Amazon’s Delivery Glasses

Amazon is developing wearable smart-glasses for its Delivery Service Partner (DSP) drivers to enable a truly hands-free delivery workflow. The glasses deploy AI-powered computer-vision and heads-up display functionality so that a driver can scan packages, follow walking turn-by-turn navigation and receive delivery instructions directly in their line of sight, thus eliminating frequent phone glances and helping maintain attention on the road or the delivery environment.

The company emphasises driver-input in the design: hundreds of Delivery Associates tested early versions of the glasses, and their feedback influenced ergonomics, display clarity, battery design and safety-features. Looking ahead, Amazon plans to evolve the platform with features like real-time defect-detection, ambient hazard alerts (e.g., pets, lighting) and seamless integration across the full delivery journey, from the station, to vehicle, to doorstep.

More information:

https://www.aboutamazon.com/news/transportation/smart-glasses-amazon-delivery-drivers

Man With Brain Implant Controls Another Person’s Hand

Researchers at the Feinstein Institutes for Medical Research and the Donald and Barbara Zucker School of Medicine at Hofstra/Northwell have demonstrated a cutting-edge BCI that allows a person with paralysis to control another person’s hand and even feel what she feels. In the experiment, a man with a spinal cord injury used implanted sensors in his motor cortex, along with AI decoding and flexible electrode patches, to send signals to another volunteer’s arm, enabling her to perform tasks such as pouring water. 

 

This inter-human neural bypass opens new possibilities for cooperative rehabilitation, where users with different degrees of mobility can work together: the paralysed man helped a woman with partial paralysis improve her hand strength, while experiencing control and touch again himself. The trial suggests that by closing both movement and sensation loops, the technology could restore more natural sensorimotor function and perhaps motivate the body to repair itself. However, the study is still limited to a small group and long-term outcomes remain to be seen.

More information:

https://singularityhub.com/2025/10/23/one-mind-two-bodies-man-with-brain-implant-controls-another-persons-hand-and-feels-what-she-feels/

Eyes-Off Driving Google AI

General Motors is targeting new software initiatives for its vehicles over the next 3 years, including an in-vehicle artificial intelligence assistant from Google and a driver-assistance system that can largely control the vehicle without human interaction or monitoring.

The conversational Google Gemini AI will begin launching in its vehicles next year, followed by the new driver-assistance system, which will allow drivers to be hands-free and take their eyes off the road under certain circumstances, in 2028.

More information:

https://www.cnbc.com/2025/10/22/gm-tech-google-ai.html

Gradient 3D Printed Designs

A new open-source software is shaking up the way engineers design and print multi-material objects. Created in the Matter Assembly Computation Lab, the tool enables engineers to design spatially varying multi-material objects with remarkable ease and precision.

The project reflects a growing interest in computational approaches that merge coding with design. The software allows users to combine complex mathematical functions and assign them as materials to different regions of a 3D object.

More information:

https://interestingengineering.com/innovation/openvcad-multi-material-3d-design

AI Actress

After an artificial intelligence-generated actress drew intense backlash from human actors, the character’s maker says it’s not meant to replace people. Tilly Norwood looks like a young woman with wavy brown hair and clear skin and is AI-generated.

Hollywood news reported that talent agents were looking to sign Tilly as an actress and that movie studios are quietly embracing AI-generated content. The Tilly Instagram account racked up hundreds of angry comments, including from some of Hollywood’s biggest names.

More information:

https://edition.cnn.com/2025/09/30/tech/hollywood-ai-actor-backlash

Mice's Motion Sensors Could Spy Users

A group of researchers from the University of California, Irvine, have developed a way to use the sensors in high-quality optical mice to capture subtle vibrations and convert them into audible data. The high polling rate and sensitivity of high-performance optical mice pick up acoustic vibrations from the surface where they sit.

By running the raw data through signal processing and machine learning techniques, the team could hear what the user was saying through their desk. Mouse sensors with a 20,000 DPI or higher are vulnerable to this attack. And with the best gaming mice becoming more affordable annually, even relatively affordable peripherals are at risk.

