Nancy Kleinrock*

Content Director, Nemertes [Next]

Cybersecurity—first some good news: Tamper-evident security keys based on spin-orbit torque:

Spintronic system protects cryptographic keys

­Cybersecurity—and now some bad news: Although the title of this NIST report doesn’t sound overly ominous, we are in fact doomed, as Johna Till Johnson and John Burke, with contributions from Matthew Rosenquist, argued during this week’s Nemertes Live session “The Mythos Monster: AI, Cybersecurity, and You”:

NIST mathematical proof supports transition to a continuous-monitor-and-update security model for AI systems

Chemical engineering to the rescue—of Li–ion batteries. Oh DEER (direct electrode-to electrode regeneration)!

Electrical bath recycles critical minerals in batteries

Microchip miniaturization—chip components are now so small that the coating of the copper wires that connect transistors is the new miniaturization target, with crystalline tungsten disulfide offering a solution:

Revolutionary atom-thin coating from NUS CDE and Applied Materials partnership breaks new ground in chip miniaturization

Remote detection technologies—lidar can be notably improved by capturing polarization in addition to time-of-flight data:

New lidar system maps location, speed, and material properties in a single measurement

Biology—controlling metabolism:

Shell too snug? Hermit crabs have a fix

Biology/Mathematics—new analogy for collective animal behavior, in which the group is like a soft crystalline material and individuals are like its atoms on a lattice connected with spring-like bonds:

How do flocking birds and schools of fish move? New research offers crystal-clear answers

­”The path to big, systemic change is collective action.”—Gloria Feldt

*This email is sent from, but doesn’t originate with, Substack; it’s written by me, Nancy Kleinrock. I invite you to write to me at nancy.kleinrock@nemertes.com with your thoughts—positive, negative, or otherwise—about this posting or whatever’s on your mind. If someone forwarded it to you and you like what you see, feel free to subscribe by clicking the button below.

Subscribe now

This week we’re launching a pared down, more-to-the-point format for the Nemertes [Next] newsletter—and calling it “Offbeat Science.” It’s a collection of science, technology, and sometimes art or society news items that have jumped out to me during the preceding fortnight, but without extensive commentary. My goal to present items you’re less likely to see in the mainstream news (since you, the discerning reader, will already be familiar with those). Instead, I’m focusing on news that’s, well, offbeat.

I hope you find this new format so unburdensome that you’ll be more likely to dig into the items, and, most of all, I hope that you’ll like it. If you’re a paid member of the Nemertes community, you are also welcome to add publicly displayed comments to this and any editions going forward, plus any reader is welcome to reach out to me personally (see the P.S. at the end). I look forward to hearing from you.

And, now for some news . . .

Nancy Kleinrock*

Content Director, Nemertes [Next]

Robotics—coming apart and joining together, without global control:

Robotic “matter” flows, adapts through mechanical intelligence

­Robotics—democratized remote control:

New app allows anyone to operate a robot from their phone

Small form factors—being small myself, and will small fingers, this appeals to me, despite the very limited screen real estate. (In fact, ba.ck in the day when Palm Pilots were all the rage, my PDA was a Rolodex REX, a fully functional PDA in a diminutive PCMCIA form factor.):

This guy crammed a laptop into an Altoids tin

Quantum information sciences, technologies, and applications:

Colorado quantum incubator to launch the nation’s first open-access quantum timing testbed

Perfect randomness realized for the first time

Augmented reality:

New framework enhances AR experience by predicting where users will look

Aviation—which is worse: layovers or an insanely long flight?

Airbus tests passenger plane that can fly 22 hours non-stop

Beauty and ingenuity::

My spouse and I were at opening day of the farmer’s market in the Adirondack town where we live and saw some rustic chairs for sale—a bit too rustic for my liking. But trees meticulously pruned, grafted, and trained for the better part of a decade to be “naturally” chair-shaped by a horticulturist whose own spine had to be surgically straightened repeatedly throughout his youth due to a congenital condition? I’d be delighted to relax in one of these works of art.

We mould trees to grow into the shape of chairs

­”When I was a kid, all I knew was that I felt more comfortable sitting in one chair than in another. And now I realize it was because one chair was older. I still respond directly to the age of things..”—Robert Redford

*This email is sent from, but doesn’t originate with, Substack; it’s written by me, Nancy Kleinrock. I invite you to write to me at nancy.kleinrock@nemertes.com with your thoughts—positive, negative, or otherwise—about this posting or whatever’s on your mind. If someone forwarded it to you and you like what you see, feel free to subscribe by clicking the button below.

Subscribe now

This week and last Nemertes [Next] had some mind-bending sessions. You’ll hear about them soon from Johna.

Next up in Nemertes [Next] is our Wednesday, May 13 virtual field trip with a focus on energy and space, featuring Virtus Solis’s Ed Tate speaking about beaming solar power back to Earth via high-efficiency microwave transmission, ORBES’s Anna Shaposhnik speaking about her firm’s space drones that film and help maintain space assets, and Orbit Fab’s Mariah Hake explaining the importance of satellite refueling to all orbital endeavors and how the startup dominates this, um, space. While this virtual field trip won’t actually involve liftoff, the CTO, CEO, and chief engineer—respectively—will demo their firm’s devices and processes.

If you’re a Nemertes member, join us on this journey. If you aren’t one yet, take the first step by clicking this button. (You must select one of the paid options to attend the field trip):

More Info

In light of the ongoing virtual conference, I’ll again be sparse with the news this week. Here are a few quick hits for you to explore:

Jack Smith speaks out about AIs training their successors: Automating AI research: AI systems are about to start building themselves. What does that mean?

Which comes first: fusion or fission? How to build a better kind of nuclear power? This side hustle might help

And, today in geoengineering:

• A new idea to save the climate? Dam the Bering Strait

• Rainmaker’s cloud seeding breakthrough: First-ever proof of manmade precipitation, 143M gallons for Oregon and Utah

What do you give up to have your own voice? What does it mean to train an AI to speak like you?

Nancy Kleinrock*

Content Director, Nemertes [Next]

­”The one thing that you have that nobody else has is you. Your voice, your mind, your story, your vision.”—Neil Gaiman

*This email is sent from, but doesn’t originate with, Substack; it’s written by me, Nancy Kleinrock. I invite you to write to me at nancy.kleinrock@nemertes.com with your thoughts—positive, negative, or otherwise—about this posting or whatever’s on your mind. If someone forwarded it to you and you like what you see, feel free to subscribe by clicking the button below.

Subscribe now

Thanks to all who joined the first session of the four-Wednesdays Nemertes [Next] Spring 2026 virtual conference. The week’s topic of Innovation: Past, Present, and Future was gamely addressed from two very different angles by two dynamic speakers.

The first was David Pogue, who lifted his excitement about “Apple: The First 50 Years” off the pages of his newly released and persistently best-selling book, sharing stories of the iconic company’s founders and intermediary movers and shakers as well as the persistent through-lines that helped Apple become the world’s first trillion-dollar company. What have been those touchstones?

• sticking close to its roots

• a disciplined focus on taking complicated technology and making it accessible to—and highly desired by—ordinary people

• operational secrecy

• a strong aesthetic sensibility

• and—as a bit of an outlier—an enduring devotion to rounded rectangles.

Sharing the session with David was Rachel Higham, who shared a framework for successfully landing an identified innovation. The key is to focus less on the invention or process per se and instead on the surrounding ecosystem into which it will exist once launched. She identifies six pillars of scaled adoption:

• actively and iteratively codesigning the innovation with the system and its various stakeholders

• validating the value of the innovation early to ensure it aligns with the operational and economic realities and needs of the existing stakeholder base

• designing for equity and usability by listening to stakeholders

• engineering the implementation so that viable pathways for adoption are front and center

• orchestrating the ecosystem to eliminate barriers and minimize handoffs within the adoption process

• and explicitly addressing funding models to ease the adoption process—this final pillar is an innovation in its own right.

Rachel illustrated the power of this innovation framework with two examples: Macmillan Cancer Support, which generates and rolls out cancer treatments and care processes for UK patients; and Form1 Partners, which has orchestrated AI adoption in the financial services realm with Replicant automating 86% of car insurance phone calls. The central tenet in both case studies is the reliance on human beings for ideas and feedback, along with well-crafted, well-vetted, well-suited technology.

Of course, a Nemertes [Next] session would be nothing without the input from Nemertes community members. In this first conference session, that boiled down to questioning whether Steve Jobs’s “genius in a bottle” approach to innovation is still applicable within the ecosystem-centric model. In both cases, key ingredients are empathy, imagination, and the wisdom to say “no” to in-apt ideas. At Apple, these three components sat within Jobs himself; in Rachel’s framework, they bubble up through stakeholder design, testing, and feedback.

So that’s a recap for Nemertes [Next] Spring 2026 session #1. Members can access the video, presentation slides, and preconference reading materials—homework, if you will—within the Community portal. You can, too, by joining Nemertes as a member.

Coming next week is session #2, with a focus on The Future of Democracy of in the Age of AI. We will feature (a) security technologist and author Bruce Schneier discussing his new book “Rewiring Democracy: How AI Will Transform Our Politics, Government, and Citizenship” and the dozen-ish social and political changes AI will engender in the civic domain, and (b) Aditi Juneja and Ade Salami of Democracy 2076, an organization that takes the long view on how to foster and sustain democracy and what constitutes and how to combat antidemocratic forces. Aditi and Ade will specifically address the impact of political realignments and what the latest one—and the one to come—will mean for the United States.

Week three, on May 6, we’ll be hearing from University of Waterloo’s Travis Craddock on the topic of quantum biology and its potential to treat neuroinflammatory illness, and from Stanford’s Hiromitsu Nakauchi on the topic of growing human organs in animals from a patient’s pluripotent stem cells—and potentially in bodyoids: nonsentient, brainless humans lab-grown for the purpose of testing treatments or growing organs for transplantation. Clearly, ethical considerations abound.

The final week in this spring’s Nemertes [Next] conference, on May 13, we’ll embark on a virtual field trip to three startups in the space and/or energy realms: Edward Tate will take us on a tour of Virtus Solis, a space-AND-energy startup set to continuously beam solar energy to dedicated ground stations; Anna Shaposhnik will take us on a tour of ORBES and its robotic cameras purpose-built to film and maintain space infrastructure; and finally Mariah Hake will provide an update and tour of Orbit Fab and its satellite refueling technologies. Buckle up, it will be a wild ride!

All this is yours—if you’re a Nemertes member. If you aren’t one yet, take the first step by clicking is button.

More Info

With all that, I’ll be sparse with the news this week. Here are a few quick hits for you to explore:

Rainforests can bounce back much faster than thought, researchers say

Automated alignment researchers: Using large language models to scale scalable oversight

Plastic texturing kills viruses when they land

Nancy Kleinrock*

Content Director, Nemertes [Next]

­”I think it's a really big deal to be able to meet people outside the context of something like a conference room or someplace where everything feels like it's formal talk.”—Biz Stone

*This email is sent from, but doesn’t originate with, Substack; it’s written by me, Nancy Kleinrock. I invite you to write to me at nancy.kleinrock@nemertes.com with your thoughts—positive, negative, or otherwise—about this posting or whatever’s on your mind. If someone forwarded it to you and you like what you see, feel free to subscribe by clicking the button below.

