Lifeboat Foundation: Safeguarding Humanity
Toggle light / dark theme

Blog

Get the latest international news and world events from around the world.

Log in for authorized contributors

Building a hill-climbing machine: Launching seven new MAI models

The title’s “hill-climbing machine” refers to Microsoft’s iterative, scientifically rigorous engineering framework. By tightly linking clean data pipelines, specialized training infrastructure, and reinforcement learning environments, they have created an optimization loop designed to steadily “climb” toward higher capabilities as compute scales.

Today we are announcing a family of seven new models developed in-house at Microsoft AI. Beyond these models, we’re building a superintelligence lab – a system and an approach we believe will define the next phase of AI.

This is an extraordinary time in technology. The compute used to train frontier models has increased by a factor of one trillion. Now we expect another thousand-fold increase over the next three years, which in turn means more advanced capabilities, and the continued rollout of ever more effective AI.

This epic compute ramp will change the nature of work, business and daily life. We all have to prepare for this reality. Our job at MAI is to help you do this – to push the frontier, and to build a hill-climbing machine to keep you at the frontier.

Explosive cytotoxicity of ruptoblasts bridges hormone surveillance and immune defense

Ruptoblasts are a cytotoxic glandular cell type that undergoes explosive cell death, ruptosis, releasing potent, broad-spectrum agents capable of killing nearby cells. Ruptosis is triggered by the hormone and cytokine activin and driven by ER-derived calcium release amplified by the actomyosin cytoskeleton.

LLMs help robots understand vague instructions and focus on key details

Imagine working at a warehouse or office sometime in the near future, and you’re asked to help a new trainee learn the basics of their job. The catch: It’s a robot. To teach them, you might want to play a game of “show and tell”—that is, physically showing how to do something a few different ways, while also explaining what you’re doing.

Let’s say you asked the robot to place some coffee on your desk without disturbing you during a Zoom call. You’ll prefer that the robot doesn’t get close to you and the laptop so that it doesn’t interrupt your meeting. To enable this behavior, the robot should be trained with data that clearly demonstrates the full task. Computer scientists have attempted to explain manipulation tasks to robots by recording lots of physical demonstrations or writing extensive directions. But if you don’t have both, the machine is likely to misunderstand what it needs to do.

It’s laborious for humans to do all that showing and telling, so researchers at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) have automated the process of teaching a robot, while clarifying instructions automatically and using nearly five times less demonstration data.

Rare meteorite provides evidence of giant early planet

Four-and-a-half billion years ago, a massive world—possibly as big as the moon or even Mars—orbited our sun before crashing into another celestial body and shattering into rubble. Now, in a paper published in the journal Earth and Planetary Science Letters, scientists report the first definitive evidence that this lost planetary embryo (protoplanet) existed. Its unique geological makeup challenges long-held assumptions about how planets evolve.

“It’s incredible to think there was once a world this large,” said Aaron Bell, an assistant research professor in the Department of Earth Science at the University of Colorado Boulder. “We only know it existed because a few fragments of it happened to land on Earth. These meteorites preserved evidence of a completely different pathway through which early planets developed.”

What gave away the lost world’s secret was a piece of its debris uncovered on Earth in the Sahara Desert, known as the Northwest Africa (NWA) 12,774 angrite meteorite.

Cutting a photon in two creates an infinite swarm of particles

By definition, elementary particles can’t be broken into smaller pieces. But in a new theoretical study published in Physical Review Letters, Johannes Skaar and colleagues have revealed what would happen if you tried anyway for a single photon. The answer is deeply strange: attempting to cut a photon in two wouldn’t produce two smaller photons, but instead conjure an infinite number of them out of thin air.

Like any quantum particle, a photon exists simultaneously as a single, localized particle, and an extended wave, spread out across space. For their investigation, Skaar’s team considered what would happen if a single photon passed through an optical shutter—essentially a very fast mirror that can be switched on and off to block part of a pulse of light. If the shutter was fast enough, it could intercept the photon mid-pulse, snipping off part of this extended wave.