More information:

https://www.tomshardware.com/tech-industry/cyber-security/high-performance-mice-can-be-used-as-a-microphone-to-spy-on-users-thanks-to-ai-mic-e-mouse-technique-uses-mouse-sensors-to-convert-acoustic-vibrations-into-speech

Smart in-Asphalt Fabric

A fabric developed by scientists from Germany's Fraunhofer Institute for Wood Research and the SenAD2 project, monitors the condition of paved roads. The material consists of flax fibers interwoven with an electrically conductive sensor wire less than 1 mm thick. The fabric is produced on a double rapier loom in a width of 50 cm (19.7 in) at whatever length is needed. Side-by-side connected lengths of the material are laid down across the base layer of a road, as it's being constructed. The surface layer is then poured over top, covering them.

At the side of the road, an external measurement unit is connected to the sensor wire. As cracks proceed to form in the asphalt over the months and years that follow, the material expands, placing pressure upon the sensor wire. This creates a change in the electrical resistance of the wire, which is detected by the measurement unit. AI algorithms determine the current extent of the damage, and to provide an estimate of its likely progression over time. The technology is currently being tested on an industrial road in Germany.

More information:

https://newatlas.com/good-thinking/electronic-fabric-asphalt-roads/

Forest Digital Twin

Scanning forests with lasers can provide three-dimensional models of forests, creating a unique way to look at these ecosystems. Researchers at the University of Helsinki are pioneers in applying this technical solution to measuring ecosystems. This technique, referred to as Terrestrial laser scanning (TLS), opens new opportunities in measuring forest structure, monitoring disturbances and simulating ecosystems.

TLS provides us with a new way of looking at forests – not just from above, but from within. It helps us understand tree growth, forest responses to disturbances and the effects of forest structure on biodiversity and capacity for recovery. A deeper understanding helps researchers, decision-makers and communities to make increasingly justified decisions – and, ultimately, to better reconcile human activity with the environment.

More information:

https://www.helsinki.fi/en/news/climate-change/digital-twin-forests-help-new-technology

3D-printed tissue Mimics Real Organs

Surgeons and doctors often rely on artificial models to practice delicate procedures. Researchers at the University of Minnesota Twin Cities have developed a new 3D printing technique that creates lifelike human tissue structures. Their work could reshape surgical training by offering models that look, feel, and respond more like real human tissue.

The Minnesota team found a way to control the shape and size of microscopic patterns inside the printed material. Those patterns directly influence the strength and stretchiness of the tissues, giving them realistic mechanical properties. They also built a mathematical formula to predict how the tissues behave under stress.

More information:

https://interestingengineering.com/innovation/3d-printed-human-tissue-surgery-training

Replicating the Electrical System of the Heart

Researchers have succeeded in creating a digital twin that for the first time replicates the heart's electrical conduction system, which coordinates heartbeats. The breakthrough could pave the way for diagnosing and tailoring treatments for conditions such as arrhythmias and heart failure. Researchers developed a computer model that digitally reconstructs the Purkinje network using a standard electrocardiogram.

The goal with the heart's digital twin is for it to be used as a virtual patient, where treatments can be tested and optimized without any risk. This would allow for better planning of procedures, greater personalization and, therefore, more effective results. The team is now working on two challenges: increasing computing capacity so that each patient's digital twin can be generated almost in real time.

More information:

https://www.upi.com/Health_News/2025/09/25/chile-heart-digital-twin-chile-study/1501758817726/

CHI Greece 2025 Article II

Yesterday, a paper I co-authored with colleagues from CYENS was presented at CHI Greece 2025. The paper is entitled “A Longitudinal Evaluation of Heart Rate Efficiency for Amateur Runners”. It presents a web-based system that uses large language models (LLMs) to automatically generate structured short-form video (i.e., reels) from lecture long-form videos while preserving instructor-authored material. It first presents Fitplotter, which is a client-side web application designed for the visualization and analysis of data associated with fitness and activity tracking devices. Next, we revisited and formalized Heart Rate Efficiency (HRE), defined as the product of pace and heart rate, as a practical and explainable metric to track aerobic fitness in everyday running.

Drawing on more than a decade of training data from one athlete, and supplemented by publicly available logs from twelve runners, we showed that HRE provides more stable and meaningful feedback on aerobic development than heart rate or pace alone. We showed that HRE correlates with training volume, reflects seasonal progress, and remains stable during long runs in well-trained individuals. We also discuss how HRE can support everyday training decisions, improve the user experience in fitness tracking, and serve as an explainable metric to proprietary ones of commercial platforms. Our findings have implications for designing user-centered fitness tools that empower amateur athletes to understand and manage their own performance data.