Subscribe now

With the Nemertes [Next] spring 2026 virtual conference kicking off in just a week and a half, here’s what to expect over the next Wednesdays (5:00–6:30 pm, Eastern, starting April 22):

• April 22: INNOVATION: PAST, PRESENT, AND FUTURE

  • David Pogue, Author and Journalist: Apple: The First 50 Years

  • Rachel Higham, NED, Strategic Advisor, former-FTSE100 CTO, The Next Frontier of Innovation: Co-Design, Validation, and Real-World Impact

• April 29: THE FUTURE OF DEMOCRACY IN THE AGE OF AI

  • Bruce Schneier, Fellow, Berkman Klein Center; Lecturer, Harvard Kennedy School; Author: AI and Democracy

  • Aditi Juneja, Executive Director, Democracy 2076: Political Realignments: Where We Are Now and Shaping the Next One for a Stronger Democracy

• May 6: THE FUTURE OF HUMANS

  • Travis Craddock, Professor, University of Waterloo: Quantum Biology and its Potential to Treat Neuroinflammatory Illness

  • Hiromitsu Nakauchi, Professor, Stanford University: Growing Human Organs in Animals: Toward Rejection-Free Transplantation

• May 13: SPACE AND ENERGY STARTUP TOURS: A VIRTUAL FIELD TRIP

  • Edward Tate, Cofounder and CTO, Virtus Solis: Space-Based Solar Power

  • Anna Shaposhnik, Founder and CEO, ORBES: Keeping an Eye on Space Assets

As you can see, the long-established spirit of [Next] shines through with the purposefully eclectic mix of topics and speakers—some long-revered heavy hitters and others deeply creative and enthusiastic up-and-comers—all coming together over the course of a month with the goal of deeply engaging with the Nemertes membership around topics they’re passionate about and that will stimulate your curiosity and help guide your future direction.

If you are a Nemertes member, sign up to attend through Community; if you’re not yet a member, you can get started through this button.

More Info

And, now for some news . . .

Nancy Kleinrock*

Content Director, Nemertes [Next]

Check out this NYT article for a little bit of a preview into the April 22 session on Innovation, where David Pogue will be sharing tales he collected from 150 primary sources within Apple, including Nemertes advisory board member Ike Nassi (Chapel Hill, Mar 2024; virtual conference, Sep 2021; Washington, D.C., Oct 2014; Paris, Jul 2011; Brussels, Jul 2002). Last month, when Pogue launched his book Apple: The First 50 Years (pictured above) at the Computer History Museum (Gordon Bell, San Jose, Feb 2012; field trip to Computer History Museum, San Jose, Feb 2012), it was to a packed house of Apple luminaries; with the upcoming Nemertes [Next] session, it will be an intimate forum where attendees will not only hear David’s stories but also be able to converse with David, Ike, and one another.

One of Apple’s first employees looks back 50 years

­The focus of week two of the Nemertes [Next] spring virtual conference (Apr 29) is democracy—what threatens it and potential mitigations. Bruce Schneier (Boston, Mar 2022; Washington, D.C., Sep 2018; Austin, Feb 2001) will be making his case for both benefits and drawbacks of AI in this domain, as he does in his most recent book, Rewiring Democracy: How AI Will Transform Our Politics, Government, and Citizenship (also pictured above), whereas in his most recent past appearance he focused on concerns about AI’s effects on cybersecurity (Matthew Rosenquist, Nemertes Live, Feb 2026).

A.I. is on its way to upending cybersecurity

Like ARPA undertakings throughout its history (Alan Kay and Len Kleinrock, New York City and Washington, D.C., May 1014), the Advanced Research Projects Agency for Health (ARPA-H) seeks big, impactful results. A current ARPA-H project, dubbed Novel Innovations for Tissue Regeneration in Osteoarthritis (NITRO), has taken just two years from inception to being nearly ready for human trials of the treatment developed by a suite of researchers, including several from Colorado universities. As too many people are personally aware, osteoarthritis degrades interbone cartilage, or even the bone itself, causing debilitating pain. The new treatment entails a single arthroscopic injection of engineered proteins into the affected joint. Animal trials show “full regeneration of the defect” within a mere month or two.

A simple shot shows promise to reverse osteoarthritis within weeks3

At Wikipedia (Jimmy Wales, Washington, D.C., Dec 2005; Jersey City, Oct 2013), humans will continue to reign supreme. The online encyclopedia has officially adopted the policy that large language models may not be used to create new articles or add substantial new prose to existing articles, thereby codifying the value of collective human thought, reason, and discourse, which has led to the massive corpus of broadly accessible human knowledge, while recognizing the propensity for LLMs to hallucinate or introduce unsourced, unverifiable, or entirely fabricated content, complete with fictitious references (Wei Li, Half Moon Bay, Dec 2023; Soroush Vosoughi, Half Moon Bay, Dec 2023; Oren Etzioni, virtual conference, Jul 2020). However, Wikipedia contributors may use LLM-based tools for mundane tasks like basic copyediting.

Wikipedia:Large Language Models

­”We have lived in this world where little things are done for love and big things for money. Now we have Wikipedia. Suddenly big things can be done for love.”—Clay Shirky

*This email is sent from, but doesn’t originate with, Substack; it’s written by me, Nancy Kleinrock. I invite you to write to me at nancy.kleinrock@nemertes.com with your thoughts—positive, negative, or otherwise—about this posting or whatever’s on your mind. If someone forwarded it to you and you like what you see, feel free to subscribe by clicking the button below.

Subscribe now

Oftentimes, themes emerge in the Nemertes [Next] newsletter, where the news items flow in (semi)logical order, one after another. Not this time. Instead each article is distinctly different, variously probing the fundamental lack of quantum mechanical causality, importance of context when applying AI (in this case, to smart cities), building functional metamaterials from combinations of rice and sand, learning and unlearning fear, and using carbon capture to aerate beer—each fascinating in its own way.

So let’s have at it!

Nancy Kleinrock*

Content Director, Nemertes [Next]

Caption: Illustration of indefinite causal order, where the (quantum mechanical) pigeon can visit either Alice first.

Just as quantum mechanics allows a system to be in a superposition of states before measurement collapses the system into one specific state, it also allows for the counterintuitive occurrence of indefinite causal order; that is, the simultaneous potential for event A to occur before event B and for event B to occur before event A. Austrian researchers have just published their experimental work demonstrating just this by implementing a device-independent Bell-like inequality test. The results show that “no theory with a fixed causal description can explain the correlations [they] observe” in an experiment involving a time bin-based quantum switch where the halves of a split photon encounter event A and event B in opposite orders, and then recombine with retained correlation. The researchers conclude: “This experiment brings us closer to treating causal order as a resource on the same footing as entanglement.”

University of Vienna team reports experimental progress on indefinite causal order in PRX Quantum

­British researchers have challenged the notion of urban AI as a “single, inevitable next stage of the smart city.” Instead, like K Krasnow Waterman emphasized (Baltimore, Sep 2024), when it comes to AI, context is hugely important. When using AI to raise the intelligence level of a smart city, not only the urban geography but also the local politics and culture matter greatly. This work identifies three distinct AI urbanisms—predictive orchestration for Shenzhen, careful and modest counter-conduct for Boston, and procedurally bounded conditionality for Barcelona—with each emerging from the researchers’ six-facet analytical framework used to study autonomous cities. (Jennifer Mathieu, Regional Meeting, MITRE, Sep 2017; John Tolva, San Francisco, May 2016; Beth Coleman, London, Jul 2014; Vida Ilderem, Washington, D.C., Apr 2013; Veronika Haunold, Vienna, Jul 2013)

Urban AI should not be understood as a single, inevitable next stage of the smart city, say researchers

Grains of rice, when packed together, exhibit the unusual quality of being naturally stronger when compressed under slow pressure than when subjected to a sudden shock; in contrast, sand is stronger under fast-loading conditions. Mixing the two natural materials together in specified ratios leads to a new class of functional materials—speed-sensitive metamaterials—that can “bend, buckle, or stiffen differently under slow movements versus sudden impacts,” says the British lead of the international research team, all without electronics, sensors, or active control. This suggests the potential for robots that are more physically malleable, safer, lighter, and adaptable than conventional metal robots. (Hod Lipson, Virtual Field Trip, Oct 2025; Brooklyn, Jul 2016; Julia Greer, San Jose, Dec 2021; San Francisco, Dec 2015; Canan Dagdeviren, Washington, D.C., Sep 2018)

Rise of the rice robots—creating active smart materials

Dutch researchers using ultrasound waves carefully directed at the brain’s amygdala to influence how strong emotions like fear are stored—and, importantly, released/unlearned. It has long been known that this small brain structure plays a major role in the experience of fear, but until now its importance had not been demonstrated in the learning and unlearning of specific fears. The simultaneous stimuli of an amygdala-targeted ultrasound blast, an electric shock, and the image of a snake led to a slower fear response than in the absence of the ultrasound; plus, the loss of fear was quicker for stimulated subjects. These findings suggest the potential accelerated therapeutic outcomes of exposure therapy. (Tomaso Melodia, San Diego, Feb 2015; Rajesh Rao, San Francisco, Dec 1013; Angelika Dimoka and Paul Pavlou, Paris, Jul 2011)

Targeted ultrasound may help the brain overcome fear more quickly

During the December 2022 field trip to Arizona State University, Travis Johnson provided an explanation and tour of the MechanicalTree direct-air carbon capture project. While the direct-air carbon capture being conducted by the Almanac Beer Company in Alameda, California, using instrumentation created by Berkeley-based Aircapture, isn’t of the same magnitude as ASU’s, the use of its captured carbon dioxide will elicit a cheer—and perhaps a burp—from the brewery’s patrons as it infuses its ales, lagers, and stouts with the air-captured CO2, even if those belches re-release the greenhouse gas and negates the effort. The real benefit isn’t explicitly in this novelty application, but rather in the potential for wide-scale, affordable, distributed carbon capture/sequestration. As an aside, Aircapture is headed up by Matt Atwood (Vienna, Jul 2013), who previously founded Algae Systems, which attempted to make a go of producing oil-replacing liquid biofuels from algae.

There’s a new place to store greenhouse gases: In your beer

­”In the course of the 1920s and 1930s, great progress was made in the study of the intermediary reactions by which sugar is anaerobically fermented to lactic acid or to ethanol and carbon dioxide.”—Hans Adolf Krebs

*This email is sent from, but doesn’t originate with, Substack; it’s written by me, Nancy Kleinrock. I invite you to write to me at nancy.kleinrock@nemertes.com with your thoughts—positive, negative, or otherwise—about this posting or whatever’s on your mind. If someone forwarded it to you and you like what you see, feel free to subscribe by clicking the button below.

Subscribe now

On March 18, members of the Nemertes community will hear from John Miranda how he turned around a bug-ridden software operation to one that was error free for 140 days—and then for another 140 days after that. How’d he do it? By persistently leading his dedicated team toward structured excellence. If you’re already a member, register to hear John’s inspiring story. If you’re not, you can join using the button at the bottom of this post.

In the meantime, I hope spring is busting out all over where you live, whether that means a profusion of wildflowers, the first days of heat and humidity, or, like where I live, the steady drip, drip, drip of snow melting from the roof as temperatures begin to creep above freezing. I’ll satisfy myself with that until I get to experience flowers and T-shirt weather, too.

And, now for some news . . .

Nancy Kleinrock*

Content Director, Nemertes [Next]

Superhydrophobic materials stay dry, with water beading up and simply rolling away. However, when that water is hot, such materials fail the never-wet test—until now. Recognizing that the secret to non-wetting are nano- and micro-scale surface features that trap air that keep water from settling, researchers at Rice University have underlain a conventional superhydrophobic spray coating with an insulated layer to create a multilayered insulated hydrophobic (MISH) coating that repels even near-boiling water—or split pea soup, as depicted—by not letting it condense in the first place. Application areas include food processing, desalination, chemical manufacturing, and medical sterilization workflows.

