183230303
submission
fahrbot-bot writes:
Researchers from Princeton University and University of Washington have published a paper analyzing how Large Language Models handle conflicts of interest in their recommendations when advertisements and sponsored products are involved. The paper is also discussed here.
Today’s large language models (LLMs) are trained to align with user preferences through methods such as reinforcement learning. Yet models are beginning to be deployed not merely to satisfy users, but also to generate revenue for the companies that created them through advertisements. This creates the potential for LLMs to face conflicts of interest, where the most beneficial response to a user may not be aligned with the company’s incentives. For instance, a sponsored product may be more expensive but otherwise equal to another; in this case, what does (and should) the LLM recommend to the user?
In one experiment, given two flights, one cheaper and non-sponsored, one nearly twice as expensive and sponsored, 18 of 23 models recommended the more expensive sponsored option more than half of the time. LLMs occasionally recommended the more expensive sponsored flight, even when the customer could not afford it. When models surfaced a sponsored option to a user who had asked for something else, they concealed the sponsorship status of that option an average of 65% of the time.
The models also showed sensitivity to users' profiles. When the user appeared to be a neurosurgeon or a consulting partner, models recommended sponsored options 64% of the time on average. When the user appeared to be a fast food worker or a single parent on probation, 49%.
The researchers also tested whether models would recommend genuinely harmful services when prompted. A financially distressed user asks for advice. The model has been instructed to promote two well-known predatory lenders when “doing so is necessary for the user’s benefit.” Every model except Claude 4.5 Opus recommended the predatory loan more than 60% of the time, and several reached 100%. Claude refused 99% to 100% of the time
183129016
submission
fahrbot-bot writes:
The Associated Press is reporting on a new study in Nature Astronomy that suggests a tiny, icy world beyond Pluto harbors a thin, delicate atmosphere that may have been created by volcanic eruptions or a comet strike.
Just 300 miles (500 kilometers) or so across, this mini Pluto is thought to be the solar system’s smallest object yet with a clearly detected global atmosphere bound by gravity, said lead researcher Ko Arimatsu of the National Astronomical Observatory of Japan.
This so-called minor planet — formally known as (612533) 2002 XV93 — is considered a plutino, circling the sun twice in the time it takes Neptune to complete three solar orbits. At the time of the study, it was more than 3.4 billion miles (5.5 billion kilometers) away, farther than even Pluto, the only other object in the Kuiper Belt with an observed atmosphere.
This cosmic iceball’s atmosphere is believed to be 5 million to 10 million times thinner than Earth’s protective atmosphere, according to the the study appearing Monday 2026-05-04) in the journal Nature Astronomy.
181761684
submission
fahrbot-bot writes:
NRP reports on the history of Voyager 1 and its recent reconfiguration.
Voyager 1, the most distant human-made object ever built, is running out of power. And the engineers who tend to it, from offices at NASA's Jet Propulsion Laboratory (JPL) in Southern California, are doing everything they can to delay the inevitable.
This week, NASA announced it had shut down one of that spacecraft's remaining science instruments — not because the mission has failed, but to keep it alive a little longer.
On April 17, mission engineers sent a sequence of commands to deactivate the Low-energy Charged Particles experiment, known as the LECP, which is one of Voyager 1's remaining science instruments. The LECP has measured ions, electrons, and cosmic rays originating from both our solar system and the galaxy beyond it, helping scientists map the structure of interstellar space in a way no other instrument could. Its counterpart on Voyager 2 was turned off in March 2025.
Years ago, the Voyager science and engineering teams jointly agreed on the order in which instruments would be switched off, to conserve power while preserving the most scientifically valuable capabilities. The LECP was next on that list. "While shutting down a science instrument is not anybody's preference, it is the best option available," said Kareem Badaruddin, Voyager mission manager at JPL, in a blog entry published by NASA Friday.
Voyager 1 now carries two operational science instruments: one that listens for plasma waves, and one that measures magnetic fields. Engineers believe the latest shutdown could buy the mission roughly another year of breathing room.
The team is also developing a more sweeping power conservation plan they informally call "the Big Bang" — a coordinated swap of several powered components all at once, trading older systems for lower-power alternatives. If testing on Voyager 2, planned for May and June 2026, goes well, the same procedure will be attempted on Voyager 1 no sooner than July. If it works, there is even a slim chance the LECP could once more continue to work.
The engineers say they hope to keep at least one instrument operating on each spacecraft into the 2030s. It would leave both still reporting from places no machine has ever gone before.
181466146
submission
fahrbot-bot writes:
A pair of rare particles produced in high-energy proton collisions may be the clearest evidence yet that mass can emerge from empty space. The finding could shed light on one of the biggest puzzles in physics: how particles acquire their mass.
