Toggle light / dark theme

Large language models (LLMs), such as the GPT-4 model underpinning the widely used conversational platform ChatGPT, have surprised users with their ability to understand written prompts and generate suitable responses in various languages. Some of us may thus wonder: are the texts and answers generated by these models so realistic that they could be mistaken for those written by humans?

Astronomers at MIT, NASA, and elsewhere have a new way to measure how fast a black hole spins, by using the wobbly aftermath from its stellar feasting.

The method takes advantage of a black hole tidal disruption event—a blazingly bright moment when a black hole exerts tides on a passing star and rips it to shreds. As the star is disrupted by the black hole’s immense tidal forces, half of the star is blown away, while the other half is flung around the black hole, generating an intensely hot accretion disk of rotating stellar material.

The MIT-led team has shown that the wobble of the newly created accretion disk is key to working out the central black hole’s inherent spin.

Centenarians, once considered rare, have become commonplace. Indeed, they are the fastest-growing demographic group of the world’s population, with numbers roughly doubling every ten years since the 1970s.

How long humans can live, and what determines a long and healthy life, have been of interest for as long as we know. Plato and Aristotle discussed and wrote about the ageing process over 2,300 years ago.

The pursuit of understanding the secrets behind exceptional longevity isn’t easy, however. It involves unravelling the complex interplay of genetic predisposition and lifestyle factors and how they interact throughout a person’s life.

Physicists have delved deeper into the enigmatic world of quantum entanglement and top quarks, bringing a new level of understanding to a phenomenon that even Albert Einstein found perplexing.

This incredible feat has the potential to revolutionize our understanding of the quantum realm and its far-reaching implications.

The experiment, conducted by a team of researchers led by University of Rochester physics professor Regina Demina at the European Center for Nuclear Research (CERN), has yielded a significant result.

Researchers from Tohoku University and Kyoto University have successfully developed a DNA-based molecular controller that autonomously directs the assembly and disassembly of molecular robots. This pioneering technology marks a significant step towards advanced autonomous molecular systems with potential applications in medicine and nanotechnology.

Details of the breakthrough were published in the journal Science Advances (“Autonomous assembly and disassembly of gliding molecular robots regulated by a DNA-based molecular controller”).

“Our newly developed molecular controller, composed of artificially designed DNA molecules and enzymes, coexists with molecular robots and controls them by outputting specific DNA molecules,” points out Shin-ichiro M. Nomura, an associate professor at Tohoku University’s Graduate School of Engineering and co-author of the study. “This allows the molecular robots to self-assemble and disassemble automatically, without the need for external manipulation.”

Internet data scraping is one of the biggest fights in AI right now. Tech companies argue that anything on the public internet is fair game, but they are facing a barrage of lawsuits over their data practices and copyright. It will likely take years until clear rules are in place.

In the meantime, they are running out of training data to build even bigger, more powerful models, and to Meta, your posts are a gold mine.

If you’re uncomfortable with having Meta use your personal information and intellectual property to train its AI models in perpetuity, consider opting out. Although Meta does not guarantee it will allow this, it does say it will “review objection requests in accordance with relevant data protection laws.”

From king’s college london, carnegie mellon, & U birmingham.

Llm-driven robots risk enacting discrimination, violence, and unlawful actions.

Rumaisa Azeem, Andrew Hundt, Masoumeh Mansouri, Martim Brandão June 2024 Paper: https://arxiv.org/abs/2406.08824 Code: https://github.com/SepehrDehdashtian/


The data and code for paper ‘The Dark Side of Dataset Scaling: Evaluating Racial Classification in Multimodal Models’ — SepehrDehdashtian/the-dark-side-of-dataset-scaling.