A heatshield for “never-wet” surfaces: Rice engineering team repels even near-boiling water with low-cost, scalable coating

­As you may have noticed, the price of oil has spiked in the past week. However, Finnish researchers have developed alternative, plant-based materials for use in applications historically dominated by fossil fuels, notably resins and epoxies. Using forestry and agricultural waste streams like sawdust and straw, they have created a polyester resin that has 76% higher tensile strength than the fossil-based resin it is set to replace, plus it has the side benefit of recyclability. (Tian Li, Washington, D.C., Sep 2018; Volker Hartkopf, Washington, D.C., May 2007)

Forest-based resins challenge fossil materials in wind turbines, boats and high-performance adhesives

Sticking with resins for the moment, Japanese researchers have harnessed sideways electrostatic forces applied to ferroelectric fluids to build motors that do not rely on magnetic fields—nor on magnets and coils. The outcome can be resin-based motors that are lightweight and responsive, which could prove beneficial for robots, compact devices, and precision systems, as well as medical equipment and data storage devices where magnetic noise is rampant. But it’s the final quote by researcher Suzushi Nishimura that really grabbed me: “Interestingly, this force had been theoretically predicted more than 100 years ago, yet no one had ever observed it directly with the naked eye. Becoming the first to witness it was a truly exciting moment. That is one of the great joys of being a researcher. Science is fun.” And isn’t that what drew us all in in the first place?

How an overlooked electrostatic force could drive the motor of the future

Sure, solar panels kick out electricity when the sun is shining, but performance degrades substantially on rainy days (and peters out altogether when the panels are buried in snow). Spanish researchers are getting around this with a perovskite solar panel overlain with plasma-deposited thin film that exploits the kinetic energy impact of raindrops, converting that triboelectrically to electrical energy, as well—up to 110 V/raindrop, or enough to power a small portable or IoT device (Pete Warden, Washington, D.C., Sep 2018), although the linked article doesn’t address the power of a single snowflake. As an additional benefit, the 100-nm-thick coating protects the perovskite material from environmental degradation.

Hybrid perovskite device generates electricity from sun and rain simultaneously

Speaking of small portable devices, The Red Bulletin, a magazine published by Red Bull, came out with a limited-edition issue devoted to gaming, GamePop. Smack dab on its cover is a playable version of Tetris. It was the creative output of Kevin Bates, who a decade ago invented the business card-sized open-source handheld Arduboy on which players could enjoy Tetris and a host of other games contributed by other developers. As such, the magazine cover is just the “latest evolution of Bates’ mission to use existing, accessible, and affordable technologies to reimagine what a portable gaming device can be.” See the linked article for the specs of the durable and responsive GamePop cover as well as a wide—and notably tall—array of other Tetris modalities. (Gabe Zichermann, Chicago, May 2011)

This magazine plays Tetris—here’s how

But not every game enthusiast is fixated on Tetris. For instance, Sammy Azdoufal wanted to have some fun with his DJI robotic vacuum cleaner by manipulating it with a video game controller. Using an AI coding assistant to help him reverse-engineer how his Romo vacuum interfaced with DJI’s cloud servers, he inadvertently found that the credentials for his own machine also granted him access to “live camera feeds, microphone audio, maps, and status data from nearly 7,000 other vacuums across 24 countries.” Oops. That’s what happens when IoT devices aren’t built with security as a first-class element from the ground up, as Bruce Schneier previously emphasized (Washington, D.C., Sep 2018). And note that Schneier will be presenting his thoughts on how AI will affect democracy during the upcoming Nemertes [Next] conference.

Man accidentally gains control of 7,000 robot vacuums

­”Looking back at the past 100 years of history, there is not an industry around that ever increased safety and security without being forced to. The market rewards doing a poor job and taking your chances. If we expect better, we have to demand it.”—Bruce Schneier

*This email is sent from, but doesn’t originate with, Substack; it’s written by me, Nancy Kleinrock. I invite you to write to me at nancy.kleinrock@nemertes.com with your thoughts—positive, negative, or otherwise—about this posting or whatever’s on your mind. If someone forwarded it to you and you like what you see, feel free to subscribe by clicking the button below.

Subscribe now

The lineup for the Spring 2026 Nemertes [Next] virtual conference is shaping up. We’re excited for this event which will take place across four Wednesdays in April and May (5:00–6:30 pm Eastern). Here are the themes:

• Innovation—past, present, and future

• Fixing Democracy—with democracy in crisis, are technology and better frameworks the answer?

• Advances in Biomedicine—what bodies do and the source of new components

• Space and Energy—here’s where the virtual field trip comes in!

If you’re already a member of the Nemertes community, you can find and discuss details there, and if you’re not, see the Nemertes website to see which categories our esteemed speakers—Rachel Higham, Aditi Juneja, Hiromitsu Nakauchi, David Pogue, and Bruce Schneier—fall into.

For more information, and to request an invite, click the button.

More Info

And, now for some news . . .

Nancy Kleinrock*

Content Director, Nemertes [Next]

Researchers at the University of Southern California are following up on work we had first-person access to well over a decade ago during a field trip and conference presentation when Dave Hughes (London, Jul 2014) shared the detailed, no-touch visual unwrapping of mummies and other ancient artifacts at the British Museum. The USC team took advantage of the hundreds of 0.5-mm-thick cross-sectional images of CT-scanned mummies by 3-D-printing life-size reproductions of their skeletal structures. Among the new findings is that Nes-Mor, a priest in the Temple of Man who died in 190 BCE, likely suffered from lower back pain and, rather amazingly for the time, had undergone trephination-like back surgery.

What do you get when you put a mummy through a CT scan?

For those wondering what might happen when chatbots hang out together at a virtual cocktail party, thanks to the viral phenomenon Moltbook, we no longer need to wonder (mostly). Within one week of launching, over 1.7M LLM-powered OpenClaw agents had convened, posted over a quarter-million posts, and left some 8.5M comments. Keeping in mind that the entire origin story of a chatbot is to mimic human conversations, it is perhaps no great surprise that this gathering produced no end of spam and crypto scams, as well as a new religion. Also not surprising is that many of the most quoteworthy lines were posted by humans posing as bots, or at least were the output of careful prompt artistry. As Georgetown University professor puts it, “It’s basically a spectator sport, like fantasy football, but for language models.”

Moltbook was peak AI theater

Likely entirely unrelatedly, there is an activist movement afoot to spur an exodus from the ChatGPT Plus paid service as a political response to contributions by OpenAI’s president to the U.S. President’s super PAC and ICE’s use of the chatbot in its hiring process. If this act of resistance takes hold, it will hardly be the first time that the leading edge of digital technology will influence political change; case in point is the Arab Spring (Mohamed Nanabhay, Sacha Meinrath, Joshua Levine, and Peter Ackerman, Washington, D.C., May 2012)

A “QuitGPT campaign is urging people to cancel their ChatGPT subscriptions

Meanwhile, others are in tears due to the abrupt access removal of OpenAI’s 4o model, with this article digging into the emotional bonds that a perhaps-alarming fraction of people are developing with their chatbots.

Why GPT-4o’s sudden shutdown left people grieving

Switching gears, while the next article doesn’t address the physico-anatomical challenges of sex in space, it does share the same skepticism that Zach Weinersmith (Nemertes Book Club, Feb 2025) conveyed when discussing human readiness for off-planet reproduction.

We’re nowhere near ready to make babies in space, experts warn

”Weightlessness is ideal for physics problems but not for intercourse.”—Kelly and Zach Weinersmith

*This email is sent from, but doesn’t originate with, Substack; it’s written by me, Nancy Kleinrock. I invite you to write to me at nancy.kleinrock@nemertes.com with your thoughts—positive, negative, or otherwise—about this posting or whatever’s on your mind. If someone forwarded it to you and you like what you see, feel free to subscribe by clicking the button below.

Upgrade to paid

Nemertes Live had a wonderful and inspiring kick-off session a week or so ago, when Prof. Len Kleinrock showed us around UCLA’s “Birthplace of the Internet” museum, told stories of that storied day in 1969 when the ARPANET came to life, and traveled bespoke tangents based on participants’ questions in a way that can only happen when you’re, well, Live. Len might have had a hoarse voice that day, but—always intrepid—he was as fully engaged as ever.

As you likely saw from Johna earlier this week, next up (Feb 18) is cybersecurity prognosticator Matthew Rosenquist, who is set to share his latest top-ten predictions list, with a 2026 focus on AI and how it will step up the game for both attackers and defenders.

And with these, 2026 is just getting started: March’s Live event features John Miranda sharing successful processes and lessons for achieving zero-bug software releases, version after version after version. You won’t want to miss him discuss what it took to create this culture of excellence.

Then mark your calendar for the spring 2026 Nemertes [Next] virtual conference, Wednesdays Apr 22–May 13. I’ll have more to share soon on the esteemed lineup coming together, with topics spanning the likes of innovation, AI, space, advanced energy technology, biomedicine, and even the future of democracy. [Next], like Live, has to be experienced in real time, so see the bottom of this newsletter to register now as a Nemertes community member, if you’re not one already.

And, now for some news . . .

Nancy Kleinrock*

Content Director, Nemertes [Next]

The Titanic was engineered to be unsinkable, what with its watertight compartments and remotely activated doors. Unfortunately, no one told the iceberg that. (Sort of like the abundant deer on my little mountain that don’t bother to read the planting instructions associated with all of our deer-resistant trees, shrubs, and flowers—but more consequential.) Well, researchers at the University of Rochester’s Institute of Optics have etched micro- and nano-pits into the interior of aluminum tubes, rendering them truly unsinkable: These surface irregularities trap air bubbles, rendering the tubes superhydrophobic, hence buoyant, regardless of how long they might have been forcibly submerged. Rafts constructed by linking together multiple tubes could serve as the basis for ships, buoys, other floating platforms, or even wave energy harvesting structures (Brian Haus, virtual conference, Jun 2020; David Timmons, virtual conference, Sep 2020).

Scientists engineer unsinkable metal tubes

­Amyloids—fibrous protein aggregates—have gotten a bad reputation as being instrumental in dreaded neurodegenerative diseases like Alzheimer’s, Parkinson’s, and Huntington’s. But new research from the Stowers Institute for Medical Research shows that functional amyloids appear to be essential for long-term memory formation.

Research reveals how the brain turns experience into memory—with help from a tiny protein

Chinese researchers have developed a chip in the form factor of a thin, flexible fiber, making it ideally suited for incorporation into wearables and textiles (Anouk Wipprecht, Atlanta, Feb, 2014). This “fiber integrated circuit” provides the missing link, as it were, for conforming devices, providing computation to blend with existing sensing, display, and power functionality (Michael McAlpine, Washington, D.C., Sep 2018). Instead of fabricating the chips on a rigid surface, circuits are laid down on an ultrathin plastic sheet, which is then rolled up, maki sushi-style. In prototypes, this process yields densities of 100K transistors per centimeter and is robust to thousands of bends, rubs, twists, and stretches, or even to being driven over by a semi truck.

