Weekly Piece of Future #152

From Self-Healing Robot Nerves to Bioprinted Livers and Machines That Lip-Sync

Jan 16, 2026

Article voiceover

0:00

-19:29

Hey there, fellow future-addicts!

Welcome to this week's edition of Rushing Robotics! As we push further into 2026, the pace of innovation shows no signs of slowing. This week brings us extraordinary breakthroughs spanning robotics, quantum computing, bioengineering, and nuclear-powered AI infrastructure. From self-healing electronic nerves that give robots human-like reflexes to 3D-bioprinted livers destined for transplantation, the future is unfolding before our eyes. Let's dive into the developments that are reshaping our world.

🤯 Mind-Blowing

Robots are gaining graded “pain” reflexes with self-healing, jelly-like electronic nerves that sense pressure in multiple intensities (not just on/off). ARPA‑H is funding bioprinted “bridge livers” designed to support patients for 22–44 weeks, potentially reducing the need for full transplants. On the compute side, a new multi-physics architecture boosts Fourier-transform speed to ~500500 billion ops/sec, while MIT’s MechStyle lets you restyle 3D models without weakening them. Quantum scaling also took a leap: metasurfaces generating up to 360,000360,000 optical tweezers could help neutral-atom systems push toward 100,000+100,000+ qubits.

🔊 Industry Insights & Updates

Meta is locking in up to 6.66.6 GW of nuclear power by 2035 to fuel AI superclusters—one of the clearest signs energy is now core AI strategy. Robots are also learning dexterous motion from minimal data, and “transparent wood” smart windows can switch from ~28%28% to ~78%78% light transmission as temperatures rise—no electricity required. Cheap 3D-printed optics are improving too, with custom super-resolution lens elements reportedly produced for under 11 dollar.

🧬 BioTech

A “womb-on-a-chip” now mimics early embryo implantation phases, offering a new window into the earliest days of pregnancy. Rice’s CLASSIC approach pairs huge DNA-circuit libraries with ML so scientists can predict new genetic designs, not just test them. MIT also built ingestible pills that signal from the stomach within ~1010 minutes to confirm dosing—promising for transplant meds and long-term therapies.

💡 Products/Tools of the Week

Note67 records meetings with on-device Whisper transcription and local AI summaries via Ollama for privacy-first notes. AcaDraw turns prompts into publication-ready scientific diagrams and editable SVGs in seconds. Feynn generates source-verified competitive and strategic research with scenario simulation for decision support. Diffray speeds up GitHub code reviews with multi-agent analysis focused on real issues (security, performance, concurrency) while avoiding style nitpicks.

🎥 Video Section

Robots are getting more expressive and more real-world-ready—lip-sync demos, humanoid platforms, and agentic operating systems for physical robots are accelerating fast.

The exciting part: we’re watching the shift from “cool prototypes” to systems that scale, self-correct, and integrate into real infrastructure. Biology and computation are increasingly borrowing each other’s tricks—healing, sensing, adapting. If this pace holds, the next few years won’t just add smarter tools; they’ll change what we consider possible to build and deploy. Stay hungry, stay futurish!

