Lifeboat Foundation

DeepMind says that it has developed an AI model, called RoboCat, that can perform a range of tasks across different models of robotic arms. That alone isn’t especially novel. But DeepMind claims that the model is the first to be able to solve and adapt to multiple tasks and do so using different, real-world robots.

“We demonstrate that a single large model can solve a diverse set of tasks on multiple real robotic embodiments and can quickly adapt to new tasks and embodiments,” Alex Lee, a research scientist at DeepMind and a co-contributor on the team behind RoboCat, told TechCrunch in an email interview.

RoboCat — which was inspired by Gato, a DeepMind AI model that can analyze and act on text, images and events — was trained on images and actions data collected from robotics both in simulation and real life. The data, Lee says, came from a combination of other robot-controlling models inside of virtual environments, humans controlling robots and previous iterations of RoboCat itself.

At the turn of this century, Jeff Bezos popularized the use of mechanical turks—low-paid workers working remotely with perhaps thousands of others on tiny parts of larger computer projects—to ensure a human perspective on mostly simple tasks that proved perplexing to computers. He termed this blending of human and digital brain power “artificial artificial intelligence.”

About a quarter million people are employed through Amazon’s Mechanical Turk marketplace, just one of many sources providing such services.

This week, researchers at Swiss-based university EPFL reported that turks who had provided important human input are now relying on AI-generated content to complete their tasks. They dubbed this phenomenon “artificial artificial artificial intelligence.”

Insects, with their remarkable ability to undergo complete metamorphosis, have long fascinated scientists seeking to understand the underlying genetic mechanisms governing this transformative process.

Now, a recent study conducted by the Institute for Evolutionary Biology (IBE, CSIC-UPF) and the IRB Barcelona has shed light on the crucial role of three genes – Chinmo, Br-C and E93 – in orchestrating the stages of insect development. Published in eLife, this research provides valuable insights into the evolutionary origins of metamorphosis and sheds new light on the role of these genes in growth, development and cancer regulation [1].

Longevity. Technology: Chinmo might sound like a Pokémon character, but the truth is much more interesting. Conserved throughout the evolution of insects, scientists think it, and the more conventionally-named Br-C and E93, could play a key role in the evolution of metamorphosis, acting as the hands of the biological clock in insects. A maggot is radically different from the fly into which it changes – could understanding and leveraging the biology involved one day allow us to change cultured skin cells into replacement organs or to stop tumors in their early stages of formation? No, Dr Seth Brundle, you can buzz off.

One hundred thousand tons of clothes dumped illegally in a Chilean desert. The tragic collapse of the Rana Plaza garment factory in Bangladesh… People are increasingly turning to a concept called “circular fashion” that may help end situations like that. Beyond Reduce, Reuse and Recycle, it encourages innovative designs and values that attract both young and old. We catch up with some of the new business models, as well as the people buying into these novel products and services.

Guest:

Yasui Akihiro (Circular-economy researcher)

Just yesterday, an unstable sunspot named AR3335 exploded, producing a solar flare that triggered blackouts over the Atlantic Ocean. The resulting solar flare was M2.5 in intensity and caused a shortwave radio blackout. Solar activity has been on the rise for the past few months, and it is expected to increase further until solar maximum, the period of greatest solar activity during the Sun’s 11-year cycle.

Solar flare risk

According to a report by spaceweather.com, NASA’s Solar Dynamics Observatory (SDO) forecasters have observed multiple streams of solar winds hurtling towards Earth from a coronal hole on the Sun’s surface, and these could reach Earth tomorrow, June 21. Moreover, a CME is also expected to deliver a glancing blow on June 22. Both these events have the potential to trigger a G1-class Geomagnetic storm. It could also result in solstice auroras at high latitudes.

Astronomers have mapped 39 interstellar clouds where high-mass stars are expected to form. This large data set shows that the accepted model of low-mass star formation needs to be expanded to explain the formation of high-mass stars. This suggests the formation of high-mass stars is fundamentally different from the formation of low-mass stars, not just a matter of scale.

High-mass stars play an important role in the evolution of the universe through the release of heavy elements and the shock waves produced when a massive star explodes in a supernova. Despite their importance, the way massive stars form remains poorly understood due to their rarity.

To better understand massive star formation a team led by Kaho Morii, Patricio Sanhueza, and Fumitaka Nakamura used the Atacama Large Millimeter/submillimeter Array (ALMA) to observe 39 infrared dark clouds (IRDCs). IRDCs are massive, cold, and dense clouds of gas and dust; and are thought to be the sites of massive star formation. The team focused on clouds showing no signs of star formation, to understand the beginning of the formation process before young stars ignite. In the 39 clouds, the team found more than 800 stellar seeds, referred to as molecular cloud cores, which astronomers think will evolve into stars.

Eventually, these molecules probably evolved a lipid (fatty) boundary separating the internal environment of the organism from the exterior, forming protocells. Protocells could concentrate and organize better the molecules needed in biochemical reactions, providing a contained and efficient metabolism.

Life on Repeat?

Abiogenesis could have happened more than once. Earth could have birthed self-replicating molecules several times, and maybe early life for thousands or millions of years just consisted of a bunch of different self-replicating RNA molecules, with independent origins, competing for the same building blocks. Alas, due to the ancient and microscopic nature of this process, we may never know.

Tesla announced it produced 10 million 4,680 battery cells at Gigafactory Texas. It is a good sign for the automaker’s production ramp-up, which relies heavily on the new cell.

The 4,680 battery cell format has taken the industry by storm since Tesla unveiled its own cell strategy at Battery Day in 2020.

The automaker claimed the potential to reduce battery cost by over 50% with the new design; it has been trying to bring it to volume production since, but it has run into some bottlenecks.

This study demonstrates that AI can be incredibly effective in helping us identify new drug candidates – particularly at early stages of drug discovery and for diseases with complex biology or few known molecular targets.

A machine learning model has been trained to recognise the key features of chemicals with senolytic activity. It recently found three chemicals able to remove senescent cells without damaging healthy cells.

Molecular structure of oleandrin. Credit: Mplanine, CC BY-SA 4.0, via Wikimedia Commons.

Continue reading “AI finds potential anti-aging molecules” »