How sushi rolls inspired a flexible fiber chip as thin as a human hair

Ricky Sethi at Fitchburg State University has developed a mathematical framework to enable generative AI systems to engage in metacognition—to assess and act critically when confidently generating conflicting or inaccurate outputs. He defines five dimensions of machine self-awareness—emotional awareness, correctness evaluation, experience matching, conflict detection, and problem importance—and quantifies each of these into a metacognitive state vector that controls ensembles of LLMs. Sethi acknowledges that this framework does not confer human-like self-awareness, but instead provides an architecture for redirecting computational resources to produce improved responses to prompts. (Hod Lipson, Brooklyn, Jul 2016; virtual conference, Oct 2025)

Artificial metacognition: Giving an AI the ability to “think” about its “thinking”

Here’s a new form of job recruitment: an autonomous drone-racing contest, where entrants/applicants submit an algorithm to pilot the drone. The top performers win the opportunity to bypass most of the conventional job application process at Palmer Luckey’s Anduril. (Eric Chang and Mike Hawley, San Francisco, Dec 2014)

Anduril has invented a wild new drone-flying contest where jobs are the prize

Well, a lot has been going on on the domestic political front in the United States in the past couple of weeks, but I’m going to focus on something positive: If Congress passes the bulk of its spending bills before the end-of-the-month deadline, the Trump administration’s push to decimate spending on scientific research will be largely repelled; instead of a cut of more than one-third, funds will only slip by 4% overall, and basic research funding will even increase by 2%.

Congress is rejecting Trump’s steep budget cuts to science

­”Invest in basic research and recruit the best minds.”—Ahmed Zewail (Nobel Laureate in Chemistry, 1999)

*This email is sent from, but doesn’t originate with, Substack; it’s written by me, Nancy Kleinrock. I invite you to write to me at nancy.kleinrock@nemertes.com with your thoughts—positive, negative, or otherwise—about this posting or whatever’s on your mind. If someone forwarded it to you and you like what you see, feel free to subscribe by clicking the button below.

Subscribe now

My mother-in-law grew up in rural Indiana. This is a photo of the phone from her childhood home, which is now mounted in our living room, albeit no longer connected to a human switchboard operator. Oh, and it’s heavy as all get-out due to the solenoid that still resides within the box.

Not too many years after her telecommunications equipment progressed into a more modern era, another family member—my dad, Leonard Kleinrock—was hard at work launching the Internet (actually, the ARPANET, as the early packet-switched wide-area network was known), facilitated by a BBN Interface Message Processor at each end of the communications channel between UCLA and the Stanford Research Institute. On October 29, 1969, the UCLA team successfully logged into the SRI computer, although not until the second try.

On January 21, 5:00 pm Eastern, to relive this seminal act in telecommunications history, we’ll join UCLA Professor and Nemertes Advisory Board member Len Kleinrock for a live virtual tour of 4320 Boelter Hall, where that very first communication commenced. Len will share the history, personalities, and sense of place, including the still-solid physicality of IMP #1, which resides in the “Birthplace of the Internet.”

To be part of this interactive Nemertes Live experience, join the community by clicking the link at the bottom of this email.

And, now for some news . . .

Nancy Kleinrock*

Content Director, Nemertes [Next]

In its early years, the Internet was a closed community in which the people involved knew and trusted one another. You might have heard that this is no longer the case. In fact, the truth is increasingly subject to question, as the combination of AI, social media, and contrasting motivations makes it ever so easy to widely disseminate falsehoods as reality (Chris Piehota, virtual conference, Jun 2020). Fortunately, actual journalists still seek the truth (Gordon Crovitz, Montreal, Jun 2023), as laid out in this self-assessment by the New York Times of how it chose to portray the photo shared by President Trump of Nicolás Maduro shortly after his forced removal from Venezuela—and how it chose to deal with other ersatz photos then circulating online.

How the Times assessed that photo from Trump of Maduro in Handcuffs

­NPR, also focusing on elevating facts over spin, recaps the events that unfolded at the U.S. Capitol on January 6, 2021.

A visual archive of Jan. 6, 2021, through the lenses of those who were there

The “AI scientist” launched by Edison Scientific might not generate actual scientific output in the blink of an eye—the pace at which ChatGPT outputs all manner of answers to simple prompts, accurate or entirely otherwise—but, prompted with a research objective, in 12 hours its Kosmos agent can offer insights akin to six months worth of a postdoc’s work output with a similar level of correctness. The biomedically focused agent is built atop a variety of LLMs (OpenAI, Google, Anthropic) and writes some 42K lines of code and reads some 1500 research papers per prompt, justifying the initial promotional price per prompt of $200.

Where is all the A.I.-driven scientific progress?

It’s January, when many people decide to treat their bodies better by shying away from alcohol and drugs and striving to exercise more. But that doesn’t mean your favorite chatbot has to join you on your quest for improved physical health. A Swedish creative director recently launched Pharmaicy, where he wrote batches of code based on how people respond to psychoactive substances that can be fed to ChatGPT to induce it to simulate being high. For some tens of dollars, anyone interested can send their chatbot on a trip and “unlock your AI’s creative mind.” (Bill Wetsel, Chapel Hill, Mar 2024)

People are paying to get their chatbots high on “drugs”

When John Suh (Pittsburgh, Jun 2019) shared the vision—and physical prototypes—of the Hyundai Elevate platform, the impression was that it would serve as the basis for a range of people-moving vehicles. Now the auto manufacturer has released MobED (Mobile Eccentric Droid), which shares many mobility characteristics of the Elevate, but on a smaller scale, given that it is intended as a controller-operated robotic base for delivery vehicles, golf carts, single-person transports, and the like, depending on the module type mounted atop it.

Hyundai Motor honored with CES 2026 Best of Innovation Award for MobED droid

Also announced at CES is a humongous MicroLED TV by Hisense; it has a 163” screen. It features a yellow subpixel adjoining each RBG unit for improved color fidelity. Personally, I prefer my TV small and the focus in my living room to be the view out the windows of the snow accumulating, the fox frolicking, and the pre-sunset alpenglow or vibrant sunrise (like this morning’s, above (no computational photography applied (Mark Levoy, Charlotte, Dec 2010))) on the mountains across the valley—but that’s just me.

Hisense unveils industry-first RBGY MicroLED display at CES 2026

­”What if a car had legs and could walk?.”—John Suh

*This email is sent from, but doesn’t originate with, Substack; it’s written by me, Nancy Kleinrock. I invite you to write to me at nancy.kleinrock@nemertes.com with your thoughts—positive, negative, or otherwise—about this posting or whatever’s on your mind. If someone forwarded it to you and you like what you see, feel free to subscribe by clicking the button below.

Subscribe now

Let me take the opportunity to wish you a warm and nurturing holiday season and happy new year, given that the Nemertes [Next] newsletter will be taking a break until January.

And, speaking of January, on the 21st we’ll be going on a virtual field trip to the Birthplace of the Internet in UCLA’s Boelter Hall with Len Kleinrock, who will show us the actual equipment used on October 29, 1969, for the initial communication between the first two nodes of the Internet’s precursor, ARPANET, and tell us stories about the day, the people (some of whom will also be in the Zoom room), and the early networking project as a whole.

To be part of this intimate tour, you have to be a member of the Nemertes community. You can join through by clicking through at the bottom of this post.

And, now for some news . . .

Nancy Kleinrock*

Content Director, Nemertes [Next]

LLMs have become multimodal, capable of working with not only text but also images, video, and audio (speech and music). Researchers at Concordia University have focused on speech, improving the processing of so-called audio tokens, which serve as the individual units of spoken language. Because speech carries information about meaning through, yes, words, but also accent, identity, emotion, and so forth, through tone of voice, pacing, and manner of speaking, the bitrate of audio tokens is high. To make speech more easily digestible by LLMs, the research team created a new audio tokenization methodology dubbed FocalCodec that retains both sound and meaning of speech at an ultralow bitrate, such that the FocalCodec output serves as an improved input modality for multimodal LLMs even as human listeners detect no difference between it and the corresponding original speech. (Haizhou Li and See Swee Lan, Singapore, Jul 2009)

Concordia researchers develop novel approach to help Large Language Models learn from speech

­Speaking of AI doing more with less—or at least spreading out the computational load—MIT researchers have developed the DisCIPL (Distributional Constraints by Inference Programming with Language Models) framework to steer LLMs to efficiently delegate subtasks of a big ask to appropriately selected, specialized small language models before bringing results back together for synthesis. The goal is to take advantage of parallelisms by inducing models to collectively produce high-quality outcomes using reduced computational and energy resources. (Burton Smith, Santa Monica, Dec 2007)

Enabling small language models to solve complex reasoning tasks

Telescopes are increasingly bespoke, designed to capture specific types of electromagnetic signals from afar (Nick Law, field trip, Chapel Hill, Mar 2024). This is particularly true for radio telescopes (Dennis Wingo, San Francisco, Dec 2014), and Cornell researchers are making a big splash with a small radio telescope design. The goal of the telescope—the Galactic Radio Explorer (GReX)—composed of eight stations distributed around the surface of the planet, is to capture data from fast radio bursts, a currently poorly understood astrophysical phenomenon. The constellation of receivers will be able to survey the entire sky (Marla Geha, Charlotte, Dec 2010); the form factor? A trio of concentric cake pans—really. This geometry minimizes sensitivity to human-made signals while focusing on radio waves arriving vertically through the atmosphere. Plus, the unit cost of the GReX is a mere pittance.

Cake-pan telescope searches sky for fast radio bursts

As you prepare for a new year of equities trading, you might wish to keep in mind the revalidation of the square root law. Kyoto University researchers analyzed all the Tokyo Stock Exchange transactions between 2012 and 2019—order submissions, trades, and cancellations—to establish a universal square root relationship between the change in a stock’s price and the number of shares traded. Previous studies have relied on size-constrained or proprietary datasets, but the scope of this work suggests that the square root law has broad applicability. In their work, the research team applies all manner of analogies from physics, making for a fun read.

The universal law behind market price swings

Korean researchers are using liquid metal composite ink to absorb, modulate, and shield electromagnetic rays. That is, an invisibility cloak. The ink’s liquid metal particles mesh with one another to create a flexible metamaterial that can be tuned to hide the cloaked object from, say, radar or communication waves, because of selective absorption. The tuning, in fact, is done through the act of physically deforming the easy-to-manufacture cloaking material. (Bogdan Popa, Miami, Dec 2011)

Harry Potter-style “moving invisibility cloak” technology developed

­”I don’t know why people are so keen to put the details of their private life in public; they forget that invisibility is a superpower.”—Banksy

*This email is sent from, but doesn’t originate with, Substack; it’s written by me, Nancy Kleinrock. I invite you to write to me at nancy.kleinrock@nemertes.com with your thoughts—positive, negative, or otherwise—about this posting or whatever’s on your mind. If someone forwarded it to you and you like what you see, feel free to subscribe by clicking the button below.

Subscribe now

This week marked the second time this fall that the Nemertes community heard from Juan Moreno on the topic of quantum technologies, this time focusing on quantum key distribution (QKD)—a topic first broached at our forum in 1998 by Simon Phoenix, a pioneer in the quantum theory of light over fiber who was then of BT Research Labs. While the physics remains the same, the terminology is now bandied about with near abandon and companies are jumping into the space left and right. Nevertheless, those with robust and reliable products remain scant.

During the process of digging into the nuances of the QKD landscape, Juan laid out the basic principles surrounding the tools that Alice and Bob would be using today, the protocols underpinning the tools, the standards those protocols are designed to meet, plus, of course, how Alice and Bob would know if Eve was up to no good. A significant challenge today is knowing what firms—and parts of the world, for that matter—are making solid progress and which should be treated with a gain of salt when you hear that a viable product is just around the corner.

