Weekly Piece of Future #156
From Bionic Vision to Wireless Brain Chips and Light-Powered Computing
Hey there, fellow future-addicts!
Welcome to this week's edition of Rushing Robotics! This edition showcases robots with insect vision, brain chips tested in space, and medical breakthroughs that stop bleeding in seconds. From AI moving to your desktop to humanoid robots joining factory floors, these innovations prove we're solving yesterday's impossible problems today.
🤯 Mind-Blowing
A 1,000-lens compound eye gives robots 180-degree vision and smell detection simultaneously, while brain chips successfully operated in orbit for the first time. A $250 desktop AI computer delivers 24 trillion operations per second without cloud dependency, and autonomous vehicles gained superhuman reflexes responding four times faster than human brains. Meanwhile, iron catalysts achieved 85% hydrogen fuel efficiency, potentially slashing fuel-cell vehicle costs by replacing expensive platinum.
🔊 Industry Insights & Updates
Light-powered computing is solving optimization problems faster than conventional systems using laser pulses at room temperature. Meta broke ground on a $10 billion AI data center in Indiana with 1 gigawatt capacity, while Apptronik secured $520 million to deploy Apollo humanoid robots already working at Mercedes-Benz and logistics facilities. Italian shipyards welcomed humanoid robot welders through a partnership targeting deployment within two years.
🧬 BioTech
MIT's magnetic mixer solved 3D bioprinting's cell-settling problem, enabling consistent functional tissue production. RNA micelles nearly eliminated metastatic cancer tumors in 26 days by combining chemotherapy with gene therapy. A revolutionary spray stops severe bleeding in under one second by transforming blood into gel, awaiting clinical trials for emergency deployment.
💡 Products/Tools of the Week
Parano.ai automatically tracks competitor activities across marketing, pricing, and product changes, delivering AI-summarized alerts to save research time. Vicoa enables seamless AI coding sessions across devices with real-time synchronization and one-tap approvals. QVeris AI provides a unified API connecting LLMs to 10,000+ third-party tools with sub-50ms latency and enterprise security. Google DeepMind's Project Genie transforms text or image prompts into interactive, photorealistic 3D environments in real-time, eliminating traditional modeling expertise requirements.
🎥 Video Section
Boston Dynamics showcases Atlas performing airborne maneuvers, Nick Builds constructs a laundry-folding robot in 24 hours, and UBTECH's autonomous logistics vehicle completes validation testing at Foxconn.
The convergence of biology, AI, and robotics is accelerating beyond prediction. These aren't concepts—they're working systems reshaping our world right now. What once seemed impossible is now deployed and operational, transforming industries and saving lives. Stay hungry, stay futurish!
🤯 Mind-Blowing
A 1,000-lens sensor eye has been developed that gives robots 180-degree vision and the ability to detect smells simultaneously. Researchers at the Chinese Academy of Sciences created an artificial compound eye modeled after the fruit fly, fitting 1,027 visual units into a sensor measuring just 1.5 millimeters using advanced laser two-photon polymer printing. The device includes tiny hair-like structures called setae between the lenses that prevent moisture accumulation and protect from dust and debris. Engineers integrated a bionic nose using an inkjet-printed chemical array that changes color when detecting harmful gases, merging vision and smell into a single lightweight unit for small robotic systems. The bio-CE system successfully identified moving objects and navigated around obstacles during testing, detecting threats from both front and sides simultaneously without requiring head rotation.
A wireless brain-computer chip has been tested in space for the first time by Chinese researchers, successfully operating in orbit to collect brain signals despite harsh space conditions. Northwestern Polytechnical University (NPU) developed and launched the implantable brain-computer interface device into orbit last December as part of a specialized space experiment platform. The operational BCI was immersed in simulated body-fluid settings during testing and successfully maintained consistent electroencephalogram collection despite extreme challenges, demonstrating that sensitive neural electronics can remain functional without degradation or short-circuiting outside Earth's atmosphere. The experiment provided crucial information about hardware resilience, noise interference effects, electrode durability in space environments, and how microgravity influences neural firing patterns. The NPU team created a flexible electrode array designed to conform to the brain's natural shapes ensuring a comfortable noninvasive fit, surpassing standard metal counterparts with dramatically improved signal stability metrics in animal testing.
