indoor technology.

The application records activities and outdoor minutes in a private local SQLite journal, maintaining user privacy. The AI runs locally on the CPU, ensuring fast inference times around 11–16 milliseconds without the need for a GPU or CUDA environment. By eliminating cloud AI API calls and telemetry, TouchGrass AI preserves user privacy and offers offline capability, making it suitable for use in places with poor cellular connectivity.

The project emphasizes open innovation, utilizing lightweight open-weight ONNX models and local inference to stay independent of proprietary AI providers.

Written by urgent.news from Dev.to'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 dev.to →

More in Tech

Rutba Vision: video surveillance for camera networks past the size one recorder copes with

First published on rutba.io . A handful of cameras is a product you buy. A hundred and fifty, writing terabytes a day, is an operations problem, and it fails the same ways each time.

  • Rutba Vision handles large camera networks without single recorder limitations
  • Software runs node agents on all machines for edge, analysis, and core roles
  • Footage recorded in self-contained fMP4 segments with SHA-256 hashing

Count transcript tags in a podcast RSS feed

Five of eight AI podcasts in a June 11, 2026 feed audit declared zero transcripts in RSS. Across the 2,148 items those feeds served, 190 carried at least one podcast:transcript tag (8.8%).

  • Five out of eight AI podcasts had no transcripts in their RSS feeds
  • 190 items out of 2,148 contained a podcast:transcript tag
  • Chain of Thought had the highest transcript coverage at 98.4%

The Measurement I Kept Skipping

For months my test notes had a gap I could not see: I always measured the device at the moment it looked most interesting.

  • Fixed mark used 12 minutes after each scene ends
  • Notes became less dramatic and more reliable
  • Battery drain explained with consistent measurements

More from Saturday 31 October →