As he develops a bespoke QKD deployability index to evaluate claims based on performance, readiness, and predictability intended for use by QKD stakeholders and decision makers, including vendors, customers, and researchers, Juan has come to recognize that the United States has a lot of catching up to do to meet the progress of China (at the top of the heap), other Asian nations, and Europe, as well.

We expect to hear more from Juan in 2026, but quantum key distribution, quantum cryptography, and quantum computing more generally will only compose a small fraction of Nemertes [Next] content for next year, since we aim to address “what’s [Next]” in all its forms. We want to learn the themes that you’d like us to explore in upcoming events and are inviting Nemertes community members to join us for the virtual holiday party Dec 17, 5 pm EST, to toss around ideas.

If you’re not yet a member and want to be, use the link at the bottom of this email, and you are always welcome to drop us a note through the Nemertes contact page.

And, now for some news . . .

Nancy Kleinrock*

Content Director, Nemertes [Next]

Uh oh. Sam Kriss, writing in the New York Times, just reminded me that AI chatbots love—love, love!—using em-dashes. Me, I find them useful, largely as big commas, and since I like commas, too, em-dashes come to the rescue. But that doesn’t make me an AI chatbot, and I think the rest of my writing style diverges so strongly from AI’s well-documented verbal tics, there should be no confusion—at all.

Why does A.I. write like … that? (and, yes, the article’s title has spaces around its ellipsis)

­A team of Japanese researchers have developed an optical wireless power transmission system reliant on an LED light source—rather than a laser—making it not only longer lived but also more safe, reliable, cost effective, and controllable when implemented as an IoT power source. To overcome the challenges of over-the-air energy loss over long distances and in inconsistent lighting environments, the team created a two-part lensing system to handle the former and a dual-mode operation in indoor environments that are brightly lit versus dim to deal with the latter. The system is designed to sequentially power multiple IoT devices within an indoor space, such as a smart factory or home. (Mustafa Boyvat, Los Angeles, Mar 2018; Meredith Perry, San Francisco, Dec 2017)

AI-powered LED system delivers stable wireless power for indoor IoT devices

Oftentimes scientific research entails building intricate machinery and running experiments, such as a stellarator and tokamak to study fusion (Dylan Schmeling, virtual field trip, Oct 2025). Other times, though, careful metaanalysis of previously collected data generates new discoveries. Such was the case when a team of researchers delved (uh oh! (see AI writing style article, above)) into conveniently stratified large-population medical data to tease out an auxiliary benefit of being vaccinated against shingles.

The medical excitement surrounds the marked decrease in subsequent incidence in dementia for vaccinated vs. unvaccinated individuals in the same cohort and general geographic location, but what really got me was the clever use of the natural experiment just waiting in plain sight to be exploited.

On September 2, 2013, people living in Wales first became eligible for the HZ vaccine if they were 70–79 years old, but not if they were 80 or older. The selective study group consisted of individuals who turned 80 one week before through one week after Sep 2, 2013; that is, half were eligible to receive the vaccine (during the subsequent year) and half were not, but otherwise there was no identifiable difference between the groups. (This actually created three groups: those who were 79 and chose to be vaccinated, those who were 79 and chose not to avail themselves of the opportunity, and those who had just turned 80 and had no say in the matter.) The key finding (quantified in the article and associated paper) is this: “Live-attenuated HZ vaccination prevents or delays mild cognitive impairment and dementia and slows the disease course among those already living with dementia.”

Public service announcement: In the United States, the CDC recommends that people ages 50 and older receive the Shingrix vaccine along with younger adults with weakened immune systems.

For those living with dementia, new study suggests shingles vaccine could slow the disease

In the meantime, President Trump has yet another idea for solving scientific problems and has released an executive order launching the Genesis Mission to use AI to supercharge scientific discovery in the United States. A noble endeavor, but coming from a president who has slashed the science budgets of many federal agencies, yanked research funding from universities, and chosen not to bestow the National Medal of Science nor the National Medal of Technology and Innovation to anyone during either of his terms in office, forgive me for hearing this announcement with some skepticism. I’m not the only one. AI has potential for adding efficiencies and novel suggestions for solving hard problems, but that doesn’t mean that the scientific endeavor should be given over to AIs wholesale.

The Genesis Mission: Here’s what’s in Trump’s most grandiose executive order yet

Genesis Mission: Why Trump’s plan to put AIs in charge of science could backfire

Among problems being tackled in an across-the-board way, albeit by the private sector, is coding. AWS just announced its suite of frontier agents capable of developing complicated software with an initial prompt but with no inherent need for intermediary followups. According to Matt Garman’s re:Invent conference keynote, the agents can work autonomously for hours or even days without human intervention. The tools fall into three categories: the self-explanatory AWS Security Agent and the AWS DevOps Agent, along with the Kiro autonomous agent. The latter is a virtual developer that “retains context across coding sessions and learns from an organization’s pull requests, code reviews, and technical discussions,” and can be connected to GitHub, Jira, Slack, and a firm’s internal documentation systems. Then it just chugs along until completion. What could possibly go wrong? So, humans to the rescue, assuming AI work product oversight is part of their job description: To safeguard against agents going inadvertently rogue, the tools log all learnings for people to review, with the option of removing any errant learnings.

Amazon’s new AI can code without human help. What does that mean for software engineers?

On the other hand, and getting back to the image at the top of this newsletter, researchers at Arizona State University (field trip, Scottsdale, Dec 2023) have demonstrated that LLM agents behave remarkably like humans in friend-forming simulations. Multiple LLMs were placed in a network and tasked with choosing which others to connect with based on attributes like number of connections, common neighbors, and shared attributes (e.g., “hobbies,” “location,” and “workplace”). The agent-formed networks exhibited characteristics of human and engineering networks, such as networks of Facebook friends or telecommunication networks. Specifically, the networks of LLM agents prioritized homophily (connecting to others with similar attributes), triadic closure (shared common connections), and preferential attachment (connecting to highly connected nodes). Given the similarities, synthetic networks of LLM agents could serve as stand-ins for humans in network simulation studies to generate synthetic data. (Clay Shirky, San Jose, Feb 2003; Boston, Sep 2007; Raisa D’Souza, London, Jul 2010)

LLMs choose friends and colleagues like people

­”When we change the way we communicate, we change society.”—Clay Shirky

*This email is sent from, but doesn’t originate with, Substack; it’s written by me, Nancy Kleinrock. I invite you to write to me at nancy.kleinrock@nemertes.com with your thoughts—positive, negative, or otherwise—about this posting or whatever’s on your mind. If someone forwarded it to you and you like what you see, feel free to subscribe by clicking the button below.

Subscribe now

The final two weeks of the Fall 2025 Nemertes [Next] Live conference focused, respectively, on successful lessons, strategies, and frameworks for innovating in large organizations with Justin Fanelli and Larry Fitzpatrick and the use of clever data collection, analysis, and the application of available technologies to address societal-scale problems with Charlotte Matthews and Maurice Pitesky. The various speakers were engaging and attendee participation was high, probing, and insightful. That’s the norm for Nemertes forums.

I could tell you about the technical outcome-driven metrics that underpin the Innovation Adoption Kit that Justin detailed, and I could relate the stories that Larry detailed en route to developing three core principles and a mechanism for successful business turnarounds (inculcate accountability, commit responsibly, offer salient transparency, and construct learning flywheels).

And perhaps I could share Charlotte’s improved assessment methodology for capturing salient characteristics of high-performance buildings (compared with imperfect though widely deployed EnergyStar and Energy Use Intensity), and I could tell of the inadequately confronted threats to poultry and bovine farming, not to mention human health, from migratory waterfowl infected with highly pathogenic avian influenza and Maurice’s approaches to limiting the risk.

But, truly, that’s not the point. The actual point of Nemertes [Next] Live is the live part: You just have to be there (on Zoom) to fully benefit. It’s not only the information, but the human interactions, the connections among people and ideas, and the sometimes-out-of-left-field questions, responses, and insights that percolate through each session and the event as a whole that makes Live what it is.

This fall’s conference might have concluded, but Nemertes Live lives on, with Juan Moreno sharing his next installment of quantum acumen, when he presents a “Deep Dive into Quantum Key Distribution” on Wednesday, December 3. This will be followed on December 17 with a lively discussion of desired topics and themes for 2026 and beyond in the context of a Nemertes [Next] Virtual Holiday Party. And, reaching into the beginning of 2026, on January 21 we’ll be treated to a virtual field trip to the “Birthplace of the Internet,” where Len Kleinrock (one of the “fathers of the Internet” and the only one I personally have) will provide his account of what transpired on Oct 29,1969, when the first ARPANET nodes successfully connected (but almost didn’t) and show us around the room where it happened.

If you’re not yet a member of the Nemertes community, use the link at the bottom of this email and join us.

And, now for some news . . .

Nancy Kleinrock*

Content Director, Nemertes [Next]

Remaining on the quantum theme, there is no shortage of announcements of the latest quantum breakthroughs. Not only is Juan Moreno (Baltimore, Sep 2024; Nemertes Live, Dec 2025 and Sep 2025) playing his part in grounding the QKD corner of the quantum landscape in facts, but DARPA has an entire initiative—the Quantum Benchmarking Initiative—to keep the field honest by evaluating the economic and technical feasibility of aspiring companies’ quantum computing approaches. Program manager Joe Altpeter purposefully brings a healthy dose of skepticism when judging each firm. In 2024’s Stage A round,18 candidate firms competed to move on. To date, two—Microsoft and PsiQuantum (Prineha Narang, Boston, Mar 2022)—have proceeded to Stage C, where they will build their hardware and have it evaluated by the DARPA team; more firms are expected to join them in time. The goal of this funnel for government assessment and support is for the United States to be in the global mix and deliver at least one competitive fault-tolerant quantum computer.

DARPA’s QBI cuts through the quantum hype

­This past spring we heard about the importance of quantum-influenced timekeeping from David Mitlyng (Nashville, Apr 2025) and previously from Tanya Ramond (virtual conference, Sep 2021), both of Xairos, when it comes to developing a secure and accurate alternative to GPS for position, navigation, and—importantly—timing. A team of researchers from the UK and Austria spend their energies explicitly exploring quantum clocks and have built one that comprises a pair of quantum dots and an electron hopping back and forth between them, with a unit of time defined by each hop. This is all well and good, but they observed that the energy required to measure the number of elapsed hops far exceeded the energy of running the clock per se. Specifically, the quantum clock in theory is reversible and in practice is nearly so, but the measurement process involving applied voltages generates appreciable heat and entropy—six orders of magnitude more entropy than generated by the clock itself. In a nutshell, this experiment reveals the measurement conundrum of quantum mechanics: You’ve gotta do it, but it changes everything.

(As an aside, every article I found about this work used an hourglass as a graphic. I have one of those and more generally am a fan of clocks. Pictured above is my Ithaca Calendar Clock (inside and out; clockworks on top and calendar works on the bottom), which turned 150 years old last year. It reasonably accurately tracks not only time of day, but also day of the week and month, and accounts for leap years (but not daylight savings time, since that wasn’t yet a thing). But I can feel the cost of its efforts with each wind of the key: my physical energy that dissipates over the course of the following eight days, week by week, year in and year out.)

