Very recommendable! Lot's of background!
Saturday, August 01, 2026
Ukraine’s Entry Into the Iran War Changes Everything | Is Caspian Sea a New Warzone?
Iran is threatening to target energy infrastructure in the Middle East
Bad, concerning news! How far will this fanatic, suicidal theocracy/dictatorship and state sponsor of terrorism and nuclear weapons seeking Iran go? World War III?
Where are the coward European powers when America needs help? No minesweepers, no navy escorts of commercial ships etc.!
"Iran is threatening to target energy infrastructure in the Middle East, likely to try to deter the United States and Israel from following through on reported plans to strike Iranian energy infrastructure.
An unspecified senior security official told Islamic Revolutionary Guards Corps (IRGC)-affiliated Tasnim News Agency on July 30 that Iran has prepared an attack plan that includes attacks on regional energy infrastructure in response to reports that US President Donald Trump has ordered new strikes on Iran.
IRGC-affiliated Fars News Agency reiterated this threat and published a list of regional energy facilities that it said Iranian missiles can reach.
The list included oil fields and other energy infrastructure in Israel, Saudi Arabia, Qatar, Kuwait, Bahrain, and the United Arab Emirates.
Iran issued these threats after Western media reported that the United States and Israel are considering conducting large-scale strikes against Iranian energy infrastructure.
An Iranian analyst close to the regime stated on August 1 that an attack on Iranian infrastructure would be a "terrifying scenario for everyone" and urged regional and international actors to prevent the "crisis." Iran likely seeks to prevent US or US-Israeli combined force operations by signaling that any attack on Iran’s energy sector would trigger Iranian retaliatory strikes against regional critical infrastructure."
Trade, Trust, and Civilisation: The Hidden Economic Message of Nolan’s 'Odyssey'
Very recommendable! I blogged here about the movie recently.
"That question lies at the heart of Johan Norberg’s recent book Peak Human, which argues that societies rise not simply because of military strength or exceptional rulers, but because they develop the institutions that make cooperation, innovation, and voluntary exchange possible. Historians once explained the fate of civilisations primarily through battles, conquests, and empires. Increasingly, however, scholarship has shifted its attention towards less dramatic but ultimately more decisive forces: secure property rights, political stability, trust between strangers, economic freedom, and the peaceful exchange of goods and ideas. Armies may conquer kingdoms, but commerce is what allows civilisations to endure."
Note: If correctly described in the article below, then it appears Johan Norberg regurgitates what August Friedrich von Hayek has said many times before in his writings. Maybe the author of this article below is not aware of this.
"Nolan’s emphasis on trade is therefore far from accidental. Archaeology has revealed that the eastern Mediterranean during the Late Bronze Age was home to one of the most sophisticated commercial systems the world had yet seen. Rather than a collection of isolated kingdoms, it formed a remarkably integrated economic network linking Egypt, Anatolia, the Levant, Cyprus, mainland Greece, and beyond. Copper, tin, wine, olive oil, timber, textiles, ivory, glass, spices, and precious metals travelled enormous distances through interconnected markets.
The most remarkable evidence for this commercial world emerged almost by accident in 1982, when a diver discovered the wreck of a merchant ship off the southern coast of modern-day Turkey. The Uluburun shipwreck, dating from the fourteenth century BC, contained more than ten tonnes of Cypriot copper, almost a tonne of tin, African ivory, Baltic amber, Mesopotamian glass, Mycenaean pottery, and gold jewellery. No primitive economy could have organised such a logistical operation. Long before the Industrial Revolution, an extensive international trading system already connected societies separated by thousands of kilometres.
Commerce alone, however, was not enough. These networks depended upon institutions of cooperation. The Amarna Letters reveal rulers addressing one another as “brothers,” exchanging gifts, negotiating alliances, and maintaining permanent diplomatic contacts.
Greek society reflected the same logic through xenia—the sacred obligation to offer hospitality to strangers. In a world where long-distance trade depended on trust between people who had never met, such norms were not merely moral ideals; they were essential economic institutions. ..."
Not all Coloradans want to share their cities with Canada geese and their poop. Really!
What about Arizonans? I can tell from my neighborhood here in the Phoenix metro area that every year we have a large number of geese poop droppings on sidewalks, trails etc.
If you don't know geese poop droppings, they are huge (maybe comparable to dog poop)!
Let me briefly hypothesize that we have generally overpopulation issues with birds in Western urban areas!