More information:

https://dl.acm.org/doi/10.1145/3749012.3749046

CHI Greece 2025 Article I

Today, a paper I co-authored with colleagues from CYENS, University of Cyprus and cognitiveux which was presented at CHI Greece 2025. The paper is entitled “The Reel Deal: Designing and Evaluating LLM-Generated Short-Form Educational Videos”. It presents a web-based system that uses large language models (LLMs) to automatically generate structured short-form video (i.e., reels) from lecture long-form videos while preserving instructor-authored material.

In a between-subject user study with 62 university students, we evaluated ReelsEd and demonstrated that it outperformed traditional long-form videos in engagement, quiz performance, and task efficiency without increasing cognitive load. Learners expressed high trust in our system and valued its clarity, usefulness, and ease of navigation. Our findings point to new design opportunities for integrating generative AI into educational tools that prioritize usability, learner agency, and pedagogical alignment.

More information:

https://dl.acm.org/doi/10.1145/3749012.3749048

Re-Inventing the Mouse

Researchers from Nazarbayev University in Kazakhstan introduced two entirely new ergonomic mouse designs. The first, called the Fleximouse, is a squeezable, pink mesh blob that allows users to move the cursor by making slight adjustments to their grip. The second, a vertical A-frame design somewhat resembling a drawing compass, was intended to keep the hand in a more comfortable position that requires less wrist movement overall.

And while the conventional computer mouse is far from perfect, testing suggests that its general adaptability to a wide range of hand shapes and sizes may contribute to its longevity. Still, the researchers say their findings could help inform the design of future mice by incorporating softer, less rigid features that more naturally align with the subtle movements of the human hand.

More information:

https://www.popsci.com/technology/squeezable-computer-mouse/

Universe Simulation on Laptop

Scientists have introduced Effort.jl, an innovative emulator that mimics the behavior of advanced cosmological models with striking accuracy, sometimes even improving on them, while running in minutes on a laptop.

Built on a neural network, the tool learns model responses and integrates known physics to cut training time, which promise to greatly deepen our knowledge of the Universe on large scales.

More information:

https://scitechdaily.com/scientists-just-found-a-way-to-simulate-the-universe-on-a-laptop/

Swarm Intelligence

A new framework has been designed to push forward swarm intelligence, the branch of AI that mimics the group behaviors of birds, fish, and bees. The coordinated movement of robots could improve search-and-rescue operations and wildfire detection.

The collective intelligence found in nature is a wonder of efficiency and coordination. Birds flock to forage. Fish school as a way to avoid predators. Bees use swarming as their method of reproduction.

More information:

https://interestingengineering.com/innovation/fish-birds-inspire-robot-swarm-control

Acoustic AI for Autonomous Driving

A new wave of acoustic technology could give vehicles the missing sense that cameras and radar can’t provide. By detecting sirens before they’re visible, picking up the chatter of pedestrians, or even transmitting urgent sounds through a driver’s headrest, researchers are teaching cars to react to the world the way humans do by listening. Unlike optical systems, which need a clear line of sight, acoustic sensors can pick up what’s happening around corners or in crowded streets. That ability could prove essential for autonomous driving, where every millisecond of awareness matters.

Sensors are designed to stand up to rain, wind, and extreme temperatures, with careful placement ensuring accurate pickup even at highway speeds. Testing has taken the car from Portugal to the Arctic Circle to stress the technology in real conditions. To make sure drivers don’t miss critical cues, important noises can also be piped directly into the cabin via the headrest. That means a siren, horn, or warning call is not just detected but delivered right to the driver’s ear, helping them respond faster. The project involves close collaboration with automotive suppliers and manufacturers.

More information:

https://interestingengineering.com/transportation/hearing-car-autonomous-driving-sound-recognition

AI-Driven Private School

An AI-driven private school, called Alpha School, is opening a Northern Virginia campus this fall, charging up to $65,000 annually. Students will spend two hours daily on academics via adaptive apps like IXL, then focus on life skills and workshops.

Instead of teachers, AI guides oversee learning and activities. Backed by billionaire investors, Alpha is expanding to 12 campuses nationwide while seeking approval to adapt its model in charter schools.

More information:

https://www.washingtonpost.com/education/2025/08/26/alpha-school-virginia-ai-education/