🤯 Mind-Blowing

An innovative electronic 'nerve' system that enables robots to exhibit human-like reflexes has been developed, featuring the ability to detect pressure at varying intensities rather than simple binary responses. Researchers from Northeast Normal University in China created these jelly-based electronic nerves that can autonomously repair themselves after sustaining damage, with sensitivity that increases following an 'injury' and diminishes during 'recovery'. The technology uses memristors with 16 stable levels to encode different pain intensities, ranging from no pain to severe, functioning like a dimmer switch rather than a conventional on/off mechanism. Central to this advancement is gelatin, a collagen-derived protein that conducts ions efficiently and exhibits self-healing properties when heated to 60°C, creating systematic conductive pathways corresponding to the 16 pain levels.
A $28.5 million grant has been obtained from the Advanced Research Projects Agency for Health (ARPA-H) in the United States to develop functional 3D bioprinted livers intended for transplantation. The initiative, named Immunocompatible Volumetric Engineering (LIVE), aims to tackle the critical shortage of donor organs by producing bioengineered tissue that can help treat acute liver failure. Carnegie Mellon University leads the project with a team comprising specialists in liver stem cells, biomanufacturing, transplantation, and regenerative medicine, collaborating with experts from the University of Pittsburgh, Mayo Clinic, and University of California. The primary objective is to develop a temporary liver that can provide support for approximately two to four weeks while the patient's own liver recuperates, thus potentially eliminating the necessity for a complete transplant. The LIVE team employs a proprietary FRESH 3D bioprinting method that allows them to print soft materials such as collagen and human stem cells into intricate structures using hypoimmune cells engineered to be "universal donor" cells that are undetectable by the human immune system.
An innovative computing architecture has been developed that enhances processing capabilities by nearly fourfold, paving the way for advancements in fields like artificial intelligence, sensing, brain-inspired computing and communication technologies. Researchers from Peking University ingeniously combined two unique devices to synchronize frequency conversion within a multi-physics-domain framework, creating a flexible system capable of executing intricate tasks including the Fourier transform. The architecture allows multiple computing paradigms to function within their optimal physical environments. the unified system leverages the advantages of two devices involved in frequency generation, modulation, and in-memory computing, ensuring precision while reducing power consumption and increasing the processing speed of the Fourier Transform from approximately 130 billion operations per second to around 500 billion.
Customizable 3D-printed items that maintain their structural integrity have become possible through an innovative generative AI platform named MechStyle developed by researchers at MIT. The system addresses a prevalent challenge in AI-assisted design where generative models can easily alter objects for aesthetic purposes, but these modifications frequently lead to issues such as snapping, bending, or collapsing post-manufacturing, with most existing AI technologies prioritizing appearance over functionality. MechStyle provides users the option to upload their own 3D models or select from pre-existing templates such as vases and wall hooks, then apply various styles through text or image prompts, with the generative AI model adjusting the geometry while a simulation assesses the impact of these alterations on strength.
An innovative method has been introduced that could enable quantum computers to expand beyond 100,000 qubits by integrating optical tweezers and metasurfaces. Researchers led by Dr. Sebastian Will and Dr. Nanfang Yu at Columbia University successfully captured 1,000 strontium atoms using metasurfaces—ultra-thin optical devices crafted from millions of nanoscale pixels that manipulate light into thousands of tightly focused laser spots at once. The team produced a 35-millimeter-wide metasurface containing over 100 million pixels capable of generating 360,000 optical tweezers—two orders of magnitude greater than current technologies. Constructed from silicon nitride and titanium dioxide, these metasurfaces can endure extremely intense lasers with optical intensities exceeding 2,000 watts per square millimeter, about a million times more intense than sunlight.

Don't keep the future all to yourself! Share a bit with your friends and family because everyone deserves a Piece of Future!

Three major energy partnerships have been announced to deliver up to 6.6 gigawatts of nuclear power by the year 2035, positioning Meta among the largest corporate nuclear energy buyers in American history. The technology powerhouse secured contracts with Oklo, TerraPower, and Constellation to provide clean, dependable electricity for Meta's next-generation AI infrastructure, particularly the Prometheus supercluster situated in New Albany, Ohio. Meta's partnership with TerraPower will support development of two new Natrium units generating up to 690 MW of power as soon as 2032, plus rights to energy from six additional units for a total of 2.1 GW of baseload capacity and 1.2 GW of integrated storage. The company is backing Oklo's establishment of a nuclear technology campus in Pike County, Ohio, expected to become operational by 2030 and contribute 1.2 GW of clean power to the PJM market using Powerhouse reactors that can process both fresh and recycled fuel.
A novel adaptive motion reproduction system has been created that enables robots to mimic human-like movements with remarkably minimal training data. Researchers in Japan, led by Akira Takakura from Keio University, devised the system using Gaussian process regression to tackle the challenge of robots struggling when object characteristics such as weight, stiffness, or texture are altered. The innovative technique transcends linear models by effectively mapping intricate nonlinear relationships even with a limited dataset, recording human grasping movements on various objects with different stiffness to discern the connection between object characteristics and the forces and positions applied by humans.
An innovative smart window made from wood has been created that adapts to light conditions, blocks harmful ultraviolet rays, and enhances insulation, all without the need for electricity. Researchers developed the transparent wood by merging modified balsa wood with polymer-dispersed liquid crystals to control visibility and thermal transfer within structures. Laboratory experiments revealed that the material transitions from opaque to clear as temperatures increase, permitting around 28% of visible light to pass through at ambient room temperature while at 40 degrees Celsius this transmittance rises to roughly 78% at a wavelength of 550 nanometers. The transparent wood provides excellent protection against ultraviolet radiation, effectively blocking nearly 100% of UVA rays due to molecular arrangement and π-π aggregation, safeguarding human skin and interior furnishings without significantly diminishing visible light.
Sophisticated optical components capable of super-resolution imaging have been created using affordable 3D printers and low-cost materials, with each custom lens produced for under $1. Researchers demonstrated that inexpensive, customizable optical elements can achieve imaging that surpasses the conventional diffraction limit by integrating 3D printing, silicone, and UV-curable resin to fabricate multi-element lenses. The team, led by Ralf Bauer from the University of Strathclyde in the UK, constructed a multifocal illumination microscope that resolved biological structures at the nanoscale.