Caveat: I did not read the article.
UK Prime Minister Keir Starmer unfortunate legacy, the Employment Rights Act of 2025
Recommendable! How stupid is this new law!
Was Keir Starmer a mediocre PM to put it mildly? I bet he was!
For decades, the government of the UK is heading in the wrong direction! Here is another example!
Once a classical liberal and largely free market economy is more and more becoming socialist since 1945.
Most of the victims of this legislation will be young individuals entering the labor market or inexperienced individuals of any age trying to change jobs etc. These adverse effects have been known for several decades!
"... Burdensome regulations, such as
day one paternity leave,
ordinary parental leave, and
statutory sick pay
might on the face of it look beneficial to workers, but not if they make employers more adverse to hiring new workers, especially young people without previous work experience. Young people who are either working flexible hours or looking for a permanent job don’t want mandated employment benefits if it means we won’t be hired in the first place.
Other regulations, such as workers being given the right to sue their employer for unfair dismissal after 6 months on the job – reduced from 2 years – will only create a more legalistic market, where employers spend more time managing legal risk than creating employment opportunities.
One benefit of having a free labour market is that it allows young and inexperienced workers the chance to gain experience and skills.
If the cost of hiring increases because of new employment rights, then it will be the youngest and least essential workers that will be laid off, or employers will simply hire less people. ..."
The two vices of many Chinese people in China
Spitting and smoking! However, one should always look to one's own faults first before judging other people! 😊
How do I know about it? I have travelled to China a few times in my life.
Too many Chinese people (women included) frequently spit anywhere and everywhere in public. Before spitting, they often make a horrible hawking sound. I also saw a person spitting e.g. into a trash can in a hallway inside a hospital.
Unfortunately, too many Chinese smokers light up their cigarettes and smoke almost anywhere in public especially also in crowded areas (e.g. busy train platform). E.g. I have seen a street food vendor smoking at the hist counter, next to the food for sale but hiding his cigarette whenever a person approached and so on.
Amazon wants to launch 5,105 satellites that would beam voice, data, and emergency service directly to smartphones beyond the reach of cell towers
Good news! More space junk? Just kidding! Competition is good, more competition is better!
"Amazon asked federal regulators for permission to launch 5,105 satellites that would beam voice, data, and emergency service directly to smartphones beyond the reach of cell towers, starting in 2028."
"... 5,100 satellites that would provide direct-to-device (D2D) services using Globalstar spectrum. ..."
"Key takeaways
- Amazon Leo has proposed a new constellation of up to 5,105 satellites to connect compatible mobile devices directly from space.
- Leo will partner with mobile network operators to extend mobile connectivity services globally, with satellite deployment beginning in 2028.
- The Leo network will use advanced signal-processing technologies to make efficient use of spectrum and minimize interference.
..."
How Amazon Leo plans to connect mobile devices from space (original news release) "Amazon Leo has filed with the FCC to deploy up to 5,105 new satellites designed to use Globalstar’s spectrum to bring voice, data, messaging, and emergency services directly to mobile devices."
Google says it fixed more Chrome bugs in June than over the past two years, thanks to AI
Good news! Is this how Google now covers up a buggy software? Just kidding!
With ML & AI software engineering will never be the same again! Rapid, unprecedented progress is to be expected!
I personally can confirm that about over the last two months or so my Google Chrome browser prompted me almost every day to perform an update.
"... The tech giant announced on Thursday [7/30/2026] that it has fixed a whopping 1,072 security bugs in the last two versions of Chrome, both released in June. That is more than the number of bugs patched in the previous 23 versions released over the last two years, which totaled 1,036 fixes. ..."
"We’re living through a massive shift in the software security industry. Large Language Models (LLMs) are unlocking unprecedented capabilities for automated vulnerability discovery, scaling far beyond the limits of human security expertise, and requiring new approaches for staying ahead of attackers. ..."
Stronger with every update: How we’re making Chrome and the web safer in the AI Era (original news release) "How Chrome is using AI to improve vulnerability discovery, triage, and patching."
Wow! Very impressive! Look at "internally found" (red line)
Nachhaltiges Bauen: „Klimaanlagen sollten nur da laufen, wo sie notwendig sind“. Wirklich!
Oder die Demagogie der Nachhaltigkeit! Wird wieder mal deutscher blinder Gehorsam erwartet!
Klimawahn und Umweltreligion (hier Stichwort Nachhaltigkeit) in der Bananenrepublik D!
