Mumbai to host Nobel Prize Dialogue in Nov
Nobel Prize laureates David MacMillan (chemistry, 2021) and
James Robinson
(economic sciences, 2024) will join experts from science, business and policymaking for the dialogues to be held in Bengaluru and Mumbai, focusing on the theme 'The Future We Want'.
—Pushkraj Vernekar
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Hans India
2 hours ago
- Hans India
AI Godfather Geoffrey Hinton Urges Caution as Tech Giants Downplay AI Risks
Renowned artificial intelligence pioneer Geoffrey Hinton, often called the 'Godfather of AI,' has voiced strong concerns over the rapid and unchecked evolution of artificial intelligence systems. In a recent episode of the One Decision podcast, Hinton criticized major tech companies for publicly downplaying the serious risks associated with advanced AI development. "Many of the people in big companies, I think, are downplaying the risk publicly," Hinton stated during the conversation. He singled out Demis Hassabis, CEO of DeepMind, as one of the few leaders in the field who 'really do understand the risks and really want to do something about it.' Hinton's remarks come at a time when AI capabilities are accelerating at an extraordinary pace. He warned that these systems are not only becoming increasingly intelligent but are also starting to learn in ways that humans can't fully comprehend. 'The rate at which they've started working now is way beyond what anybody expected,' he said. In 2024, Hinton was awarded the Nobel Prize in Physics along with John J. Hopfield for their groundbreaking work on artificial neural networks, which laid the foundation for today's generative AI and deep learning technologies. Yet, Hinton now finds himself reflecting on the unintended consequences of these advancements. 'I should have realized much sooner what the eventual dangers were going to be,' Hinton admitted. 'I always thought the future was far off and I wish I had thought about safety sooner.' Hinton spent over a decade at Google before stepping down in 2023. His resignation sparked widespread speculation that he was protesting against the company's aggressive AI strategies. However, he set the record straight in the podcast, saying the media narrative around his departure was overblown. 'There's a wonderful story that the media loves — this honest scientist who wanted to tell the truth so I had to leave Google. It's a myth,' Hinton explained. 'I left Google because I was 75 and I couldn't program effectively anymore, but when I left, maybe I could talk about all these risks more freely.' He further elaborated on the challenges of remaining objective while being employed by a tech giant. 'You can't take their money and then not be influenced by what's in their own interest,' he said. In the same discussion, Hinton praised Demis Hassabis, founder of DeepMind — acquired by Google in 2014 — for his commitment to AI safety. Hassabis, who now leads AI research at Google, has frequently raised alarms about the misuse of powerful AI tools. In an earlier interview with CNN, Hassabis expressed his own worries about AI, not in terms of job displacement, but in regard to potential misuse by malicious actors. 'A bad actor could repurpose those same technologies for a harmful end,' Hassabis said. 'And so one big thing is how do we restrict access to these systems, powerful systems, to bad actors but enable good actors to do many, many amazing things with it?' As the AI landscape continues to evolve, Hinton's message is clear — greater transparency, responsibility, and proactive safety measures are critical to ensuring AI serves humanity without unforeseen consequences.