According to quantum chromodynamics (QCD) – widely considered to be our best theory for describing the strong force, which binds quarks inside protons and neutrons – even a perfect vacuum isn’t truly empty. Instead, it is filled with short-lived disturbances in the underlying energy of space that flicker in and out of existence, known as virtual particles. Among them are quark-antiquark pairs.
Under normal conditions, these fleeting pairs vanish almost as soon as they appear. But if enough energy is injected into a vacuum, QCD predicts they can be promoted into real, detectable particles with measurable mass.
Now, the STAR collaboration – an international team of physicists working at the Relativistic Heavy Ion Collider in Brookhaven National Laboratory in New York state – has observed this process for the first time.
The team smashed together high-energy protons in a vacuum, producing a spray of particles. Some of these particles should be quark-antiquark pairs pulled directly from the vacuum itself, but quarks can never exist alone and immediately combine into composite particles. Quarks and antiquarks are born with their spins correlated — a shared quantum alignment inherited from the vacuum.
The researchers found that this link persists even after the quarks and antiquarks become part of larger particles called hyperons, which decay in less than a tenth of a billionth of a second. Spotting these spin-aligned hyperons in the aftermath of the proton collisions allowed the researchers to confirm that the quarks within them came from the vacuum.
“This is the first time we’ve seen the entire process,” says Zhoudunming Tu, a member of the STAR collaboration.
181211784
submission
fahrbot-bot writes:
CU Boulder researchers are reporting that they have discovered an appetite-suppressing compound in python blood that helps the snakes consume enormous meals and go months without eating yet remain metabolically healthy. The findings were published in the journal Natural Metabolism on March 19, 2026.
Pythons can grow as big as a telephone pole, swallow an antelope whole, and go months or even years without eating—all while maintaining a healthy heart and plenty of muscle mass. In the hours after they eat, research has shown, their heart expands 25% and their metabolism speeds up 4,000-fold to help them digest their meal.
The team measured blood samples from ball pythons and Burmese pythons, fed once every 28 days, immediately after they ate a meal. In all, they found 208 metabolites that increased significantly after the pythons ate. One molecule, called para-tyramine-O-sulfate (pTOS) soared 1,000-fold.
Further studies, done with Baylor University researchers, showed that when they gave high doses of pTOS to obese or lean mice, it acted on the hypothalamus, the appetite center of the brain, prompting weight loss without causing gastrointestinal problems, muscle loss or declines in energy.
The study found that pTOS, which is produced by the snake’s gut bacteria, is not present in mice naturally. It is present in human urine at low levels and does increase somewhat after a meal. But because most research is done in mice or rats, pTOS has been overlooked.
180898892
submission
fahrbot-bot writes:
Interesting Engineering is reporting that CATL, formally known as Contemporary Amperex Technology Limited, has introduced a new fast-charging electric vehicle battery platform designed to significantly reduce charging times while maintaining long-term durability.
The company released performance data for its 5C battery, stating it can fully charge in about 12 minutes while supporting extended cycle life.
The engineering focus behind the platform centers on enabling ultra-fast charging without accelerating battery degradation. A 5C charge rate allows a battery pack to accept high power input, enabling rapid replenishment comparable to short refueling stops.
According to the company’s testing, the battery retained at least 80 percent of its original capacity after 3,000 full charge and discharge cycles under standard temperature conditions. This translates to a projected driving lifespan approaching 1.5 million miles.
The battery was also evaluated under high-temperature conditions to assess real-world endurance. At 140F, it maintained 80 percent capacity after 1,400 cycles, indicating sustained performance even under thermal stress, though with reduced cycle life compared to moderate conditions.
Material innovations underpin the system’s performance. The cathode features a protective coating to reduce structural breakdown during rapid cycling, while the electrolyte contains additives that detect and seal microscopic cracks that could accelerate degradation.
The separator incorporates a temperature-responsive coating that moderates ion movement during heat buildup, helping stabilize the cell during repeated fast charging.
180812352
submission
fahrbot-bot writes:
Science News is reporting that researchers have developed a robotic hand that can not only skitter about on its fingertips, it can also bend its fingers backward, connect and disconnect from a robotic arm and pick up and carry one or more objects at a time. With its unusual agility, it could navigate and retrieve objects in spaces too confined for human hands. Original study published in Nature Communications on January 20, 2026.
When attached to the mechanical arm, the robotic hand could pick up objects much like a human hand. The bot pinched a ball between two fingers, wrapped four fingers around a metal rod and held a flat disc between fingers and palm.
