Lifeboat Foundation: Safeguarding Humanity
Toggle light / dark theme

Blog

Category: robotics/AI – Page 15

A new flexible AI chip for smart wearables is thinner than a human hair

The promise of smart wearables is often talked up, and while there have been some impressive innovations, we are still not seeing their full potential. Among the things holding them back is that the chips that operate them are stiff, brittle, and power-hungry. To overcome these problems, researchers from Tsinghua University and Peking University in China have developed FLEXI, a new family of flexible chips. They are thinner than a human hair, flexible enough to be folded thousands of times, and incorporate AI.

A flexible solution

In a paper published in the journal Nature, the team details the design of their chip and how it can handle complex AI tasks, such as processing data from body sensors to identify health indicators, such as irregular heartbeats, in real time.

Tiny silicon structures compute with heat, achieving 99% accurate matrix multiplication

MIT researchers have designed silicon structures that can perform calculations in an electronic device using excess heat instead of electricity. These tiny structures could someday enable more energy-efficient computation. In this computing method, input data are encoded as a set of temperatures using the waste heat already present in a device.

The flow and distribution of heat through a specially designed material forms the basis of the calculation. Then the output is represented by the power collected at the other end, which is a thermostat at a fixed temperature.

The researchers used these structures to perform matrix vector multiplication with more than 99% accuracy. Matrix multiplication is the fundamental mathematical technique machine-learning models like LLMs utilize to process information and make predictions.

New light-emitting artificial neurons could run AI systems more reliably

Over the past decades, computer scientists have developed increasingly advanced artificial intelligence (AI) systems that perform well on various tasks, including the analysis or generation of images, videos, audio recordings and texts. These systems power various highly performing software, including automated transcription apps, large language model (LLM)-powered conversational agents like ChatGPT, and various other platforms.

A Breakthrough That Cuts Blockchain Delays Nearly in Half

The idea of a fully connected digital world is quickly becoming real through the Internet of Things (IoT). This expanding network includes physical devices such as small sensors, autonomous vehicles, and industrial machines that collect and exchange data online.

Protecting this data from tampering is essential, which has led engineers to explore blockchain as a security solution. Although blockchain is widely known for its role in cryptocurrencies, its core function is as a decentralized digital ledger. Instead of data being controlled by a single organization, information is shared and maintained across many computers.

Hugging Face abused to spread thousands of Android malware variants

A new Android malware campaign is using the Hugging Face platform as a repository for thousands of variations of an APK payload that collects credentials for popular financial and payment services.

Hugging Face is a popular platform that hosts and distributes artificial intelligence (AI), natural language processing (NLP), and machine learning (ML) models, datasets, and applications.

It is considered a trusted platform unlikely to trigger security warnings, but bad actors have abused it in the past to host malicious AI models.

Unlike traditional #AI that just generates static images or videos

Genie is a “World Model.” It doesn’t just show you a scene; it simulates the physics, the depth, and the logic of a world you can actually control and navigate in real-time.

Stay up to date with the latest Google AI experiments, innovative tools, and technology. Explore the future of AI responsibly with Google Labs.

Researchers Show AI Robots Vulnerable to Text Attacks

“I expect vision-language models to play a major role in future embodied AI systems,” said Dr. Alvaro Cardenas.

How can misleading texts negatively affect AI behavior? This is what a recently submitted study hopes to address as a team of researchers from the University of California, Santa Cruz and Johns Hopkins University investigated the potential security risks of embodied AI, which is AI fixed in a physical body that uses observations to adapt to its environment, as opposed to using text and data, and include cars and robots. This study has the potential to help scientists, engineers, and the public better understand the risks for AI and the steps to take to mitigate them.

For the study, the researchers introduced CHAI (Command Hijacking against embodied AI), which is designed to combat outside threats to embodied AI systems, including misleading text and imagery. Instead, CHAI employs counterattacks that embodied Ais can use to disseminate right from wrong regarding text and images. The researchers tested CHAI on a variety of AI-based systems, including drone emergency landing, autonomous driving, aerial object tracking, and robotic vehicles. In the end, the researchers discovered that CHAI successfully identified incoming attacks while emphasizing the need for enhancing security measures for embodied AI.

When AI Builds AI

Leading artificial intelligence companies have started to use their own systems to accelerate research and development, with each generation of AI systems contributing to building the next generation. This report distills points of consensus and disagreement from our July 2025 expert workshop on how far the automation of AI R&D could go, laying bare crucial underlying assumptions and identifying what new evidence could shed light on the trajectory going forward.

Stephen Wolfram: computation is the universe’s OS

Mathematica creator Stephen Wolfram has spent nearly 50 years arguing that simple computational rules underlie everything from animal patterns to the laws of physics. In his 2023 TED talk, he makes the case that computation isn’t just a useful way to model the world — it’s the fundamental operating system of reality itself.

Wolfram introduces “the ruliad,” an abstract concept encompassing all possible computational processes. Space and matter, he argues, consist of discrete elements governed by simple rules. Gravity and quantum mechanics emerge from the same computational framework. The laws of physics themselves are observer-dependent, arising from our limited perspective within an infinite computational structure.

On AI, Wolfram sees large language models as demonstrating deep connections between semantic grammar and computational thinking. The Wolfram Language, he claims, bridges human conceptualization and computational power, letting people operationalize ideas directly — what he calls a “superpower” for thinking and creation.

/* */