A pyramid-shaped desktop computer has been launched that runs AI tasks locally rather than on the cloud, offering 24 trillion operations per second at a price of $250. M5Stack introduced the AI Pyramid Pro, an edge-AI device equipped with 8GB RAM, 32GB internal storage, HDMI 2.0, four USB 3.0 ports, and two Type-C ports designed for real-time computer vision, speech comprehension, and object recognition. The device uses an Arm CPU and dedicated neural processor instead of a GPU to maintain energy efficiency and low power consumption while running small to medium-sized language models and managing multiple camera feeds simultaneously. Local AI processing provides enhanced privacy, reduced latency with near-instantaneous responses, elimination of per-use cloud fees, and the ability to operate without continuous internet connection. The Pyramid Pro targets edge-AI developers, embedded systems engineers, and creators working on advanced AI projects rather than general consumers.
A new autonomous car vision system has been developed that responds to threats four times faster than the human brain. Researchers from China, the UK, Hong Kong, Saudi Arabia, and the US created a hardware-based reflex mechanism designed to accelerate decision-making in automated driving. The system uses a two-dimensional synaptic transistor array functioning as a motion detection chip that identifies changes in images in just 100 microseconds, significantly outpacing human perception which requires roughly 150 milliseconds to react. The technology employs a filter-then-process methodology inspired by human visual processing, disregarding entire images and recording only moving objects before relaying selected signals to computer vision algorithms.
A new catalyst has been developed that achieves 85% hydrogen fuel efficiency by solving the decades-old problem of stabilizing iron catalysts for fuel cells. Scientists at Washington University in St. Louis created a technique using chemical vapor of gases to stabilize iron catalysts during thermal activation, making them viable for proton membrane fuel cells. This breakthrough could dramatically reduce fuel-cell vehicle costs, as platinum catalysts currently account for about 45% of a fuel cell stack's expense. Iron being plentiful and inexpensive presents a viable alternative to scarce platinum, though it previously struggled with stability in the acidic conditions required. The technology targets heavy-duty vehicles like transport trucks and buses that operate from centralized locations, making hydrogen refueling logistics more manageable.
🔊 Industry Insights & Updates
A new light‑powered machine has been built that solves complex optimisation problems faster than conventional computers, using pulses of laser light to emulate magnetic spins in an Ising model. Researchers at Queen’s University, led by associate professor Bhavin Shastri, assembled a room‑temperature optoelectronic oscillator system that can process up to 256 coupled “spins” with just five optical components, and already outperforms older optical Ising machines by a wide margin. The device, demonstrates how billions of operations per second can be achieved without the need for cryogenic cooling, offering a promising pathway for tackling real‑world tasks such as drug discovery, cryptographic analysis, and logistics optimisation.
Construction has started on a $10 billion AI data center in Lebanon, Indiana, by Meta that will provide 1 gigawatt of power capacity. The facility represents one of Meta's most substantial infrastructure investments and will serve as the company's second data center in Indiana, engineered to manage digital platforms and rapidly expanding AI workloads. The project is anticipated to generate over 4,000 jobs during peak construction phases with approximately 300 permanent positions once operational. Meta committed to matching electricity consumption entirely with clean energy, achieving LEED Gold certification, utilizing closed-loop liquid-cooling systems, and returning 100 percent of water consumed back to local watersheds through collaboration with Arable on irrigation technology for farmers.
A $520 million funding round has been secured by US-based robotics company Apptronik to deploy its Apollo humanoid robots in manufacturing settings, achieving a $5 billion valuation with backing from Google and Mercedes-Benz. Apollo robots are already operational at partner facilities including Mercedes-Benz, GXO Logistics, and others where they assist in optimizing manufacturing and warehouse tasks. The flagship Apollo robot stands 5'8" tall, weighs 160 pounds, can handle loads of up to 55 pounds, and functions for four hours on a single battery charge with hot-swappable battery packs for continuous operation. Google DeepMind recently demonstrated Apollo's capabilities powered by Gemini AI, showcasing its ability to manage unfamiliar items, follow spoken instructions, and adjust in real-time to varying circumstances without retraining in new environments.
A humanoid robot welder has been introduced to work alongside humans in Italian shipyards through a partnership between Generative Bionics and Fincantieri. The robot will function as a welder enhancing safety, boosting efficiency, and improving production quality in naval production facilities. Generative Bionics will develop the robot with artificial intelligence and advanced capabilities in manipulation, perception, and vision to oversee welding seams and operate in intricate environments. The four-year collaboration targets swift development with initial testing slated for completion by the end of 2026, with operational capabilities expected within the first two years followed by broader implementation at Fincantieri's Monfalcone shipyard.