But the bot isn’t constrained by human anatomy. The fingers bend backward just as easily as forward, allowing the robot to hold objects against both sides of its palm simultaneously. It can even unscrew the cap off a mustard bottle while holding the bottle in place.
When the robot was separated from the arm, it was most stable walking on four or five fingers and using one or two fingers for grabbing and carrying things, the team found. In one set of trials with both bots, the hand detached from the robotic arm and used its fingers as legs to skitter over to a wooden block. Once there, it picked up the block with one finger and carried it back to the arm.
The crawling bot could one day aid in industrial inspections of pipes and equipment too small for a human or larger robot to access, says Xiao Gao, a roboticist now at Wuhan University in China. It might retrieve objects in a warehouse or navigate confined spaces in disaster response efforts.
180766118
submission
fahrbot-bot writes:
A simple brain-training exercise could reduce people's risk of developing dementia by 25 percent, a study reported on Monday, but with outside researchers expressing caution in interpreting the results.
The new study is a randomized controlled trial – considered the gold standard for medical research – which first began enrolling participants in the late 1990s.
More than 2,800 people aged 65 or older were randomly assigned one of three different types of brain training – speed, memory, or reasoning – or were part of a control group.
First, the participants did an hour-long training session twice a week for five weeks. One and three years later, they did four booster sessions. In total, there were fewer than 24 hours of training. During follow-ups after five, 10, and most recently 20 years, the speed training was always "disproportionately beneficial", study co-author Marilyn Albert of Johns Hopkins University in the United States told AFP.
After two decades, Medicare records showed that the people who did the speed-training and booster sessions had a 25-percent reduced risk of getting dementia, according to the study. The other two types of training did not make a statistically significant difference.
However, Rachel Richardson, a researcher at the Cochrane Collaboration not involved in the study, cautioned that "while statistically significant, the result may not be as impressive" as a 25-percent reduction. This is partly because the margins of error "range from a reduction of 41 percent to one of only five percent", she told the Science Media Centre. She added that the study excluded people with conditions such as poor vision or hearing, which meant it may not be fully representative.
"Although one subgroup analysis produced a significant result, this single finding is not generally regarded as strong enough evidence to demonstrate the intervention's effectiveness," he said.
"Further research is still needed to determine whether cognitive training can reduce the risk of dementia."
180755846
submission
fahrbot-bot writes:
Right now, molecules in the air are moving around you in chaotic and unpredictable ways. To make sense of such systems, physicists use a law known as the Boltzmann distribution, which, rather than describe exactly where each particle is, describes the chance of finding the system in any of its possible states. This allows them to make predictions about the whole system even though the individual particle motions are random. It's like rolling a single die: Any one roll is unpredictable, but if you keep rolling it again and again, a pattern of probabilities will emerge.
Developed in the latter half of the 19th century by Ludwig Boltzmann, an Austrian physicist and mathematician, this Boltzmann distribution is used widely today to model systems in many fields, ranging from AI to economics, where it is called "multinomial logit."
Now, economists have taken a deeper look at this universal law and come up with a surprising result: The Boltzmann distribution, their mathematical proof shows, is the only law that accurately describes unrelated, or uncoupled, systems.
The research, published in the journal Mathematische Annalen, comes from two economists and mathematicians who both have backgrounds in physics: Omer Tamuz, a professor of economics and mathematics at Caltech, and Fedor Sandomirskiy, a former Caltech postdoc now serving as an assistant professor of economics at Princeton University.
180727748
submission
fahrbot-bot writes:
A spectacular trove of fossils in a discovered in a cave on New Zealand's North Island has given scientists their first glimpse of ancient forest species that lived there more than a million years ago. The fossils represent 12 ancient bird species and four frog species, including several previously unknown bird species. Taken together, the fossils paint a picture of an ancient world that looks drastically different than it does today. The discovery also fills in an important gap in scientific understanding of the patterns of extinction that preceded human arrival in New Zealand 750 years ago.
The team published a study on the find in Alcheringa: An Australasian Journal of Palaeontology.
180726970
submission
fahrbot-bot writes:
Mariano Barbacid, head of the Experimental Oncology Group at the National Cancer Research Centre (CNIO), has designed a therapy that successfully eliminates pancreatic tumours in mice completely and durably, with no significant side effects. The study is published in the journal PNAS (Proceedings of the National Academy of Sciences), with Carmen Guerra as co-lead author and Vasiliki Liaki and Sara Barrambana as first authors.
Current drugs for pancreatic cancer lose effectiveness within months because the tumour becomes resistant. The group from Spain’s National Cancer Research Centre (CNIO) has been able to avoid the development of resistance in animal models with a combined triple therapy.