Christine Lemaitre (Quelle) Hat sie vergessen mit dem Zeigefinger zu drohen? 😊☛
Wind turbine trash Is a problem onshore and offshore
Recommendable! Wind power is not only intermittent and unreliable! Wind turbines are noisy, spoil nature or change local climates!
Renewable energy is a disturbing euphemism!
"... wear of turbines, resulting from high winds, rain, hail, and other weather phenomena that erode the turbine blades and cause them to deposit tons of tiny bits of plastic/resin/polymer waste on the land surrounding land-based facilities in Scotland and Sweden. Leading-edge wear is even worse for offshore wind facilities, polluting the sea and the food chain.
Trash from wind is even more obvious and damaging when the turbines explode or collapse, which is not uncommon, as numerous videos demonstrate. ...
The material from each wind turbine has value. But if recycled, the value is less than 10 percent of the cost of decommissioning land-based turbines and just 3 percent of the cost of decommissioning offshore wind. And because the materials in the blades are of even less value and bound in a fashion that makes them difficult to recycle, wind turbines, blades, towers, and ancillary equipment such as nacelles containing gears are piling up, cluttering vast acres of land in some areas of the country.
Images are easily found that show broken and abandoned wind turbines in Hawaii rusting and falling apart, abandoned for years. In Texas, in lieu of promised proper disposal and recycling, the companies involved in decommissioning wind turbines have found it easier to stockpile the debris on open land, creating an illegal nuisance eyesore and a potential health danger. ..."
Japan finalises joining the EU’s flagship research programme Horizon Europe
Good news! When East meets West and vice versa!
"Twenty-three countries outside the EU have now affiliated to the research funding scheme with Japan being the most populous
Japan has officially associated to Horizon Europe, becoming the 23rd non-EU country to join the EU’s flagship research funding programme. The agreement entered effect on 30 July and means that Japanese researchers can now access parts of the Horizon programme on the same terms as EU researchers. ..."
"... The agreement allows Japanese researchers and organisations to take part in Horizon Europe on equal terms with those from EU Member States and other associated countries. They will be able to join and lead international research consortia, receive EU funding, and work alongside top research institutions to help address global challenges. ..."
Japan officially joins Horizon Europe (original news release)
Horizon Europe "Research and innovation funding programme until 2027. How to get funding, programme structure, missions, European partnerships, news and events."
Why not more online facial age estimation to distinguish children from adults?
I bet the technology could be improved!
As a safeguard, facial age estimation ought to be clearly separated from personal identification or login requirements! This could be achieved by e.g. third party providers who do nothing else but facial age estimation.
There are other means of age verification available too in the digital age. Why is e.g. my Google account not age verified! Google are you paying attention!
I bet privacy concerns (like the article below) are overblown here! Why would a selfie video be necessary? Does not a close up photo of your eyes/face suffice.
For example, porn websites could use it to keep children away, but not adults.
"Online facial age estimation is moderately reliable for broad categories like telling adults from children, but it is imprecise for guessing exact numbers. Top algorithms from groups like the National Institute of Standards and Technology have a mean absolute error rate of about 2 to 3 years, making them much better than average human guesses, but still prone to error based on lighting, skin care, and camera quality." (Google Search)
"... However, this is quickly shifting a broad policy problem into a biometric one, where access to ordinary online services can depend on literally showing your face [???]. ...
While facial age estimation may offer platforms a convenient way to respond to youth online safety demands, that convenience comes with serious privacy costs. Required facial scans lower the baseline of privacy for everyone, especially teens, and create a regime where access to everyday social spaces depends on individuals surrendering sensitive information about their faces and, in many cases, their identities. ...
Meta’s implementation of Yoti, a digital identity and age assurance company, on Instagram and Facebook shows how this works.
If Instagram believes an account holder is under 18 years old and the user tries to change that, it can send that user to Yoti for analysis of a selfie video [???].Yoti then estimates whether the person appears to be over 18 and sends that result back to Meta.
Its tests have expanded across the United Kingdom, Europe, Canada, Australia, and Japan, and similar flows have begun appearing on Facebook in Australia. Yoti also presents facial age estimation as part of a broader age assurance system that can include document-based checks, while highlighting that it estimates age rather than identifying the person. ..."
How did mammals from North America move into South America millions of years ago
Amazing stuff!
"... researchers have found fossil evidence that animals from different latitudes of North America and South America mixed long before the Panamanian land bridge permanently linked the two continents about 3 million years ago.