🧬 BioTech
A magnetic mixer called MagMix has been developed by MIT researchers that creates the most uniform 3D-printed tissues by preventing cell settling during the printing process. The device addresses the problem where living cells denser than hydrogels sink to the bottom of printer syringes, causing blockages and uneven tissue distribution. MagMix consists of a small magnetic propeller inside the bioprinter syringe and an external motor that controls the propeller without direct contact with bio-ink, preventing contamination. Testing showed MagMix prevented cell settling for over 45 minutes of continuous printing while maintaining high cell viability, enabling production of functional muscle tissues. Ritu Raman, Eugene Bell Development Professor at MIT, stated the system allows mixing speeds to be fine-tuned for different bio-inks while exerting minimal stress on cells.
Metastatic colorectal cancer tumors in mouse lungs were almost completely depleted within 26 days using newly developed RNA micelles that deliver chemotherapy and gene therapy directly to cancer cells. Researchers at The Ohio State University created these tiny molecular clusters by combining the chemotherapy drug gemcitabine with small interfering RNA that silences survivin, a protein helping cancer cells survive, and attached a ligand molecule to enhance binding to cancer cell receptors. The nanoparticles spontaneously target tumors without causing toxicity or immune responses. The study, showed the micelles attacked cancer in two ways by accumulating in tumor blood vessels and entering cells with the ligand's help. Co-author Daniel Binzel, a research assistant professor at Ohio State, explained that the chemotherapy kills cells while the small interfering RNA blocks survival gene expression, eliminating cancer through multiple pathways simultaneously.
A revolutionary spray has been developed that stops severe bleeding almost instantly by transforming blood into a gel-like substance in under one second. Researchers at Korea Advanced Institute of Science and Technology (KAIST) created this hemostatic agent designed for military and emergency medical use where traditional methods prove ineffective. The spray contains three natural ingredients: alginate from seaweed that forms gel on contact with blood, gellan gum from bacteria that stabilizes the gel structure, and chitosan from crustacean shells that attracts blood cells to accelerate clotting. The powder absorbs approximately seven times its weight in blood and works without requiring direct pressure on wounds. The technology awaits clinical trials and regulatory approval before deployment in military medical kits and hospitals.
💡Products/tools of the week
A new AI-powered competitive intelligence platform is now available to help businesses automatically track competitor activities and reclaim valuable research time. Parano.ai continuously monitors competitors' marketing, product, pricing, reviews, hires, funding, and website changes, detecting meaningful shifts with visual diffs and delivering AI-summarized alerts directly to email, Slack, or Microsoft Teams. The platform's AI analyzes patterns and context to explain why changes matter, enabling GTM, product, and leadership teams to act faster on strategic moves without wasting hours on manual research. Companies can set up tracking for competitor sets within minutes and receive comprehensive weekly summaries with executive insights, all without needing a credit card to start.
A new cross-device AI coding companion has been launched to enable seamless remote development workflows. Remote AI coding sessions can now be managed from any device through Vicoa, a platform that runs AI agents like Claude Code and Codex remotely and provides a clean interface with real-time session synchronization, searchable history, push notifications, and one-tap code diffs and approvals. Developers can start a coding session on a laptop and seamlessly continue on a phone or tablet without losing context, while Vicoa handles session management, approvals, and secure connections. The tool installs as a simple CLI command and keeps AI central to the workflow, allowing agents to write, review, and interact with code while developers avoid babysitting long-running tasks. This makes Vicoa ideal for developers who want to maintain productivity on the go and keep their AI-driven workflows synced across all devices.
QVeris AI is a Tool OS for AI agents that provides a single, OpenAI‑SDK‑compatible API and SDKs to discover, route, sandbox, and execute 10,000+ third‑party tools (search, finance, OCR, weather, etc.) so LLMs can automatically find, select, and call real‑world services; it includes plugins (Cursor), a QVerisBot assistant option, and examples to get started quickly. It’s built for production with distributed routing, smart failover, edge nodes (<50ms avg latency), 99.9% SLA, and enterprise security (encryption and fine‑grained access control). Developers and teams use it to eliminate adapter work, accelerate agent development.
A new experimental research prototype has emerged that transforms simple text or image prompts into photorealistic, interactive 3D environments generated in real time. Google DeepMind developed Project Genie using Genie 3, a general-purpose AI world model that enables users to sketch worlds, build characters, and immediately explore navigable scenes that dynamically expand around them. The system simulates physics, remembers prior interactions, and allows creators, researchers, and prototypers to rapidly test game levels, storytelling scenes, and robotics or simulation concepts without traditional 3D modeling expertise. By emphasizing AI-driven on-the-fly world generation and remixing, this tool removes technical barriers and accelerates creative iteration for diverse applications across entertainment and research.
