Technology

Walking faster, hanging out less

Source

MIT Technology Review

Published

Oct 21, 2025

TL;DR

AI Generated

A study by MIT scholars using computer vision technology reveals changes in pedestrian behavior in three northeastern US cities from 1980 to 2010. The average walking speed increased by 15%, while the number of people lingering in public spaces decreased by 14%. Factors like cell phones and indoor meeting spots like coffee shops may contribute to this shift. The findings could aid urban designers in creating or modifying public spaces to foster more encounters and community engagement.

Read Full Article

Autonomous Driving: Assessment Of YOLO Algorithms (RMIT et al.)

A technical paper titled “Advances in You Only Look Once (YOLO) algorithms for lane and object detection in autonomous vehicles” was published by RMIT University, Kyungpook National University, Deakin University, and the RCA Robotics Laboratory. The paper discusses the importance of accurate perception in Autonomous Vehicles (AVs) for lane detection and object recognition. It highlights the benefits and limitations of YOLO algorithms in real-world environments, emphasizing the need for improvements in detecting lane markings, adapting to various driving conditions, and handling environmental factors. The paper provides insights for researchers, engineers, and policymakers to enhance AV navigation systems for increased safety and reliability.

SemiEngineering•

3 months ago

Motorized aim assist system uses a moving mousepad for laser-accurate headshots — a cannibalized GRBL drawing system and computer vision software deliver perfect shots

Nick from the Basically Homeless YouTube channel created a motorized aim assist system using a moving mousepad that leverages a GRBL drawing system and computer vision software for precise headshots in video games. By combining woodworking, hardware hacking, and software skills, Nick developed a system that tracks targets on the screen and adjusts the mousepad underneath the mouse to achieve accurate shots. The mechanism, repurposed from an Xdraw A4 drawing board, smoothly moves the mousepad to enhance aiming skills. The computer vision model monitors the screen, calculates necessary movements, and sends commands to the mousepad, resulting in laser-accurate headshots.

Tom's Hardware•

5 months ago

Why Vision AI Models Fail

This white paper delves into the reasons behind the failure of vision AI models, highlighting common failure modes such as poor data quality, underrepresented edge cases, and model bias. Real-world case studies from companies like Tesla, Walmart, and TSMC illustrate how these failures can lead to significant business losses. The document emphasizes the importance of data curation, model evaluation, and analysis in preventing these failures before they impact production. Strategies for detecting, analyzing, and preventing model failures, as well as approaches for production monitoring to track data drift and model confidence over time, are also discussed.

IEEE Spectrum•

5 months ago

Faster, Smaller AI Model Found for Image Geolocation

A new, lightweight AI model has been developed for image geolocation, matching street-level photos to aerial images more efficiently. The software, created by Jingyu Liu, Kaifei He, and others, uses computer vision, hashing algorithms, and image analysis to enhance navigation systems and location software. This advancement in artificial intelligence aims to improve image analysis processes.