🧬 BioTech

A miniature womb on a chip has been developed that accurately simulates how human embryos attach to and penetrate the uterine lining during the crucial first days of pregnancy. Researchers from the Chinese Academy of Sciences created the breakthrough technology by embedding human endometrial cells within gel-like layers inside a microfluidic chip, allowing the cells to organize into a structure that closely resembles actual uterine tissue. The team tested the artificial uterine lining using both real human blastocysts and lab-generated blastoids, which successfully completed all three phases of implantation: initial contact with the uterine surface, stable attachment, and active tissue penetration.
An innovative method has been introduced that has the potential to transform the way DNA is engineered for medical and biotechnological uses by employing machine learning alongside extensive libraries of DNA designs. Researchers from Rice University unveiled a breakthrough termed CLASSIC—short for "combining long and short range sequencing to investigate complexity"—that enables scientists to generate hundreds of thousands to millions of DNA designs simultaneously, far exceeding previous capabilities. This was the first instance where AI and machine learning could be applied to analyze circuits and accurately predict untested ones because until now no one had been able to create libraries as expansive as theirs.
Innovative pills capable of transmitting signals from within the stomach have been created by scientists from the Massachusetts Institute of Technology (MIT), allowing the pill to notify when it has been ingested. The technology features a biodegradable radio frequency antenna made of zinc encased within a cellulose particle, with the antenna coiled and housed within a capsule alongside the medication. Once ingested, the capsule’s coating constructed from gelatin and coated with glucose and molybdenum tungsten disintegrates, allowing the drug and RF antenna to be released, with the antenna detecting an RF signal from an external receiver and transmitting a confirmation signal within ten minutes. The innovative system holds promise for tracking transplant recipients who must adhere to immunosuppressive medication regimens, as well as individuals suffering from chronic infections like HIV or tuberculosis who require long-term treatment

💡Products/tools of the week

Note67 captures meeting audio and screen content while performing Whisper-powered on-device transcription with advanced speaker separation and echo handling. The application generates private AI summaries using local LLMs via Ollama, ensuring all transcripts and summaries remain exclusively on the user's Windows or macOS machine. Note67 runs completely locally, making it ideal for users who want real-time AI-powered notes with complete data privacy without any cloud dependencies.
AcaDraw transforms plain-language prompts in English or Chinese into publication-ready scientific illustrations, diagrams, and editable SVGs within seconds. The AI-powered science drawing tool leverages advanced image models including Nano Banana and Seedream 4.0 to deliver high-fidelity strokes and vector-style outputs. Researchers, educators, and students can utilize instant live previews and an interactive chat for iterative refinements, generating precise presentation- and journal-quality figures without requiring manual drawing skills.
An AI-driven strategic intelligence platform delivers multi-agent, source-verified research on companies, competitors, suppliers, and industries, generating concise briefs while running scenario simulations to recommend traceable, action-ready decisions. Specialized agents map entities, surface real-time signals, and model trade-offs through a decision engine with guided game-theory simulations. Feynn minimizes hallucinations by linking every insight to verified sources, enabling leaders and strategy teams to move quickly with defensible foresight.
Speeds code reviews and reduces false positives by analyzing GitHub pull requests with full codebase context, diffray's coordinated multi-agent system of specialized AI agents surfaces high-confidence, actionable issues in security, concurrency, data consistency, testing, and performance while deduplicating results and avoiding style nitpicks. The AI-powered code review tool integrates through a GitHub app, enforces "culture as code" team rules, and ensures privacy by not storing code or using it for AI training, helping development teams catch real bugs faster.

Discussion about this post

Ready for more?