Since around 10 million years ago, mammals — primarily small ones like sloths and coatis — found a way across the watery barrier. Only larger mammals had to wait for the permanent land corridor. ...
Biologists thought that only after the isthmus formed did animals surge north and south, what biologists refer to as the Great American Biotic Interchange or GABI. That resulted in the extinction of a lot of South American species as North American species took over.
Yet this picture of mammal movements across continents is too simple, ... paleontologists argue. After assembling and analyzing a new database of fossil mammals discovered in Mexico, they conclude that northern North American animals had already migrated into southern North America by about 10 million years ago, where they waited around in a kind of holding pen until opportunities arose for some of them to continue crossing to the south starting about 7 million years ago. ...
During that pre-GABI time, smaller mammals — heralded by racoon-like coatis and ringtails and small rodents — moved to South America via some kind of intermittent corridor, by island hopping through the Caribbean or by rafting on vegetation. Only one major South American group, the sloths, moved to North America. ..."
From the abstract of the Perspective:
"Whenever plants and animals newly come together, there will be some winners and some losers. This was the case for the Great American Biotic Interchange (GABI), which occurred ~2.8 million years ago (Ma), when the emergence of the Isthmus of Panama connected the continents of South and North America. Although species moved both northward and southward during the GABI, a massive taxonomic asymmetry arose between the two continents, with a much greater proportion of species in South America having a North American origin than the opposite. ... Tseng et al. report that a large fossil dataset from Mexico helps to uncover the events that primed mammals for movement across the Isthmus."
From the editor's summary and abstract:
"Editor’s summary
For millions of years after the breakup of the supercontinent Gondwana, South America was disconnected from other landforms, a time period that has been referred to as its “splendid isolation.”
This isolation allowed for the evolution of mammals with distinctive features, including xenarthrans, sprassodont carnivores, and notoungulates.
The formation of the Panamanian land bridge facilitated connection of North and South American faunas, with a bias from north to south. Tseng et al. looked at fossils from low latitudes in México and found that northern mammals accumulated in the region before the connection (see the Perspective by McGuire and Williams). This allowed for more rapid dispersal south and further supports that the land bridge first formed during the Pliocene rather than the Miocene. ...
Abstract
The movement of taxa between North and South America across the Panamanian land bridge, known as the Great American Biotic Interchange (GABI), established modern-day mammal assemblages in the Americas, heralding the decisive end of what has been called the “splendid isolation” of the South American fauna.
Lacking substantive low-latitude records, previous work on this pattern relied on South American and temperate North American fossils to conclude that pulsed interchanges began about 2.8 million years ago (mya).
Based on new low-latitude fossil data from México, we found evidence for a prolonged, triphasic interchange.
Phase 1 consisted of the initial accumulation of northern higher-latitude mammals in a Mexican “holding pen” beginning 10 mya.
Dispersal began in phase 2, with a strong north-south gradient and increasing biogeographic connections (7 to 3 mya).
The final phase consisted of the already well-described taxonomic pulses (from ~2.8 mya)."
When mammals crossed between continents (no public access)
North and South America weren’t always connected, but their mammals still found a way to intermingle (original news release) "Before the Panamanian land bridge formed about 3 million years ago, mammals on the two continents were isolated — or so biologists thought. New fossils show that they found ways to mix long before that."
A 10-million-year biogeographic holding pen primed the Great American Biotic Interchange (no public access) "Before the Panamanian land bridge formed about 3 million years ago, mammals on the two continents were isolated — or so biologists thought. New fossils show that they found ways to mix long before that."
This illustration depicts (left) large North American mammals — counterclockwise from top, mammoths, llamas, tapirs, peccaries and saber-toothed cats — waiting in a kind of holding pen in Mexico for the bridge to form.
In South America, the mammalian interchange involved, counterclockwise from top, ground sloths, marsupials, capybaras, armadillos, glyptodonts and, incidentally, terror birds. While North American mammals took over in South America, only about 10% of North American mammals today are derived from mammals that originated in South America.
English for trippers: Mugged by a mug in muggy weather
Did you take a mug shot? Not your cup of tea? 😊
Researchers break diffraction barrier in super-resolution microscopy
Good news! This could be a breakthrough!
"... have developed a groundbreaking super-resolution imaging technology that allows scientists to visualize molecular structures with sub-angstrom-level localization precision — three orders of magnitude beyond the nanometer limits of standard fluorescent dyes — while drastically simplifying the imaging process.
