Every image you’ve ever seen of the sun is looking at its equator, because Earth’s orbit sits there with a 7.25-degree tilt. That means humans have never had a good angle to view the sun’s north and south poles until now. The European Space Agency has released images of the sun’s north and south poles, giving everyone their first glimpse at the top and bottom of our nearest star.

The imagery was taken by the ESA’s Solar Orbiter, which began its trek to view the sun’s polar regions in 2020. To do this, the Orbiter engaged its boosters, made some adjustments, and slingshotted itself around Venus at a staggering 27,000 miles per hour. 

Once it reached its destination, it took images using its Polarimetric and Helioseismic Imager (PHI), the Extreme Ultraviolet Imager (EUI), and the Spectra Imaging of the Coronal Environment (SPICE) instrument. 

«Today we reveal humankind’s first-ever views of the sun’s poles,» said Professor Carole Mundell, ESA director of science, in a blog post. «The sun is our nearest star, giver of life and potential disruptor of modern space and ground power systems, so it is imperative that we understand how it works and learn to predict its behavior. These new, unique views from our Solar Orbiter mission are the beginning of a new era of solar science.» 

See the Solar Orbiter’s stunning sun view

The images are viewable above or in a YouTube video on ESA’s channels. In the video, you can see the view that we Earthlings generally see before the video transitions to the solar orbiter’s viewpoint and zooms in so you can see the bottom of the sun in all of its hot, fiery glory. The video is only 50 seconds long, but it’s 50 seconds of footage that humans have never seen before. 

Most of the ESA’s images and videos are of the sun’s south pole, but the blog post includes imagery of the north pole as well. For the most part, scientists had no idea what to expect from the data, given that this is the first time any human had seen it before. 

The full dataset from the Orbiter’s first pole-to-pole adventure is set to reach Earth by October 2025, which will give scientists much more to work with in terms of understanding how the sun works. Future orbits will include measurements from all 10 of the Orbiter’s tools, so even more information is coming over the next few years.

A new generative AI model Meta released this week could change how machines understand the physical world, opening up opportunities for smarter robots and more, the company said. 

The new open-source model, called Video Joint Embedding Predictive Architecture 2, or V-JEPA 2, is designed to help artificial intelligence understand things like gravity and object permanence, Meta said. 

«By sharing this work, we aim to give researchers and developers access to the best models and benchmarks to help accelerate research and progress,» the company said in a blog post, «ultimately leading to better and more capable AI systems that will help enhance people’s lives.»

Current models that allow AI to interact with the physical world rely on labeled data or video to mimic reality, but this approach emphasizes the logic of the physical world, including how objects move and interact. The model could allow AI to understand concepts like the fact that a ball rolling off of a table will fall.

Meta said the model could be useful for devices like autonomous vehicles and robots by ensuring they don’t need to be trained on every possible situation. The company called it a step toward AI that can adapt like humans can. 

One struggle in the space of physical AI has been the need for significant amounts of training data, which takes time, money and resources. At SXSW earlier this year, experts said synthetic data — training data created by AI — could help prepare a more traditional learning model for unexpected situations. (In Austin, the example used was the emergence of bats from the city’s famed Congress Avenue Bridge.) 

Meta said its new model simplifies the process and makes it more efficient for real-world applications because it doesn’t rely on all of that training data. 

The next steps for world models include training models that are capable of learning, reasoning and planning across different time and space scales, making them better at breaking down complicated tasks. Multimodal models, that can use other senses like audio and touch in addition to vision, will also help future AI models understand the real world.