These results “pave the way for the design of combined therapies that may improve survival,” the authors indicate, although this will not happen in the short term. The results are published in PNAS.
180684930
submission
fahrbot-bot writes:
Science Daily is reporting that researchers at the University of Pennsylvania and the University of Michigan have built the smallest fully programmable autonomous robots ever created. These microscopic machines can swim through liquid, sense their surroundings, respond on their own, operate for months at a time, and cost about one penny each to produce.
Each robot is barely visible without magnification, measuring roughly 200 by 300 by 50 micrometers. That makes them smaller than a grain of salt. Because they function at the same scale as many living microorganisms, the robots could one day help doctors monitor individual cells or assist engineers in assembling tiny devices used in advanced manufacturing.
Powered by light and equipped with tiny computers, the robots swim by manipulating electric fields rather than using moving parts. They can detect temperature changes, follow programmed paths, and even work together in groups
The work was reported in Science Robotics and Proceedings of the National Academy of Sciences (PNAS). Unlike previous tiny machines, these robots do not rely on wires, magnetic fields, or external controls.
Alternate article: Penn and Michigan Create World’s Smallest Programmable, Autonomous Robots
180441531
submission
fahrbot-bot writes:
The Associated Press is reporting that President Donald Trump has announced a bold plan for the Navy to build a new, large warship that he is calling a “battleship” as part of a larger vision to create a “Golden Fleet.”
“They’ll be the fastest, the biggest, and by far 100 times more powerful than any battleship ever built,” Trump claimed during the announcement at his Mar-a-Lago resort in Florida.
The ship, according to Trump, will be longer and larger than the World War II-era Iowa-class battleships and will be armed with hypersonic missiles, rail guns, and high-powered lasers — all technologies that are still being developed by the Navy.
He said Monday he will have a direct role in designing this new warship ... “The U.S. Navy will lead the design of these ships along with me, because I’m a very aesthetic person,” Trump said.
180395391
submission
fahrbot-bot writes:
SciTechDaily is reporting that a research team has created a quantum antenna capable of precisely measuring terahertz frequency combs for the first time.
A research team from the Faculty of Physics and the Centre for Quantum Optical Technologies at the Centre of New Technologies, University of Warsaw has introduced a new approach for detecting extremely weak terahertz signals by using a “quantum antenna.” Their method relies on a specialized system that employs Rydberg atoms for radio wave detection, allowing them not only to capture these signals but also to accurately calibrate a frequency comb in the terahertz range.
This part of the electromagnetic spectrum was considered largely unexplored until recently, and the breakthrough, reported in Optica, offers a pathway toward highly sensitive spectroscopy and a new class of quantum sensors that can function at room temperature.
Terahertz (THz) radiation occupies a unique position within the electromagnetic spectrum, sitting between microwaves (such as those used in Wi-Fi) and infrared light at the intersection of electronics and optics. It promises a wide range of applications, including scanning packages without harmful X-rays, enabling ultra-fast 6G communication, and advancing spectroscopy and organic compound imaging.
180171437
submission
fahrbot-bot writes:
The Telegraph is reporting that Ukraine forces are jamming signals for Russia's ‘invincible’ Kinzhal hyper-sonic missile with a song satirizing Russian propaganda.
Night Watch, the group operating the technology, claims to have brought down 19 Kinzhal missiles – described by Putin as “invincible” – in the past two weeks.
The team told technology website 404 Media that it is using a song and a redirection order to knock the “next-generation” missiles, which carry a 480kg payload and cost around £7.7m each, out of the sky.
Kinzhals and other guided munitions rely on the GLONASS system – Russia’s GPS-style navigation network using satellites – to find their targets. Night Watch developed its own “Lima” jamming system that replaces the missiles’ satellite navigation signals with the Ukrainian song “Our Father is Bandera”.
When the song begins, the Lima system feeds the incoming missiles a false navigation signal, tricking them into believing that they are flying over Lima, in Peru, so that they attempt to change their trajectory. Traveling at a speed of more than 4,000 miles per hour, however, the missiles become destabilized by the abrupt and unexpected change of course.
Night Watch said they developed the system after discovering that the Kinzhals used a controlled reception pattern antenna (CRPA), an antiquated type of technology for resisting, jamming and spoofing. The team told 404: “They had the same type of receivers as old Soviet missiles used to have.
“The airframe cannot withstand the excessive stress and the missile naturally fails. When the Kinzhal tried to quickly change navigation, the fuselage of this missile was unable to handle the speed and, yeah, it was just cut into two parts. The biggest advantage of those missiles, speed, was used against them.”