Unlike traditional dyes that fade rapidly under illumination and limit data collection, the platform, called U-STORM (Upconversion enabled Stochastic Optical Reconstruction Microscopy) utilizes a new class of compositionally engineered upconverting nanoparticles (UCNPs) that blink spontaneously and indefinitely. ..."
From the abstract:
"Single-molecule localization microscopy enables high-resolution biological imaging, but its precision is limited by the rapid photobleaching of conventional fluorophores.
Multicolour imaging is further constrained by the need for spectrally distinct dyes requiring separate excitations or sequential acquisition.
Here we show that small (~10 nm) upconverting nanoparticles can be compositionally tuned to exhibit spontaneous, sustained blinking under single near-infrared excitation without optical or chemical modulation.
By adjusting sensitizer (Yb3+)–emitter (Tm3+/Er3+) ratios, we identify a regime with intrinsic ON–OFF switching and low duty cycles (~0.9%) without photobleaching or statistical aging, enabling repeated localizations and sub-ångström precision (0.62 Å over 88,000 localizations) in upconversion-enabled stochastic optical reconstruction microscopy.
By elucidating the underlying physical mechanism of this blinking, we engineered blue- and red-emitting probes for multicolour upconversion-enabled stochastic optical reconstruction microscopy.
This technique enables the resolution of tightly packed UCNPs and the visualization of epidermal growth factor receptor dimers and multimers on cell membranes at single-protein resolution, all achieved with a simple optical setup without imaging buffers."
Spontaneous and indefinite blinking in upconverting nanoparticles for ångström-precision multicolour super-resolution imaging (open access)
Fig. 1: U-STORM workflow and route to ångström-level precision.
Fig. 3: Super-resolution imaging with high precision and accuracy using blue-emitting U-STORM probe.
Over 21 million Americans used marijuana daily or near-daily in 2025 and other addictions
Bad news! For how long has this been going on? A nation on high! 😊
How many individuals kick their habit eventually? I bet quite a few.
On a positive note: Adolescent use of marijuana seems to be trending downward since 2021. Similar substance use disorders. (see infographics below).
"... This week the Substance Abuse and Mental Health Services Administration released the results from its latest National Survey on Drug Use and Health. These showed that around 21.4 million Americans used marijuana daily or near-daily last year. Compare that to the 19.9 million Americans who smoked daily and the 17.2 million who drank alcohol at least 20 days a month.
Can a country built on whiskey, beer, and tobacco thrive when it becomes a nation fueled by weed? ..."
SAMHSA Releases Annual National Survey on Drug Use and Health (original news release)
2025 National Survey on Drug Use and Health (NSDUH) Releases "Conducted annually, the National Survey on Drug Use and Health (NSDUH) provides nationally representative data on the use of tobacco, alcohol, and illicit drugs; substance use disorders; mental health conditions; suicidal thoughts and behavior; and receipt of substance use and mental health treatment among the civilian, noninstitutionalized population aged 12 or older in the United States. NSDUH estimates allow researchers, clinicians, policymakers, and the general public to better understand and improve the nation’s behavioral health. These reports and detailed tables present national estimates from the 2025 National Survey on Drug Use and Health (NSDUH)."
The source of the charts below.
Bundesland NRW streicht 90% der etwa 4.000 deutschen Baunormen und digitale Bauanträge werden Pflicht
Gute Nachrichten! Es geschehen auch manchmal Wunder in der Bananenrepublik D!
"... Auch bisher galten nie alle 4000 Normen gleichzeitig ...
Für diese Projekte künftig keine Baugenehmigung mehr nötig
Eine Beschleunigung versucht die neue Landesbauordnung zu erreichen, in dem sie zahlreiche Projekte davon befreit überhaupt eine Baugenehmigung einholen zu müssen. Konkret betrifft das:
- Schaffung von Wohnraum im Bestand (Dach-Ausbau)
- Wohnungsbauprojekte mit weniger als 5000 m2 innerhalb eines gültigen Bebauungsplanes und falls die Gemeinde kein Veto einlegt.
- Ladestationen für Elektromobilität, jetzt inklusive der damit verbundenen technische Nebenanlagen.
