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

On the Business Strategies of Platform Biotech Companies

Platform biotech companies face a different strategic landscape relative to single-asset biotechs. Here’s my short educational writeup on the fundamentals of the platform biotech business approach.

Investors may sometimes favor platform biotech companies, though usually only when markets are strong. While opinion tends to swing with the markets, there are certain real advantages to platforms. Particularly, investors find attractive structures with more routes to a very large return outcome, which platforms can offer.

A single-asset company typically will generate a return if it shows a clinical success (often successful phase 2 or even phase 3 clinical trial data) with its sole program and is acquired. By contrast, platform companies are not always killed by a single program failure. So, platforms provide more shots on goal in a sense. That said, if the core platform (e.g. delivery system) fails in humans, the entire company will die [1]. Because of this, it is still critical that the platform company’s first asset program goes well.

Platforms feature routes to preclinical revenue which are much more difficult to achieve for assets. Pharma companies will pay for access to enabling platform technology even when only preclinical data are available. Programs by Dyno Therapeutics (AAVs) [2] and Evox Therapeutics (exosomes) [3] represent examples of the phenomenon. This provides third-party validation and non-dilutive capital before the clinical data, which can support the company during the otherwise most challenging period in a biotech’s lifetime.

Robotok: Mining Internet Videos for Dexterous Robot Learning

What if robots could learn dexterous skills by watching videos on the internet?

Teaching robots to manipulate objects is challenging because collecting training data is expensive, time-consuming, and difficult to scale across different tasks and environments. Humans, on the other hand, demonstrate countless everyday skills in videos uploaded online. The challenge is finding the right demonstrations that can help a robot learn how to perform a specific action.

A new system called RoboTok tackles this problem by turning internet videos into a searchable resource for robot learning. Given a reference video of a person performing a manipulation task, RoboTok searches for other videos showing similar hand movements and interactions with objects. Rather than relying mainly on visual similarities, it analyzes the motion trajectories of the hands in three-dimensional space, using a representation centered on the person performing the action.

This approach allows the system to recognize similar manipulation behaviors even when videos differ in camera angles, backgrounds, visual appearances, or when parts of the person’s body are obscured. Because the motion representation is compact, RoboTok can also index and search large collections of online videos efficiently.

The researchers evaluated the system using video-retrieval benchmarks and real-world robot experiments. Their results showed that RoboTok retrieved more relevant demonstrations than existing approaches and helped robots perform manipulation tasks more successfully. In real-world tests, RoboTok-guided robot policies achieved an average improvement of 42.2 percentage points in task success rates compared with the strongest baseline for each task.

The broader implication is significant: the internet could become a vast, continuously growing source of training data for robots. Instead of relying exclusively on costly, specially collected robot demonstrations, researchers could extract useful manipulation examples from videos of humans performing everyday tasks.

RoboTok does not mean robots can automatically master every skill simply by watching videos. They still need suitable learning methods and robot-specific training. But by making human demonstrations easier to find and reuse, this approach offers a promising path toward more scalable robot learning and more dexterous machines.

#ArtificialIntelligence #robotics #ImitationLearning #DexterousManipulation.

Continue reading “Robotok: Mining Internet Videos for Dexterous Robot Learning” | >

Astronomers Have Detected A Mysterious Signal Coming From Empty Space. Nobody Knows What It Is

In the southern constellation of Fornax lies a tiny patch of sky just like any other. Your eyes may have brushed past it a hundred times without a thought.

Yet this otherwise unremarkable speck of the cosmos is among the most well-cataloged in the sky. The Hubble Space Telescope stared at it for more than 270 hours to obtain its famous Ultra Deep Field (HUDF) – and contained therein are nearly 10,000 galaxies extending billions of light-years into the Universe.

When the James Webb Space Telescope (JWST) launched, it was only natural that its golden mirrors would turn to the same region of space, using its vastly more sensitive infrared capabilities to find what Hubble may have missed.