To find out what would happen afterward, the researchers applied quantum equations that describe how the photon’s underlying electromagnetic field behaves at the quantum level. Specifically, their analysis tracked precisely how the photon’s quantum state would be transformed by the shutter’s intervention.

Terahertz imaging maps spatial chirality in materials with 100-micrometer resolution

In nature, there exist structures that are mirror images of each other but cannot be perfectly superimposed. These are known as chiral objects, derived from the Greek word for “hand,” since left and right hands share the same relationship. Although similar in structure, chiral molecules exhibit different behaviors, and chirality is central to life itself. DNA has a twisted chiral structure, and living organisms prefer one handedness over the other. This distinction is equally important in drug design, materials science, and nanotechnology.

One way to distinguish chiral molecules is by measuring their response to circularly polarized light in the terahertz (THz) region. THz waves lie between microwaves and infrared light and are especially sensitive to subtle collective motions and twisting structures in materials. However, conventional THz measurements average the signal across an entire sample, making it impossible to determine how chirality varies across different locations.

Now, researchers from Chiba University, Japan, and Tohoku University, Japan, have shown that this limitation can be overcome, allowing chirality to be visualized as two-dimensional images, much like creating a map of chirality across a material. The work appears in ACS Photonics.

Plate tectonics shaped the Cradle of Civilization by merging two ancient rivers, study suggests

The Euphrates River is the longest river in Western Asia and runs through the eastern side of the Fertile Crescent. Flowing over 1,700 miles from Turkey through Syria and Iraq, the river played a crucial role in sustaining the region known as the “Cradle of Civilization.” Yet, researchers aren’t sure about the river’s origins or how tectonic activity might have shaped its evolution. A new study, published in Nature Geoscience, suggests that two ancient rivers, diverted by shifting plate tectonics, merged to form this vital river.

Previous research proposed two main hypotheses about where the early Euphrates river ended: the first was that the river ended in lakes in Anatolia or the Mediterranean; the second was that it flowed southeast into Arabia.

The team involved in the new study used seismic reflection and topographic data to map out and study ancient river paths and sediment deposits. They identified two ancient rivers, the Paleo-Karasu and Paleo-Murat, which appeared to flow into the Mediterranean Sea until around 3.6 million years ago. They found that the rivers were emptying into the Mediterranean in the Late Miocene, during a period in which the Mediterranean was partially dried up, referred to as the Messinian Salinity Crisis (MSC), lasting from 5.97–5.33 million years ago.

How a Richard Feynman formula could explain your dining habits in a new city

One of the dilemmas facing anyone in a new and unfamiliar city is where to dine out. You might consult guides, speak to locals, check reviews, and ultimately, try your luck. But if you’re there for a while, at some point you’re going to be asking yourself whether to visit new eateries or stick to the ones you’ve already tried and liked.

This is known as a classic explore-exploit dilemma and was something the late physicist and Nobel laureate Richard Feynman pondered during a restaurant meal with a friend in the 1970s. His companion was debating whether to order his favorite dish or try something new. Feynman turned the question into a math problem and solved it there and then, scribbling his workings on pieces of paper.

Feynman, who died in 1988, never published his solution, but researchers came across his handwritten notes and not only deciphered them, but also put the solution to the test.

Britain’s oldest cave art may have been rediscovered in Bacon Hole cave

The oldest cave art in Britain may have been discovered, or more likely rediscovered, in a cave on the Gower Peninsula in South Wales, possibly dating back around 17,000 years.

The red stripe markings on the walls of a cave called Bacon Hole were first spotted in 1912. They were hailed by their finders, Professor William Sollas and Henri Breuil, as the first Paleolithic cave art in the British Isles. However, skeptics dismissed the discovery, arguing that the markings were merely natural mineral deposits.

Over time, the claim and subsequent debate drifted out of public consciousness and have largely been forgotten about. That was until scientists decided to reexamine the marks.

/* */