- Drohnenlandeplätze und Anlagen, die der Landes- und Bündnisverteidigung dienen
- Bauliche Anlagen der Telekommunikation sowie der allgemeinen Versorgung mit Elektrizität, Gas, Öl, Wärme, Wasser, Wasserstoff und Anlagen zur Ver- oder Entsorgung von Kohlenstoffdioxid (wie Trafo- oder Pumpstationen) sind bis zu einer Grundfläche von 20 m² und einer Höhe von 5 m nun verfahrensfrei
- Solaranlagen unter bestimmten planungsrechtlichen Voraussetzungen im Außenbereich
- Landwirtschaftliche Gebäude zum vorübergehenden Schutz von Pflanzen und Tieren bis 4 m Höhe, sowie Kompostanlagen und Fahrsilos.
..."
BauCode NRW: Drittes Update zur Landesbauordnung macht Bauen einfacher, schneller und digitaler (original Pressemitteilung, viel zu viel Text)
Die Heldin: Ina Scharrenbach, Ministerin für Heimat, Kommunales, Bau und Digitalisierung des Landes Nordrhein-Westfalen (Quelle)
RNA editing for stem cell health and longer life
Good news! Behind the secrets of aging!
"Intestinal stem cell function declines with age, and dysregulation of these cells is associated with a shortened lifespan. Zhang et al. found that the RNA-editing enzyme adenosine deaminase acting on RNA (ADAR) restrained intestinal stem cell proliferation in fruit flies, and its abundance decreased with aging and after gut injury.
Loss of ADAR in these cells in young flies caused intestinal hyperproliferation and reduced lifespan, whereas ADAR overexpression in aged flies improved gut function and increased lifespan.
Reductions in colonic ADAR2 were associated with aging in humans and with intestinal regeneration in mice, suggesting that ADAR mediated regulation of intestinal stem cells may be conserved."
From the editor's summary and abstract (a very technical abstract):
"Editor’s summary
Intestinal stem cell (ISC) function declines with age, and ISC dysregulation is associated with shortened life span.
Zhang et al. found that the RNA editing enzyme ADAR restrained ISC proliferation in fruit flies, and its abundance and activity decreased with aging and after gut injury.
ISC-specific loss of ADAR in young flies caused intestinal hyperproliferation and reduced life span, whereas ADAR overexpression in aged flies improved gut function and increased life span.
ADAR inhibited proliferative MAPK signaling in ISCs because its RNA editing activity generated a translational repressor of MAPK pathway components. Reductions in colonic ADAR2 and a protein that stabilizes it were associated with aging in humans and with intestinal regeneration in mice, suggesting that ADAR-mediated regulation of ISCs may be conserved. ...
Abstract
Aging impairs intestinal stem cell (ISC) function, disrupting epithelial homeostasis and regenerative repair.
Loss of ISC quiescence promotes intestinal dysfunction and contributes to organismal aging.
Here, we report an epitranscriptomic mechanism through which a decrease in adenosine-to-inosine (A-to-I) RNA editing by the adenosine deaminase ADAR in ISCs during aging disrupts a conserved signaling axis that maintains ISC quiescence.
In Drosophila melanogaster, ISC-specific loss of dADAR triggered hyperproliferation and tissue aging, whereas dADAR overexpression reduced age-related gut dysfunction and extended life span.
Intestinal injury induced a decrease in dADAR, which was necessary for tissue regeneration, and in the zinc finger protein dZn72D, which stabilized dADAR to support intestinal homeostasis during aging. dADAR edited transcripts encoding the RNA binding protein Pumilio (dPUM) to generate an isoform that inhibited translation of two components of the mitogen-activated protein kinase (MAPK), dEGFR and dERK, thereby maintaining ISC quiescence.
The abundance of ADAR2 and of the dZn72D ortholog ZFR decreased in human colonic crypts with aging and in mouse colon after injury, suggesting that attenuation of ADAR activity may play a conserved role in ISC aging and in facilitating colonic regeneration.
Our work identifies a conserved mechanism of epitranscriptomic regulation that safeguards tissues and the decay of ADAR-mediated RNA editing as a form of age-related epigenetic information loss that disrupts ISC homeostasis."
Fig. 1. A-to-I RNA editing in ISCs declines during aging.
Fig. 2. dADAR inhibits ISC proliferation in a cell-autonomous manner.
Fig. 4. dADAR loss drives gut aging and reduces life span in Drosophila.
Fig. 6. dADAR maintains ISC quiescence by inhibiting the MAPK signaling pathway.
Chinese military researchers tap US AI models to train defense systems
Bad news! Why should AI not be dual use?
