Lifeboat Foundation

In 2020, OpenAI introduced GPT-3 and, a year later, DALL.E, a 12 billion parameter model, built on GPT-3. DALL.E was trained to generate images from text descriptions, and the latest release, DALL.E 2, generates even more realistic and accurate images with 4x better resolution. The model takes natural language captions and uses a dataset of text-image pairings to create realistic images. Additionally, it can take an image and create different variations inspired by original images.

DALL.E leverages the ‘diffusion’ process to learn the relationship between images and text descriptions. In diffusion, it starts with a pattern of random dots and tracks it towards an image when it recognises aspects of it. Diffusion models have emerged as a promising generative modelling framework and push the state-of-the-art image and video generation tasks. The guidance technique is leveraged in diffusion to improve sample fidelity for images and photorealism. DALL.E is made up of two major parts: a discrete autoencoder that accurately represents images in compressed latent space and a transformer that learns the correlations between language and this discrete image representation. Evaluators were asked to compare 1,000 image generations from each model, and DALL·E 2 was preferred over DALL·E 1 for its caption matching and photorealism.

DALL-E is currently only a research project, and is not available in OpenAI’s API.

It’s 85 miles wide, which is larger than the US state of Rhode Island. And it’s coming this way.

We can travel in time, just not like in the movies Earendel was observed using an effect where the fabric of space-time is warped by gravity – a phenomenon predicted by Einstein. This causes light to bend as it passes by objects with large masses, like planets, suns, or even galaxies, allowing us to see around and even behind these objects.

NASA’s Perseverance rover is hightailing it to a fascinating river delta region in the Jezero Crater on Mars. But to get there, it first had to pass near its original landing site. Images from there are a trip down memory lane, back to when Percy dramatically landed on the red planet in February 2021.

Steve Ruff, Arizona State University associate research professor and Mars geologist, runs the Mars Guy channel on YouTube. He posted a video on Sunday recapping the rover’s arrival on mars and what happened to the parachute and back shell — two key components of the landing system that delivered Percy safely to the surface.

Can the US and Russia still collaborate in space?

Subscribe and turn on notifications 🔔 so you don’t miss any videos: http://goo.gl/0bsAjO

Continue reading “What Russia’s war means for the International Space Station” »

Like something out of a sci-fi movie, the Hubble Space Telescope discovered the most distant star ever seen—through a cosmic fluke.

Nicknamed “Earendel,” the star’s light took 12.9 billion years to reach Earth, and scientists believe it emitted light throughout the first billion years of our universe’s existence. Beside its age, Earendel’s position is also curious. According to NASA, Hubble was able to detect the star thanks to gravitational lensing, a phenomenon in which a massive object (like a galaxy) warps the fabric of space. When light reaches the warped space, it can act like a magnifying glass and highlight objects we’d otherwise have a difficult time seeing.

And that’s how Earendel—whose Old English name is akin to “morning star” or “rising light”—was found (though Tolkien fans will recognize it as a riff on Eärendil, the half-Elven seafarer who carried a Silmaril).

Vast object is falling towards the middle of the solar system – but will stop before it reaches us.

The Lunar Crater Radio Telescope is now moving into Phase 2 of development. It’ll be built by robots, into a natural bowl-shaped crater on the moon’s far side. If completed, it’ll be the largest radio telescope in the solar system.