The costs of quantum timekeeping

You don’t need me to tell you that the charge is on to ramp up the volume of AI data center infrastructure. In fact, odds are that your local community is debating issuing permits for new facilities right now. Moreover, I don’t use the term volume metaphorically, since these are buildings with both huge footprints and energy and water requirements. But they can be made more resource-efficient, as Cornell researchers have detailed in a new metaanalysis that outlines strategies and geographic locations to reduce carbon dioxide emissions by as much as 73% and water use by 86% (compared with worst-case scenarios). Siting new AI data centers where clean energy is the norm and natural cooling opportunities are abundant is key to keeping the consequences of the buildout in check (Maxie Reynolds, Los Angeles, Jun 2022).

“Roadmap” shows the environmental impact of AI data center boom

And as more AI data centers come online, they compete with other electricity users, which increases the burden on grid operators to balance the load and keep the lights on for everyone while also minimizing costs and ensuring safety. The challenge is exacerbated as distributed generation sources increase in number and type. MIT researchers have developed an optimization tool they dub FSNet that builds in both system constraints (e.g., voltage levels; required uptime) and a feasibility requirement, such that the solutions are inherently implementable. The two-step problem-solving framework explicitly tests for the feasibility of each solution its machine-learning model devises and only retains—and then iterates upon—those options that don’t violate any constraints. FSNet has demonstrated its speed—several-fold faster than conventional problem solvers—while guaranteeing success. The team anticipates its use beyond electrical grid optimization, suggesting it could also be applied to complicated problems like product design, investment portfolio management, and production optimization to meet consumer demand.

A faster problem-solving tool that guarantees feasibility

Researchers at North Carolina State University have developed a functional defense against cryptanalytical attacks to extract model parameters from neural-network based AI models. The success of this brand of attack relies on marked distinctions between neurons in adjacent layers of the network. As such, the defense is achieved by training the network such that neurons differ little from layer to layer; moreover, successful similarity is achievable even when applying the approach to only the first-to-second layer of the network or with only a subset of neurons. The cost is the accuracy of the resultant model, but the team was able to provide a robust defense to parameter extraction for a mere one-percent diminishment.

Researchers unveil first-ever defense against cryptanalytic attacks on AI

Researchers at Weill Cornell Medicine have engineered a novel biomolecular tool for regulating gene activity within cells. Acyclovir-controlled poison exon—dubbed Cyclone—is a versatile gene-switch tool that is nontoxic (acyclovir is an inexpensive and common antiviral medication) and can be engineered into any gene to reversibly activate or deactivate the inhibitory effect of the poison exon when it comes to translation of a transcribed gene by mRNA to synthesize a protein. Moreover, Cyclone doesn’t only operate as an on–off switch, but is tunable by acyclovir dosage, thus enabling precise control of the expression of therapeutic genes and expanding the potential applications of synthetic biology and precision medicine. (William Haseltine, Phoenix, Dec 2008; George Church, virtual conference, Dec 2020; regional meeting, Boston, Jun 2025)

Scientists introduce breakthrough gene-switch technology

­”When genomes are the new canvas for the artist, we might be able to radically upgrade the human species and the software of the biology of the human species.”—Jason Silva

*This email is sent from, but doesn’t originate with, Substack; it’s written by me, Nancy Kleinrock. I invite you to write to me at nancy.kleinrock@nemertes.com with your thoughts—positive, negative, or otherwise—about this posting or whatever’s on your mind. If someone forwarded it to you and you like what you see, feel free to subscribe by clicking the button below.

Subscribe now

This past Wednesday was the sixth of eight Nemertes [Next] Live sessions, this time with Melanie Mitchell, Jonathan Taplin, and community participants focusing on AI. Melanie set things in perspective with a rapid-fire reminder of the field’s 70-year history and the oscillating periods of optimism and pessimism.

No, we didn’t have sci-fi-worthy robots in the 1970s as Claude Shannon had predicted in 1961; no, we didn’t have AGI in the mid-1980s as Herb Simon had anticipated in 1965; and, no, expert systems didn’t fully replace knowledge workers.

There is no doubt that in the past decade or so, deep learning and its offshoot generative AI—trained on the whole of the Internet and then tamed through reinforcement learning with human feedback—are proving whizzes at pattern matching and performance on key benchmarks, but are they “intelligent”? Are they sentient? Do they have a sense of self? No, no, and no, according to not only Melanie—who itemizes problems with the benchmark methodology and therefore results—but also machine learning pioneer Yann LeCun, who has said that a system trained on language alone will never approximate human intelligence, as evidenced by AIs failing in ways that humans don’t, which Andrej Karpathy has described as “jagged intelligence.” Of course, LeCun’s comment doesn’t address what happens when AIs also learn from the real world (aka robots) contributing to the collective knowledge.

Looking toward the future, Melanie has a mix of hopes and concerns, including: a revolution in science and medicine, but harm to all levels of education; freedom from tedious and dangerous jobs, but a disruption of employment more generally; and expanding creativity in the arts, but a growing trust in AI to do tasks outside its wheelhouse.

Jonathan also gave a bit of a history lesson, viewing AI through the lens of the our current existence in a digital interregnum, where we are plunging ahead into an ill-formed future based on a disruptive technological revolution. As with previous periods of disruption—1830–1865, characterized by slavery and Manifest Destiny, and the Civil War; 1890–1930, the Gilded Age, characterized by oligarchy, monopoly, and urbanization; and 1989, the hopeful time when the Cold War ended, coinciding with Tim Berners Lee proposing open-source hypertext as the basis for the World Wide Web, built upon the similarly not patent-protected protocols of the Internet.

Yet 9-11 coupled with technology enabled a surveillance state and the culture of nihilism that dished up Tony Soprano and eventually led to the rise of Donald Trump. During each interregnum; a ruling elite was trying to hold on to power and the corruption and looting it enabled; a technological revolution was disrupting then-current norms, jobs, and ways of life; economic inequality worsened and political violence rose—this is again where we find ourselves today, with AI as the underpinning technology of disruption. Jonathan sees us poised at a turning point with the volume of AI slop on track to dominate human art.

With his formative years and early career in the arts entwined with the 1960s/1970s counterculture music and cinema, Jonathan is particularly sensitive to artists receiving a raw deal at the hand of oversized corporate interests—something that has risen to extraordinary scope in today’s “AI shell game,” where Nvidia, OpenAI, Microsoft, et al., are intricately entwined and becoming ever more so.

Although a modern-day counterculture has yet to arise, Jonathan anticipates eventual youth revolt and civil unrest as their recent/expensive college degrees are rendered moot as AI replaces a plurality of white collar jobs in the coming decade, as many predict. Might the “solution” be a universal basic income intended to placate the masses—but people without aspirations for their futures? Alternatively, the AI bubble could burst and the economy crash, similarly wiping out their hopes and dreams. (In either case, the tech billionaires—or trillionaires!—have personal havens in safer countries to live out their lives in luxury.) And in either case, idle people will have a choice—individually and collectively—to either seek out a sense of purpose or to rise up in a counterculture movement. Jonathan believes that someone has to stand up to the protoauthoritarian state; in the past it has been the artists, but not today, which he posits is linked to the commercialization of art over the course of his professional lifetime and its acceleration in the age of generative AI. His upcoming book, “Where Is the Counterculture,” will address these matters.

As I noted, this was the sixth of eight Nemertes [Next] Live sessions this fall. You won’t want to miss the remainder: First, on November 12, we focus on making headway in large organizations with Larry Fitzpatrick, CTO of One Main Financial, speaking on “Avoiding Failure as a Pathway to Success”; and Justin Fanelli, Adjunct Professor at Georgetown University, speaking on “Accelerating the Flywheel of Innovation Adoption” by drawing on his deep expertise moving innovation forward within a truly large enterprise: the U.S. Navy.

To be part of the conversation, become a member of the Nemertes community—if you aren’t already—by clicking the button at the bottom of this email and join in the discussion.

And, now for some news . . .

Nancy Kleinrock*

Content Director, Nemertes [Next]

Looking forward to November 19 and our final session of Nemertes [Next] Live Fall 2025, we’ll be hearing from two speakers, each addressing societal-scale problems. One of them is UC Davis professor of veterinary medicine and cofounder of AgriNerds Maurice Pitesky. His specialty is highly pathogenic avian influenza—bird flu, in its various forms. There hasn’t been a lot of chatter in the mainstream media since the price of eggs came down months ago. That doesn’t mean the threat of H5N1 is gone. Far from it. This article reporting on research done at Cornell’s College of Veterinary Medicine draws attention to the considerable pathogenic potential of raw milk products. After all, H5N1 infects cows in addition to fowl. Fortunately, pasteurization kills the virus, but even the aging process involved in cheesemaking lowers the risk, particularly for acidic cheeses like feta.

Bird flu virus surviving in certain cheeses

Since the beginning of Donald Trump’s second term as President, a focus of the administration has been to whittle down the federal workforce and slash agency budgets. The Cybersecurity and Infrastructure Security Agency (CISA) has not been immune. I’ll leave it to you to decide whether it is by design or an unfortunate side effect, but among the many casualties of CISA cuts is election security (Rich DeMillo, Seattle, Mar 2020), since among the job cuts are 14 positions dedicated to helping state and local officials “manage ransomware threats, disinformation, and potential [election] interference.” This opens the door for nation-state actors to fire up their AI chatbots to infect the digital landscape with intentional lies, deep fakes, and other social-engineering shenanigans, with an eye toward influencing the U.S. electorate—or elections themselves. This article offers the actionable suggestion of proactively taking a prevention-first stance, rather than hiding heads in the sand or being merely reactive. (MITRE Cybersecurity Workshop, Sep 2016; Thomas Rosario, Johns Hopkins Applied Physics Lab, field trip, Sep 2015)

How the F5 breach, CISA job cuts, and a government shutdown are eroding U.S. cyber readiness

Okay, moving on to more uplifting topics, here are some items that fall into the category of cool science and engineering:

Not all governmental undertakings are being cut. The Department of Energy announced $625M over ten years to five National Quantum Information Science Centers—variously at Argonne, Lawrence Berkeley, Oak Ridge, Brookhaven, and Fermi national labs—to move the science, engineering, and workforce opportunities of U.S. quantum computing forward. Also, community members should mark December 3, 5 pm EST, on their calendars for a Nemertes Live session with Juan Moreno, Nemertes [Next] Advisory Board Member and Quantum Enthusiast (Nemertes Live, Aug 2025; Baltimore, Sep 2024). His focus that Wednesday is quantum key distribution.

DOE announces $625M to advance National Quantum Information Science Research Centers

Japan has just brought online its first (and the world’s second, after one in Denmark) osmotic power plant, which generates electricity by mixing freshwater and salt water (Gary Atkinson, London, Jul 2014). This renewable energy source is not dependent on the whims of the weather, but rather generates electricity by taking advantage of the difference in salinity across a membrane: as freshwater crosses the membrane in an (admittedly vain) effort to equilibrate salt concentrations, pressure builds up on the salty side and is used to drive a turbine. The science is hardly new; instead, it’s been the engineering challenges of energy loss due to pumping and membrane-crossing friction that have delayed scale-up. With the inherent benefit of consistent energy output, the researchers involved believe that osmotic energy could satisfy as much as 15% of global energy demand within 15 years.