Common leukemia mutation derails protein positioning during cell division, study reveals

Experts from the University of Nottingham have discovered how a common genetic mutation disrupts the positioning of a key protein during cell division, revealing a new mechanism behind acute myeloid leukemia and a potential avenue for future treatment.

Acute myeloid leukemia (AML) is an aggressive blood cancer that disrupts the production of healthy blood cells. Around one-third of patients carry a mutation in a gene called NPM1, which causes its protein to accumulate in the wrong part of the cell and plays a key role in driving the disease.

Until now, research has largely focused on how the mutated protein moves between the cell’s nucleus and the surrounding cytoplasm. But a new study, led by Dr. Daniel Booth and Dr. Claire Seedhouse at the University of Nottingham’s School of Medicine, reveals that what happens during cell division itself is also critical.

Linus Torvalds says AI is a gateway drug for coding

Linus Torvalds is using AI to develop Linux kernel code now https://www.zdnet.com/innovation/linus-torvalds-uses-ai-to-w…nel-patch/

Linux creator Linus Torvalds says AI is like a “gateway drug”

Linus Torvalds AI stance has softened as he says tools boost productivity but still create false bug reports. Read what his comments reveal.

I Tested AI Cloning… and It’s Way Worse Than You Think!

I used AI to reverse engineer real binaries, understand how they work, and then clone their functionality into brand-new applications. In this video I show the full process — what the AI can actually figure out, where it falls short, and how far you can push the technique of turning an existing binary into something new.

This is the practical side of AI-assisted reverse engineering, not theory.

What you’ll see:

How AI helps tear apart and understand binaries.
Extracting logic, behavior, and structure.
Cloning that functionality into a clean new application.
Real examples of what works and what still needs a human engineer.
The limits, risks, and surprising capabilities of current tools.

If you’ve ever wondered how far AI can go in reverse-engineering and recreating software from binaries, this is the deep dive.

For a free download of Task Manager TMOG, visit: https://tmog.org

Continue reading “I Tested AI Cloning… and It’s Way Worse Than You Think!” | >

Jamming the flow to enable memory in quantum materials

Understanding and controlling phase transitions are fundamental to physics and central to many technological revolutions. There is currently strong interest in materials with coupled structural and electronic phase transitions, where controlling the transition plasticity naturally yields built-in memory—a key step in emulating neurons and synapses for “neuromorphic” or brain-like computing technologies.

A team of UC San Diego researchers led by physics professor Oleg Shpyrko developed a new approach to coherent X-ray diffraction—a microscopy technique that is highly sensitive to specific nanoscale domain configurations —that allowed them to observe dynamics within a neuristor (a man-made device that mimics the brain’s neurons) made of vanadium dioxide while electricity was applied.

By taking measurements while operating the neuristor, they discovered a strategy to make the material learn and forget a specific state, opening an avenue toward developing energy-efficient brain-inspired hardware.

Rocks from hydrothermal vents can spontaneously combust, posing potential risks for deep-sea mining operations

Projects involving the mining of seafloor massive sulfide (SMS) deposits around hydrothermal vents already face criticism over their destructive nature, and now these projects may face a new concern: spontaneous combustion. A new study, published in Scientific Reports, describes how certain samples of rock brought up from SMS deposits have unexpectedly caught fire. The study indicates that conditions at the surface, combined with processing activities, may turn some of these rocks into a hazard.

Hydrothermal vents on the ocean floor release heated water rich in minerals that result in metal-rich deposits over time. These deposits can become valuable sources of copper, zinc, gold and silver. However, the deposits also contain sulfur-bearing minerals that can release heat when they react with oxygen in the air. If that heat builds up, the material can spontaneously combust.

Spontaneous combustion has been occasionally documented in sulfide mines on land, referred to as volcanogenic massive sulfide deposits (VMS). Some safety practices have been developed for land-based mines, but seafloor rocks are more porous and likely require different safety practices.