Lifeboat Foundation: Safeguarding Humanity
Toggle light / dark theme

Blog – Page 3

In a surprising twist, a graduate student at UMass Amherst discovered a strange new fluid behavior that seems to defy thermodynamics.

While experimenting with oil, water, and magnetized nickel particles, he found that no matter how hard the mixture was shaken, it would always return to the same elegant urn shape. This behavior sparked curiosity among physicists, who eventually traced the cause to unusually strong magnetism altering the way the fluids interact. Though it has no immediate use, the finding opens new frontiers in soft-matter science.

A surprising discovery in soft matter.

Read more | >

A surprising link between the immune system and brain behavior is emerging, as new research reveals how a single immune molecule can affect both anxiety and sociability depending on which brain region it acts upon.

Scientists found that IL-17 behaves almost like a brain chemical, influencing neuron activity in ways that alter mood and behavior during illness. These findings suggest the immune system plays a much deeper role in shaping our mental states than previously thought, opening new doors for treating conditions like autism and anxiety through immune-based therapies.

Immune Molecules and Brain Behavior.

Read more | >

Scientists have uncovered the strongest evidence yet for the existence of elusive intermediate-mass black holes (IMBHs), long thought to be the missing link between stellar-mass and supermassive black holes. By tracking a hypervelocity star, J0731+3717, that appears to have been ejected from the

Read more | >

Scientists in Japan have uncovered a surprising twist, literally, in how molecules organize themselves. By introducing tiny leftover fragments from previous assemblies, they discovered a way to flip the direction of helical molecular structures.

Using specific intensities of UV and visible light, they controlled whether these molecules formed left-handed or right-handed spirals, revealing a new method to fine-tune optical and electronic properties. This groundbreaking insight could unlock novel ways to engineer smarter, more responsive materials.

Revealing the power of molecular self-assembly.

Read more | >

The threat actor known as Paper Werewolf has been observed exclusively targeting Russian entities with a new implant called PowerModul.

The activity, which took place between July and December 2024, singled out organizations in the mass media, telecommunications, construction, government entities, and energy sectors, Kaspersky said in a new report published Thursday.

Paper Werewolf, also known as GOFFEE, is assessed to have conducted at least seven campaigns since 2022, according to BI.ZONE, with the attacks mainly aimed at government, energy, financial, media, and other organizations.

Read more | >

The clone websites identified by DTI include a carousel of images that, when clicked, download a malicious APK file onto the user’s device. The package file acts as a dropper to install a second embedded APK payload via the DialogInterface. OnClickListener interface that allows for the execution of the SpyNote malware when an item in a dialog box is clicked.

“Upon installation, it aggressively requests numerous intrusive permissions, gaining extensive control over the compromised device,” DTI said.

“This control allows for the theft of sensitive data such as SMS messages, contacts, call logs, location information, and files. SpyNote also boasts significant remote access capabilities, including camera and microphone activation, call manipulation, and arbitrary command execution.”

Read more | >

Microsoft is testing a new Defender for Endpoint capability that will block traffic to and from undiscovered endpoints to thwart attackers’ lateral network movement attempts.

As the company revealed earlier this week, this is achieved by containing the IP addresses of devices that have yet to be discovered or onboarded to Defender for Endpoint.

Redmond says the new feature will prevent threat actors from spreading to other non-compromised devices by blocking incoming and outgoing communication with devices using contained IP addresses.

Read more | >

Large Reasoning Models like DeepSeek-R1 mark a fundamental shift in how LLMs approach complex problems. Instead of directly producing an answer for a given input, DeepSeek-R1 creates detailed multi-step reasoning chains, seemingly ‘thinking’ about a problem before providing an answer. This reasoning process is publicly available to the user, creating endless opportunities for studying the reasoning behaviour of the model and opening up the field of Starting from a taxonomy of DeepSeek-R1’s basic building blocks of reasoning, our analyses on DeepSeek-R1 investigate the impact and controllability of thought length, management of long or confusing contexts, cultural and safety concerns, and the status of DeepSeek-R1 vis-à-vis cognitive phenomena, such as human-like language processing and world modelling. Our findings paint a nuanced picture. Notably, we show DeepSeek-R1 has a ‘sweet spot’ of reasoning, where extra inference time can impair model performance. Furthermore, we find a tendency for DeepSeek-R1 to persistently ruminate on previously explored problem formulations, obstructing further exploration. We also note strong safety vulnerabilities of DeepSeek-R1 compared to its non-reasoning counterpart, which can also compromise safety-aligned LLMs.

Read more | >