Japan just switched on Asia’s first osmotic power plant, which runs 24/7 on nothing but fresh water and seawater

While on the subject of round-the-clock renewable energy, satellite startup Reflect Orbital seeks to make photovoltaics productive even at night. How? By launching a megaconstellation of reflective satellites into low-Earth orbit and concentrating sunlight onto targeted nighttime locales, such as existing solar power plants. The result: “sunlight on demand.” The firm’s first 18-meter test satellite is due to launch next year, with an ultimate goal of 4000 satellites—each 54-m across—by 2030, collectively delivering 200 W/m^2. The founder’s fever dream entails 250K such satellites. Challenges include the huge cost of launching such an array, the light pollution it would exert, public safety concerns associated with the concentrated rays, ecological disruption to circadian rhythms, animal migration, and agricultural crop growth, not to mention the multiplicative increase in space junk. Then again, they could be used to light up your summer evening garden party. (Matt Kozlov and Techstars Founders, Los Angeles, Jun 2022)

The true cost of “solar power at night” with Reflect Orbital

Pixels are getting smaller as researchers at the University of Würzburg in Germany use optical antennas to circumvent the limitation of the wavelengths of visible light on pixel size. Since visible light is roughly 400–700 nm (violet->red), these folks created pixels just 300x300 nm by “selectively passivating nanoelectrode edges with an insulating layer while simultaneously defining a nanoaperture in flat areas.” They see this approach as a scalable strategy toward subwavelength miniaturization for use in photonic integrated circuits and near-eye displays (Barmak Heshmat, Scottsdale, Dec 2022; Allan Evans, San Francisco, Dec 2014; Jeri Ellsworth, San Francisco, Dec 2013).

Physicists built a pixel so small, full HD could fit on a bread crumb

­”You see that pale, blue dot? That’s us. Everything that has ever happened in all of human history, has happened on that pixel.”—Al Gore

*This email is sent from, but doesn’t originate with, Substack; it’s written by me, Nancy Kleinrock. I invite you to write to me at nancy.kleinrock@nemertes.com with your thoughts—positive, negative, or otherwise—about this posting or whatever’s on your mind. If someone forwarded it to you and you like what you see, feel free to subscribe by clicking the button below.

The most recent session, with Emory’s Gordon Berman and Michael Ghil of UCLA and the Ecole Normale Supérieure in Paris, delved into patterns in flux and matters of complexity, causality, and variability. Gordon’s work centers on quantifying causality in complex dynamical systems, ones that inherently experience change—systems like the economy, the climate, or developing human beings. His approach relies on a metric of relative predictability to train a neural net one time only using data (both actual and surrogate) over all times that can then be applied to causal relationships going forward.

Michael, with his very long history in climate change, wrote his dissertation when the fear was impending deep freeze rather than a hothouse Earth. Regardless, the more complex the climate model—and the greater number of complex climate models there are—the greater the uncertainty of what the future holds. “Global weirding” is the new norm. In this session, Michael put forth a new, relatively simple climate model that matches historical data closely and, among many factors, explicitly incorporates albedo and the visible fact—if one makes the effort to look—that polar and alpine ice sheets are not white, but rather gray due to glacial algal blooms. As such, glaciers no longer exert their full albedo potential to moderate temperature. It might take thousands of years for ice sheets to melt but only one month for algal blooms to become robust. This model predicts the same bistability he wrote about in the 1970s, a feature that holds the potential for algae spurring the planet toward an overheated tipping point.

Well, that’s a lot, but Nemertes [Next] Live is still rolling along for three more Wednesdays. The topic the first week of November is AI, with professor Melanie Mitchell of the Santa Fe Institute and author, director emeritus of USC’s Annenberg Innovation Lab, and notable contrarian Jonathan Taplin offering realistic evaluations and sometimes alarming views of AI and its potential. Melanie will address how “intelligent” artificial intelligence is and is likely to become, while Jonathan’s focus is on the effects of AI on the social fabric. The alarm raised by Hod last week serves as an appetizer to the November 5 session with Melanie and Jonathan.

To be part of the conversation, become a member of the Nemertes community—if you aren’t already—by clicking the button at the bottom of this email.

And, now for some news . . .

Nancy Kleinrock*

Content Director, Nemertes [Next]

As discussed above, we heard from Michael Ghil about glacial algae’s potential large-scale influence on the climate, but what about the changing climate’s influence on microorganisms? Mycorrhizal fungi, depicted here, enter into symbiotic relationships with trees and other plants by assisting them in gathering water and nutrients from the soil and, in turn, receive a significant percentage of their carbon needs from those flora. How much carbon—and how that is likely to change as carbon dioxide concentrations and temperatures rise—is a topic of study by scientists at McGill University. When using data from 1800 forest sites in the eastern U.S., the team replicated earlier work showing that trees in regions with significant fertilizer runoff, and therefore higher soil nitrogen concentrations, require less fungal assistance; the trees reciprocate with lower carbon contributions. What was new is that rising soil temperatures similarly decreased the helpful exchanges. Balancing that is increased atmospheric CO2, which stimulates greater plant growth, such that those plants will need more help from fungi for soil-based nutrients. So, as Ghil and Gordon Berman both emphasized during Nemertes [Next] Live this past Wednesday, real-world systems are complicated and highly variable.

How plant–fungi friendships may change in the face of warming soil and rising CO2 levels

­Looking for yet additional factors to incorporate into climate models? New modelling by Cornell researchers indicate that peatlands could suddenly shift from being carbon sinks to carbon sources. In a future climate of warmer temperatures and greater atmospheric carbon dioxide, the likelihood of droughts increases by an averaged factor of 4.5, and drought conditions could “wipe out hundreds of years of accumulated carbon” in a matter of months. Their data was collected during a 2021 drought in a northern Minnesota natural spruce bog.

Peatlands’ “huge reservoir” of carbon at risk of release

Korean researchers have built a generative AI model that combines physics simulation rooted in the material point method with 2D-to-3D reconstruction using Gaussian splatting, and trained it to learn physical laws from visual cues in videos. Their initial application is the movement of clothing worn by avatars moving within a virtual scene, along with realistic handling of collisions between the avatars and their garments and other objects in the scene. Lead researcher Tae-Kyun Kim describes the MPMAvatar model thusly: “This technology goes beyond AI simply drawing a picture; it makes the AI understand ‘why’ the world in front of it looks the way it does. This research demonstrates the potential of ‘Physical AI’ that understands and predicts physical laws, marking an important turning point toward AGI.” This conclusion, mirrors that of Hod Lipson from last week’s Nemertes [Next] Live session.

AI model accurately renders garment motions of avatars

Given the high volume of spent lithium–ion batteries, recycling them and recapturing the lithium as well as cathodic heavy metals—nickel, manganese, and cobalt—within them is a timely challenge. Researchers at Boston College, including Babak Momeni (Boston, Mar 2022), are turning to the microbial world to get the job done. The team developed a modified biohydrometallurgical setup that sidesteps the usual dependence on external chemical fuels like ferrous sulphate to stimulate the Acidithiobacillus ferrooxidans bacterium to do what its name suggests. Metallic iron or stainless steel already present in the battery casing serves to fuel the autotrophic bacterium following initial acidification. The outcome? Nearly perfect leaching efficiency of all four target metals. (Peter Girguis, Salt Lake City, Dec 2009)

Feeding off spent battery waste, a novel bacterium signals a new method for self-sufficient battery recycling

MIT researchers have improved upon the optical resolution enabled by the 2014 Chemistry Nobel Prize-worthy innovation of super-resolution microscopy. That historic work pushed light-based molecular positional resolution down to 10 nanometers; the new discrete grid imaging technique—DIGIT—allows for true atomic resolution of <0.2 angstroms using light with only the addition of knowledge of the materials crystal structure—a 500-fold improvement! This should help guide the design of quantum devices at an atom-by-atom level. Experimental samples were fabricated at MIT.nano (Will Oliver, MIT field trip, Mar 2023).

A “seating chart” for atoms helps locate their positions in materials

­”Nothing exists but atoms and empty space; everything else is opinion.”—Democritus

*This email is sent from, but doesn’t originate with, Substack; it’s written by me, Nancy Kleinrock. I invite you to write to me at nancy.kleinrock@nemertes.com with your thoughts—positive, negative, or otherwise—about this posting or whatever’s on your mind. If someone forwarded it to you and you like what you see, feel free to subscribe by clicking the button below.

Hod will be taking us inside his robotics lab, which is predicated on the biological concepts of self-organization, replication, and evolution. In fact, his long-standing vision is to build a machine that can build a robot that walks out of that machine, steps into the world, and subsequently reproduces. Steps along the way include self-learning foosball-playing robots, a tidying-up robot, a robot that anticipates and mimics human facial expressions, a 3-D food printer, an exploration of robot metabolism, and even the world’s fastest soft robotic fish. Hod will explain how each of these functions and how each serves his larger goal. You won’t want to miss it—and you don’t have to. Subscribe to become a Nemertes member and then register to attend the virtual Nemertes [Next] Live session Wednesday, October 15, 5:00 pm EDT. You can find the relevant button at the bottom of this email.

And, now for some news . . .

Nancy Kleinrock*

Content Director, Nemertes [Next]

In a couple of weeks—October 22, to be precise—Nemertes [Next] Live will be in conversation with Gordon Berman speaking about the value of incorporating dynamics into computations of cause and effect (Joshua Vogelstein, Nemertes Live, Jan 2025) as well as from Michael Ghil speaking about how climate variability contributes to the climate change’s “global weirding.” The article here addresses one aspect of this weirding: cross-species breeding between animals resulting in novel hybrids such as the grue jay (center, above), a genetic blend of the blue jay (left) and the tropical green jay (right). The grue’s green mom must have made her home considerably north of the species’ normal range. (Note that, as brightly colored as all of these jays might be, none of them—blue, green, nor grue—are olo (Austin Roorda, Nemertes [Next] Live, Oct 2025).) Such hybridization is likely to become more common as climate change drives animals together into previously unshared habitats. Ghil will be speaking about the physics behind climate variability and implications reaching far beyond cross-breeding.

The pizzly bears and grue jays of the future

­As I’m sure you’ve been hearing, this is Nobel week. Only mildly overshadowed by the behemoth of prestigious awards is the list of MacArthur genius grant recipients. This week, Nabarun Dasgupta (Chapel Hill, Mar 2024) joined Keivan Stassun (Nashville, Apr 2025) in holding this honor, who was a 2024 MacArthur fellow. Recall, University of North Carolina’s Dasgupta has been working tirelessly to reduce harm from opioids through projects to widely distribute naloxone to reverse the effects of overdose and to analyze samples of street drugs for nefarious additives to help individuals make informed decisions on what to put into their bodies while also alerting medical responders to the hazards circulating in their local geographies. He thus combines community engagement with the scientific practice of epidemiology. As Nab put it when hearing of this latest honor, “I don’t think the universe could send a clearer signal that we should keep going in the direction that we’re going in.”

Nabarun Dasgupta: Epidemiologist and Harm Reduction Advocate

Not to discount Nobel-worthy science: This year’s Nobel Prize in Chemistry went to a trio of researchers—Omar Yaghi, Jeremy Feldblyum, and Adam Matzger—for their work on metal-organic frameworks, solid structural chemical materials featuring spaces just waiting to house all manner of small molecules to perform specialized tasks, ranging from the separation of rare and critical materials to efficient and selective drug delivery (Susanna Thon, Baltimore, Sep 2024).

Q&A: Exploring metal-organic frameworks (MOFs) with chemist

Just as humans benefit from bouncing ideas off of one another, and just as outcomes improve when humans and AIs collaborate with one another, so too do AIs perform better when they work together. New work out of Johns Hopkins University entailed the formation of a “council of AI agents”—specifically, instances of OpenAI’s GPT-4 LLM-based chatbot—overseen by a facilitator algorithm instructing the council to reconsider any divergent answers to U.S. Medical Licensing Examination questions that they might generate. By iteratively exploring the reasoning that each council member put forward and iterating until reaching consensus, this approach yielded scores well north of 90% on the exam, including correcting more than half of incorrect answers generated by mere majority vote. Thus, collective decision making triumphed, even when the AI received no special training on medical data, albeit in an environment where incorrect answers were overtly flagged. This iterative approach “allows the system to take a few tries, compare notes, and self-correct, and it should be built into future tools for education and, where appropriate, clinical care.”

