Toggle light / dark theme

Today, we’re excited to share V-JEPA 2, the first world model trained on video that enables state-of-the-art understanding and prediction, as well as zero-shot planning and robot control in new environments. As we work toward our goal of achieving advanced machine intelligence (AMI), it will be important that we have AI systems that can learn about the world as humans do, plan how to execute unfamiliar tasks, and efficiently adapt to the ever-changing world around us.

V-JEPA 2 is a 1.2 billion-parameter model that was built using Meta Joint Embedding Predictive Architecture (JEPA), which we first shared in 2022. Our previous work has shown that JEPA performs well for modalities like images and 3D point clouds. Building on V-JEPA, our first model trained on video that we released last year, V-JEPA 2 improves action prediction and world modeling capabilities that enable robots to interact with unfamiliar objects and environments to complete a task. We’re also sharing three new benchmarks to help the research community evaluate how well their existing models learn and reason about the world using video. By sharing this work, we aim to give researchers and developers access to the best models and benchmarks to help accelerate research and progress—ultimately leading to better and more capable AI systems that will help enhance people’s lives.

Planning for a future of intelligent robots means thinking about how they might transform your industry, what it means for the future of work, and how it may change the relationship between humans and technology.

Leaders must consider the ethical issues of cognitive manufacturing such as job disruption and displacement, accountability when things go wrong, and the use of surveillance technology when, for example, robots use cameras working alongside humans.

The cognitive industrial revolution, like the industrial revolutions before it, will transform almost every aspect of our world, and change will happen faster and sooner than most expect. Consider for a moment, what will it take for each of us and our organizations to be ready for this future?

One of the biggest mysteries of evolution is how species first developed complex vision. Jellyfish are helping scientists solve this puzzle, as the group has independently evolved eyes at least nine separate times. Different species of jellyfish have strikingly different types of vision, from simple eyespots that detect light intensity to sophisticated lens eyes similar to those in humans.

Biologists have studied jellyfish eye structure, light sensitivity, and visual behavior for decades, but the exact genes involved in jellyfish eye formation remain unknown.

Aide Macias-Muñoz, a professor of ecology and , is exploring how eyes and light detection evolved using genetic tools. Her lab has just completed a high-quality genome sequence of Bougainvillia cf. muscus, a small jellyfish-like animal in the Hydrozoa group that has an astonishing 28 eyes.

We’ve questioned that model and tackled questions from a different angle – by looking inward instead of outward.

Instead of starting with an expanding universe and asking how it began, we considered what happens when an over-density of matter collapses under gravity.

Prof Gaztanaga explained that the theory developed by his team of researchers worked within the principles of quantum mechanics and the model could be tested scientifically.

Astronomers from the International Center for Radio Astronomy Research (ICRAR), in collaboration with international teams, have made a startling discovery about a new type of cosmic phenomenon.

The object, known as ASKAP J1832-0911, emits pulses of radio waves and X-rays for two minutes every 44 minutes.

The paper, “Detection of X-ray Emission from a Bright Long-Period Radio Transient,” is published in Nature.

On planet Earth, fire is usually a transient phenomenon—even the strongest of wildfires will eventually succumb to human and/or meteorological intervention. But the same can’t be said for the Darvaza gas crater in Turkmenistan, known colloquially as the “Gates of Hell.” This natural gas field has been burning continuously for decades thanks to its steady supply of seeping methane, and in that time, this devilish pit has become one of the country’s most popular tourist attractions despite its location in the middle of the Karakum desert, roughly 160 miles north of the capital city of Ashgabat.

Turkmenistan’s authoritarian leader, Gurbanguly Berdymukhamedov, has previously stated that he wants to snuff out the Gates of Hell once and for all (though the latest pronouncement came years after he reportedly went off-roading around the crater). Now, a new report suggests those efforts may be bearing fruit. Last Thursday, officials in Turkmenistan said that gas being emitted from the pit has diminished three-fold, though the Agence France-Presse (AFP) reports that no timeframe for this gaseous decrease was provided. This news is in line with previous reports last year that satellite observations of the Gates of Hell showed a 50 percent decline in emissions.