Maybe US AI models should train Chinese defense systems how to malfunction in case of a military conflict? 😊
What about the US spying on China how they use US AI models?
"Chinese military researchers have used outputs from leading U.S. artificial intelligence models developed by OpenAI and Anthropic to train domestic AI systems to advance China’s defense capabilities, according to a Reuters review of more than 80 Chinese academic papers and patents. ...
One paper published last year by researchers in PLA Unit 96941, a military intelligence and cyber-warfare unit in Beijing, described using OpenAI’s GPT-3.5 to process sensitive military source code. ..."
"Summary
- Chinese military researchers distil US models, documents show
- Distilling done to create smaller and more secure versions of a model
- Distillation has become a flashpoint in US-China ties as AI competition intensifies
- Distillation cannot replace massive compute needed to develop frontier AI from scratch
Chinese military researchers have used outputs from leading U.S. artificial intelligence models developed by OpenAI and Anthropic to train domestic AI systems to advance China's defence capabilities, according to a Reuters review of more than 80 Chinese academic papers and patents. ..."
"Executive Summary:
- Distillation has become essential to the rise of Chinese artificial intelligence (AI) models. Chinese academic and industry papers published in the period 2024–2026 detail how distillation takes place—including adversarial distillation—and provides insight into the kinds of entities involved in the work of using leading models to build capabilities that rival the world’s best models.
- The prize is frontier reasoning: the step-by-step “chain of thought” that is the most expensive capability to build and the one on which U.S. laboratories lead most clearly. Chinese research literature documents a deliberate, premeditated effort to distill that capability.
- Some published Chinese research also focuses on engineering distillation techniques to evade detection. Much of this work is conducted by researchers affiliated with the People’s Liberation Army, the defense industry, and the Chinese Academy of Sciences.
- Distilled models are either already being used, or are proposed to be used, for public security applications such as surveillance and monitoring systems, and for military and cyber operations.
..."
Spy vs. spy (remember the famous MAD cartoons, source) MAD = mutually assured destruction
Self-Supervised Learning of Structured Dynamics from Videos
This could be an interesting new, but narrowly focused paper by Andrew Zisserman and his team!
From the abstract:
"Understanding motion in video is a fundamental challenge for visual learning, as frame-to-frame change entangles two sources of dynamics:
camera motion and object motion.
This decomposition has remained underexplored in representation learning, partly because these factors are tightly coupled in natural videos and difficult to supervise separately.
Yet recovering it is important for learning robust motion representations that separate meaningful object dynamics from camera-induced variation. We study whether such structured motion representations can be recovered from frozen features of a pretrained image vision transformer.
We propose the Structured Dynamics Model (SDM), which explicitly separates the dominant source of temporal change from residual dynamics through future-feature prediction, rather than representing video change with a single entangled latent or with unstructured, spatially dense transition tokens.
Training combines self-supervised learning on real video with weak supervision of scene dynamics on synthetic Kubric data.
We evaluate SDM on ProbeMotion, a new evaluation suite spanning synthetic and real videos with camera motion, object motion, and combined dynamics.
SDM outperforms backbone baselines using global CLS or average-pooled features, and compares favorably to strongly supervised representations such as VGGT on several probes, despite using substantially weaker supervision.
These results suggest that pretrained image models can be readily repurposed into structured video-dynamics representations, providing a useful inductive bias for learning and analyzing latent video dynamics."
Robostral Navigate
This could be an interesting new paper in robotics by over 140 authors!
This work appears to rely heavily on generating synthetic data and it focuses on indoor navigation.
Here is the official news release (a blog) by Mistral AI: Introducing Robostral Navigate "Summary: Robostral Navigate is an 8B model that enables robots to autonomously navigate complex environments using only a single RGB camera, achieving 76.6% success on unseen R2R-CE benchmarks—outperforming multi-sensor approaches while being more efficient. Built entirely in-house with simulated data and token-efficient techniques, it generalizes across robot types and adapts to real-world obstacles unseen during training. The model combines pointing-based navigation with reinforcement learning for continuous improvement, paving the way for unified embodied AI in robotics."
From the abstract:
"Deploying navigation systems at scale requires a recipe that minimizes sensor assumptions, generalizes across robot embodiments, and trains efficiently.
Yet, today's best systems depend on depth sensors, multi-camera rigs, or pre-built maps, limiting the hardware they support and increasing deployment cost.
