Toggle light / dark theme

In a nutshell: In the 21st century, there’s a new space race, primarily between Tesla (SpaceX) and Amazon (Blue Origin). Both are concentrating on traditional methods of launching satellites into space—namely, big rockets. However, a small space startup called SpinLaunch is developing a cheaper and environmentally cleaner way of launching satellites.

SpinLaunch is using a gigantic centrifuge to shoot stuff into space. By “stuff,” we mean things that can withstand the G-force created by being spun at 5,000 miles per hour (over 10,000 Gs), which is a category of stuff that does not yet include satellites. However, it did launch a missile-like projectile tens of thousands of feet into the air last month, using only 20 percent of the accelerator’s power.

The design is relatively simple. A carbon fiber tether holds the projectile inside a span vacuum chamber as it spins up to speed. Once the centrifuge has reached the desired velocity, the launch vehicle is released out a tube taller than the Statue of Liberty (50.4 meters). It is not unlike launching your friends off the merry-go-round when you were a kid. A more mature and controlled application of the principle would be the hammer throw event in the Olympics.

Google’s own Deepmind has just released a new revolutionary Artificial Intelligence model which will likely power many AI Applications in the future. They named it Perceiver and it’s meant to replace the most popular AI transformer models.

Perceiver is meant to bring a more general approach to problems from AI models which some would call Artificial General Intelligence. Whether or not Deepmind will manage to create the best AI in the future is yet to be seen. But one thing is for sure, the future is quite weird for AI but also amazing to see.

If you enjoyed this video, please consider rating this video and subscribing to our channel for more frequent uploads. Thank you! smile

TIMESTAMPS:
00:00 A new kind of AI
01:46 What this new model does.
03:40 How Perceiver works.
05:40 Problems with Perceiver.
06:28 Last words.

#deepmind #ai #agi

The most intelligent AI Scientists in the world are becoming increasingly worried about Artificial Intelligence programs becoming more unpredictable and incomprehensible as they become more powerful. AI is also overtaking powerful positions in the government, healthcare and defense which could prove dangerous as an Artificial Super Intelligence is coming very close as the Singularity approaches in the future of 2045. People like Elon Musk and Ray Kurzweil have long warned us about AI beating Humans in anything we can imagine. Nvidia and Meta are also working on specially made hardware and software in the form of pytorch and 2022 GPU’s. Artificial General Intelligence is a real dangers and here are some solutions to it.

TIMESTAMPS:
00:00 The Dawn of incomprehensible AI
01:33 The Dangers of Artificial Intelligence.
03:03 A Possible solution.
04:23 What ASI means for Society.
07:15 So, is all hope lost?
09:03 Last Words.

#ai #asi #agi

Summary: A newly developed AI algorithm can directly predict eye position and movement during an MRI scan. The technology could provide new diagnostics for neurological disorders that manifest in changes in eye-movement patterns.

Source: Max Planck Institute.

A large amount of information constantly flows into our brain via the eyes. Scientists can measure the resulting brain activity using magnetic resonance imaging (MRI). The precise measurement of eye movements during an MRI scan can tell scientists a great deal about our thoughts, memories and current goals, but also about diseases of the brain.

“If you want to measure something with very high precision, you almost always use an , because light makes for a very precise ruler,” says Jaime Cardenas, assistant professor of optics at the University of Rochester.

Now, the Cardenas Lab has created a way to make these optical workhorses even more useful and sensitive. Meiting Song, a Ph.D. student, has for the first time packaged an experimental way of amplifying interferometric signals—without a corresponding increase in extraneous, unwanted input, or “noise”—on a 1 mm by 1 mm integrated photonic . The breakthrough, described in Nature Communications, is based on a theory of weak value amplification with waveguides that was developed by Andrew Jordan, a professor of physics at Rochester, and students in his lab.

Crew Dragon resulted from a partnership between NASA and SpaceX over the last decade. Anticipating the space shuttle’s retirement, NASA worked with SpaceX and Boeing to privately develop launch systems to carry astronauts into low Earth orbit. When the final contracts were signed in 2014, it was expected that SpaceX and Boeing would each fly one mission a year. Boeing, however, has run into technical challenges with the development of its Starliner spacecraft, so Crew Dragon has had to perform double duty from the start.

“I think we’re incredibly grateful with the partnership that we’ve had,” Kathy Lueders, chief of human spaceflight operations for NASA, said during a post-launch news conference of the NASA-SpaceX team. “You know, when I first started in commercial crew, six or seven years ago, it would have been a dream to me that we would have flown these four missions back to back. Because it’s a really tough thing to do. So I’m incredibly proud of this joint team.”

The cadence is even more impressive considering that SpaceX also recently debuted an upgraded version of its Cargo Dragon spacecraft. Including crew and cargo missions, SpaceX has either launched or landed a Dragon spacecraft every month in 2021 except for February and March.