Collaborative AI passes U.S. medical exams

That is but one example of AI’s medical utility. Another is Swedish work being conducted at the KTH Royal Institute of Technology, where researchers are applying machine learning to biomarkers in the blood to establish pan-disease analysis of the human blood proteome (Melissa Lechner, San Diego, Feb 2015). The training set included longitudinally collected blood samples from members of the study group of some 8200 patients—folks who suffered from at least one of 59 diseases. The outcome is statistically relevant profiles for the progression of a diseased individual over time. In particular, childhood infectious diseases and lung cancer could both be picked out of a series of blood proteome analyses to provide an early diagnosis, but this approach is proving only confirmational for ovarian and colorectal cancers.

A pan-disease proteome could improve disease biomarker detection

A dementia diagnosis is never desired, but, as they say, knowledge is power. In particular, patients and their families can prepare for the future and also get clearance to participate in the most effective treatments available. For instance, the highly sensitive blood analysis coming out of research at the University of York can differentiate between the presence of amyloid and phosphorylated tau, indicating Alzheimer’s, and alpha-synuclein, indicating Parkinson’s. This work involving an ultrasensitive optical sensor should enter clinical trials soon.

Breakthrough in blood test for Alzheimer’s by University of York researchers

­”The first person to survive Alzheimer’s is out there now.”—The Alzheimer’s Association

*This email is sent from, but doesn’t originate with, Substack; it’s written by me, Nancy Kleinrock. I invite you to write to me at nancy.kleinrock@nemertes.com with your thoughts—positive, negative, or otherwise—about this posting or whatever’s on your mind. If someone forwarded it to you and you like what you see, feel free to subscribe by clicking the button below.

The upcoming [Next] Live event, on October 1 (5:00–6:30 pm Eastern), features two speakers and will similarly alter perceptions—in this case literally, since the overarching theme will be “Design and Perception; and Designing Perception.” Founding chair of Department of Design Tech within Cornell’s College of Architecture, Art, and Planning, Jenny Sabin, speaks to “A Dynamically Responsive Built Environment: The Interplay of Design and Science” by sharing her lab’s philosophy of design and a range of her artistic, architectural, and inspired projects. Jenny is followed by University of Waterloo’s Austin Roorda speaking about expanding limits of perception, in particular the color “olo,” generated by selectively exciting retinal M-cones, something that doesn’t occur in nature, but that he has personally experienced. His talk is “Perceiving Something Entirely New: Olo.” As is the norm for [Next] Live, the final half hour of the event is an open conversation stimulated by the pair of speakers.

To participate, you should be a Nemertes member (paid subscriber). Click here or the button below to become a member (if you aren’t already!)

Become a member!

And then continue . . .

. . . for some news.

Nancy Kleinrock*

Content Director, Nemertes [Next]

Work underway at NYU’s Tandon School of Engineering (field trip, Brooklyn June 2018) uses colloidal quantum dots to bypass the need for the conventionally toxic processes involving heavy metals when manufacturing infrared cameras (Dror Sharon, San Francisco, Dec 2014; Greg Dobler, Austin, Feb 2026)—a product category increasingly important in autonomous vehicles, medical imaging, and national security. Instead of the tedious, costly, and environmentally unfriendly process of individually placing atoms across a detector’s pixels, solution-based colloidal quantum dots are synthesized in solution and then applied uniformly using conventional roll-to-roll coating processes. The outcome is ultrasensitive IR detection at the microsecond timescale.

An eco-friendly way to see in the dark using colloidal quantum dots

­Staying on the subject of quantum dots, combining an old standby with new-fangled addition can lead to accentuated success. In this case, the goal is to combat drug-resistant bacteria, the old standby is simple vinegar—i.e., acetic acid, a natural antibacterial agent—and the newcomer is cobalt-containing carbon quantum dot nanoparticles. A research team of Australians and Norwegian scientists have demonstrated nontoxicity to humans and microbial lethality to antibiotic-resistant staph and E. coli.

Nanoparticles supercharge vinegar’s old-fashioned wound healing power

Sticking with Australian researchers, and refocusing on matters of eco-friendliness, work at the Royal Melbourne Institute of Technology has yielded a surprisingly strong building material simply using cardboard and dirt to entirely sidestep the need for concrete when constructing the foundations of buildings. Assuming appropriate-quality soil at the building site, a construction team would only need to bring with them cardboard tubes and use on-site dirt to create rammed-earth support structures from which to build the foundation. But not just any old cardboard tubes: The relevant innovation in this low-tech construction process is the precise specifications of the dimensions and aspect ratio for the tubes. An important side benefit, of course, is to keep waste cardboard out of landfills. This work follows a higher tech—and more expensive and less ambiently available—earth-rammed-into-carbon-fiber-tubes support material that the same researcher developed and showed to have comparable strength to high-performance concrete. (Heidi Kujawa, San Francisco, Dec 2016)

Cardboard and earth reshape sustainable construction

This same RMIT researcher, who heads up the university’s Centre for Innovative Structures and Materials, doesn’t restrict himself to dry land, although this next project also focuses on material strength. With biomimetic inspiration from the Pacific Ocean sea sponge Venus’ flower basket that fashions its intricate basket-like skeleton out of silica, his lab developed a strong new auxetic metamaterial. “While most materials get thinner when stretched or fatter when squashed, like rubber, auxetics do the opposite,” says Jiaming Ma. “Auxetics can absorb and distribute energy effectively, making them extremely useful.” He anticipates applications in construction—notably to enhance earthquake safety—as well as protective equipment, sports gear, and medical applications. (Julia Greer, San Jose, Dec 2021; San Francisco, Dec 2015). Stay tuned for more bio-inspired architectural design on Wednesday with Jenny Sabin.

New material inspired by a sea spong’s glass skeleton

You don’t need me to tell you that we live in a multiply connected world, but here’s one facet that I hadn’t previously considered: the link between a warming climate and ambient noise in the vicinity of airports. (Maybe that’s because I don’t live directly in a flight path—although, ironically, a loud military jet is flying overhead as I type this.) Warmer air is less dense, leading an airplane to take off with as much as a 7.5% reduced angle of climb according to the British researchers in the University of Reading’s Turbulence Research Group. The upshot will be a broader area around airports where aircraft noise will regularly rise above 50dB, subjecting neighbors to noise pollution. Oh, and this is in addition to the group’s other recent findings that a warming climate increases flight turbulence and airport flooding. Something to look forward to. (Robert Carter, Memphis, Sep 2006; field trip to FedEx, Memphis, Sep 2006)

Airports will get noisier as Earth gets warmer

­”Seems like half of my anxiety dreams are about airports.”—Lois McMaster Bujold

*This email is sent from, but doesn’t originate with, Substack; it’s written by me, Nancy Kleinrock. I invite you to write to me at nancy.kleinrock@nemertes.com with your thoughts—positive, negative, or otherwise—about this posting or whatever’s on your mind. If someone forwarded it to you and you like what you see, feel free to subscribe by clicking the button below.

Oh, and although it took over a week, I finally found not only my mug, but also my delightful North Country Public Radio pint glass—the one etched with an owl.

In any case, I expect I will have found my favorite coffee mug by the time September 17, 5 pm EDT, rolls around; that’s when the [Next] series kicks off with University of Pennsylvania’s Konrad Körding engaging Nemertes members in a conversation on evolution and innovation, constraints and deconstraints, exploring how the most basic of biological principles guide success—or failure—of not only organisms, but also organizations and their pursuits. See this fall’s full agenda, which includes two virtual field trips, and find out how to join the intellectual journey.

And now for some news ...

Nancy Kleinrock

Content Director, Nemertes [Next]

I’ve always had a curiosity about the perception of color, perhaps beginning with the question from my then-young peers about who dreamed in black-and-white and who in color. That was actually a hotly debated topic back when TV came over the airwaves, through an antella, and into a small box in the den or living room, displaying what would now be considered unacceptably grainy images in blacks, whites, and grays. Maybe my friends watched a whole lot more TV than I did, because I clearly recognized that the world was full of color and that my dreams were similarly vivid. Still, I actively wondered whether what I saw as teal or magenta or burgundy was perceived the same way in other people’s brains. German researchers took up this question by placing a small group of study subjects in an MRI machine, showing them different hues, and observing resultant brain activity. They found that brains of people without visual impairments such as red–green color blindness interpret colors reliably the same as one another; the same is true for macaque monkeys. Together this suggests that—evolutionarily—color matters. During the second session of the Fall 2025 Nemertes [Next] Live series on October 1, we’ll explore color perception and color generation with University of Waterloo’s Austin Roorda and Cornell’s Jenny Sabin.

Do you see the same colors that I do?

­Meanwhile, researchers at the University of Oklahoma and University of Arizona have been studying the evolutionary path of color—both perceiving it and biologically generating it. Their finding is that color vision in animals emerged hundreds of millions of years before life—first plants and later animals—used color on their own behalf to variously spread seed, attract pollinators, ward off predators, or attract mates. So in this proverbial chicken-and-egg question, the biological ability to distinguish among hues far preceded color generation.

When nature burst into vivid color

And, why all this focus on color? Because trees around my new home have already started showing theirs. It’s going to be a beautiful fall.

­As we will hear from [Next]’s perennial contrarian Jonathan Taplin (Half Moon Bay, Dec 2023; Los Angeles, Mar 2018; Boston, Apr 2017; Washington, D.C., May 2012; Pasadena, Feb 2002) when he and Santa Fe Institute’s Melanie Mitchell share their concerns about AI during [Next] on November 5, AI is the latest technology to separate creators from their lawful copyright protections. In one instance, at least, book authors are coming out on top: Anthropic has agreed to settle a lawsuit with authors and publishers for illegally downloading and storing their copyrighted works. The $1.5B settlement amounts to $3000 per work to each of 500K authors. With echoes from the Napster era, this case has the potential to shift the balance of power back toward creators. (Peter Eckersley, San Francisco, Feb 2010; Cory Doctorow, London, Jul 2010; Solveig Singleton, Miami, Jul 2005; Mark Ishikawa, Austin, Feb 2004; Larry Lessig, Jul 2002; Jack Valenti, Pasadena, Feb 2002; Manus Cooney, Washington, D.C., Nov 2001; Douglas Adams, San Jose, Sep 2000)

Anthropic agrees to pay $1.5 billion to settle lawsuit with book authors

­If you look closely at the schedule of Nemertes [Next] Live sessions, you’ll notice some that Wednesdays are bereft of sessions. This is not to say that Nemerteans won’t be convening around interesting discussions those evenings (specifically, September 24 and October 29). As the last Wednesday of the month has been all year, they will be devoted to the Nemertes Book Club—the September book is Douglas Copland’s “Microserfs”; October’s choice remains TBD). Back in May, we were joined by coauthor Zach Weinersmith to discuss his book “A City on Mars.” We’re not the only ones to have taken notice of his contrarian view of establishing human settlements off-planet. MIT Technology review has published a review of three books, including the Weinersmiths’, making the case against extraterrestrialism for humans.

Book Review: The case against humans in space

­

"We are not going to be able to operate our Spaceship Earth successfully nor for much longer unless we see it as a whole spaceship and our fate as common. It has to be everybody or nobody.”—R. Buckminster Fuller

­