We introduce Robostral Navigate, an 8B vision-language model built around this scalability objective. The model consumes only a stream of monocular RGB images - the most ubiquitous sensor across robotic platforms and predicts waypoints by pointing to the next target location in the current camera view.
Operating purely in image space, rather than robot-specific coordinates, makes the policy naturally robust to changes in camera intrinsics and scene scale, enabling deployment across wheeled, legged, and aerial robots without recalibration.
We generate 2.4 million trajectories across 350k simulated scenes to reduce the reliance on real-world data collection and scale easily. We further introduce a prefix-caching training recipe that packs entire episodes into single training sequences, reducing training tokens by 22x and cutting training time from months to days.
A tree-based attention mask prevents conditioning on previous ground-truth actions, encouraging visually grounded action prediction, and reinforcement learning is used to further improve exploration and recovery capabilities.
On the Room-to-Room and Room-Across-Room in Continuous Environments (R2R-CE and RxR-CE) benchmarks, Robostral Navigate sets a new state of the art. On R2R-CE, it achieves a 77.4% success rate, surpassing the best monocular method by 10.5 points and the strongest depth- or multi-camera system by 5.3 points despite using only a single RGB camera. On RxR-CE, it reaches 75.1% success rate, outperforming all monocular baselines."
Attributes Should Come from Images, Not Class Names: Distribution-Conditioned Attribute Selection for Vision-Language Models
This could be an interesting research paper by Deepak Pathak and his team!
From the abstract:
"A popular route to interpretable zero-shot classification asks a large language model (LLM) to describe each class name and prompts CLIP with the resulting descriptors.
We show that these descriptors carry little visual evidence of their own: removing the class name from the prompt collapses ImageNet accuracy from 59.5% to 15.5%.
The diagnosis is that the descriptors are conditioned on the label rather than on the images, so they describe the concept in general and mislead exactly when the data shifts; an LLM insists that strawberries are red, but every strawberry in ImageNet-Sketch is a colorless line drawing.
We therefore select attributes from the target image collection instead: we score a large attribute pool against the images in CLIP's joint embedding space and keep the top-scoring attributes per class.
Selected this way, class-name-free attribute prompts reach 23.8% on ImageNet (against 15.5% for LLM descriptors), the gain holds on four shifted ImageNet variants, and reselecting from the LLM's own pool isolates the selection mechanism as the cause.
With one image per class, the selected attributes outperform the prompt-tuning method CoOp by 3 points while fitting in under a minute instead of 14 hours, with no learned soft prompt to obscure the decision. Because the attribute set is chosen by the data, it doubles as a readable summary of a dataset, which we use to describe distribution shift in words."
Patch Policy: Efficient Embodied Control via Dense Visual Representations
This could be an interesting new paper by Yann LeCun and his team!
From the abstract:
"Pretrained dense visual features from Vision Transformers (ViTs) are powerful yet have been underutilized in robot learning. Modern robot policies either compress each observation into a single global token, or rely on visual backbones trained from scratch, sacrificing both fine-grained spatial detail and the benefits of large-scale visual pre-training.
While there exist policies that do operate on dense patch features like large vision-language-action models (VLAs), they tend to be heavy and slow, inheriting the full cost of a billion-parameter vision-language model (VLM) backbone.
We close this gap with Patch Policy, a minimal architectural extension that enables transformer-based policies to consume dense pre-trained patch tokens directly without the computational overhead of a full VLM.
At its core is a block-causal attention mask that preserves the temporal causality of standard policies while letting the model attend over many patch tokens per observation, alongside other state information. Patch Policy is lightweight, fast, and highly effective.
Across four simulated and three real-world environment suites, our method achieves a 40% relative improvement over policies using state-of-the-art global-pooled representations.
Furthermore, it surpasses fine-tuned OpenVLA-OFT by 18% while using roughly 0.7% of the parameters.
We believe Patch Policy provides a pipeline for the robotics community to readily leverage continuing progress in visual representation learning, without sacrificing the training efficiency or inference speed required for high-frequency, reactive control. Videos can be viewed at this https URL"
Hidden Civilisation Discovered Beneath Amazon's Wilderness
Very recommendable! Amazing stuff!
I already blogged here about this discovery.
Friday, July 31, 2026
Maurice Ravel: Boléro / Petrenko · Berliner Philharmoniker
Enjoy! This is a mild, subdued version of Bolero!
Subscribe to:
Posts (Atom)