India Today
2 hours ago
- India Today
AI godfather warns tech giants are downplaying AI risks, says only DeepMind's Demis Hassabis gets it
Geoffrey Hinton, widely known as the 'Godfather of AI,' is raising concerns about the rapid development of AI and the fact that major technology companies are downplaying its dangers. Speaking on the One Decision podcast, Hinton said many corporate leaders are aware of the risks but are avoiding taking meaningful action. 'Many of the people in big companies, I think, are downplaying the risk publicly,' Hinton said. 'People like Demis, for example, really do understand the risks and really want to do something about it.'advertisementHinton, who is also a Nobel Laureate, was awarded the 2024 Nobel Prize in Physics alongside John J. Hopfield for their work on artificial neural networks. In fact Hinton's decade-long research paved the way for today's rapid advancement in artificial intelligence. However, he is now warning that advanced AI systems are becoming smarter and smarter and are even learning in ways humans don't fully understand. 'The rate at which they've started working now is way beyond what anybody expected,' he also admitted that he regrets not recognizing these dangers earlier in his career: 'I should have realized much sooner what the eventual dangers were going to be. I always thought the future was far off and I wish I had thought about safety sooner.' Hinton left Google in 2023 after more than a decade at the company. His departure was widely interpreted as his protest against its aggressive AI push. However in the podcast Hinton clarified that this narrative was not true and instead exaggerated. 'There's a wonderful story that the media loves this honest scientist who wanted to tell the truth so I had to leave Google. It's a myth,' Hinton said. 'I left Google because I was 75 and I couldn't program effectively anymore but when I left, maybe I could talk about all these risks more freely.'He added that staying at Google would have inevitably meant some level of self-censorship. 'You can't take their money and then not be influenced by what's in their own interest,' he the podcast, Hinton also spoke about Demis Hassabis and praised hims as one of the few leaders who 'really wants to do something about' the risks of AI. Hassabis, who sold DeepMind to Google in 2014 now heads its AI research arm. While he talks about their development, he has also long expressed concern about the potential misuse of advanced AI this year, in an interview with CNN, Hassabis admitted he is worried about AI. But he said he is less concerned about AI replacing jobs and more focused on the possibility that the technology could fall into the wrong hands.'A bad actor could repurpose those same technologies for a harmful end,' Hassabis told CNN's Anna Stewart. 'And so one big thing is how do we restrict access to these systems, powerful systems, to bad actors but enable good actors to do many, many amazing things with it?'- Ends
The Hindu
9 hours ago
- The Hindu
New microscope reveals molecular jostling faster than ever before
More than a century ago, a 26-year-old Albert Einstein explained Brownian motion in one of four papers he published in his annus mirabilis, the miraculous year, called because these papers shot him to fame. Brownian motion is the random jittering of small particles in a fluid, caused because they're constantly colliding with molecules around them. Now, scientists at the California Institute of Technology (Caltech) have developed a breakthrough imaging technique that enables real-time filming of these molecular motions. Their findings were published in Nature Communications. 'Surreal experience' Conventional microscopes are invasive and have limited fields of view. Other microscopes still can't distinguish individual molecules, which are around tens of angstroms in size (1 angstrom = 0.0000000001 m). To compare, one human hair is about a million angstrom thick. The Caltech team has now found a way to indirectly detect molecules by observing their interactions with light. Their technique also taps into the Brownian motion of particles. Using the device they have reported that they can see down to tens of angstroms. 'It was a surreal experience to visualise molecular sizes in real-time at the angstrom scale,' Yogeshwar Nath Mishra, who co-led the study when at Caltech's Jet Propulsion Laboratory and who is now an assistant professor at IIT-Jodhpur, said. 'Even more remarkable was the realisation that no existing technique can achieve this level of detail.' Need for speed The more massive a particle, the slower its Brownian motion. '[It] is like watching how much a spinning object twists after being nudged by light. Small molecules spin fast and scramble the light more. Big molecules spin slowly and keep it aligned,' Lihong Wang, director of the Caltech Optical Imaging Laboratory and who supervised the study, said. So by measuring how fast a molecule changes the properties of light, they could determine its size. The Egyptian-American chemist Ahmed Zewail from Caltech was the first to measure particle motion at super-short time scales. This work allowed his team to observe chemical reactions as they happened for the first time. He was awarded the Nobel Prize for chemistry in 1999. 'While traditional techniques often rely on time-consuming point-by-point scanning, our approach captures the scene in a single shot,' Wang said. 'We also achieved imaging speeds of hundreds of billions of frames per second, making it possible to observe molecular interactions in unprecedented slow motion.' The device is thus the world's fastest single-shot microscope. 'Finally, unlike [traditional methods] which require extensive sample preparation and often damage the specimen, our method is non-intrusive, enabling direct, in-situ measurements,' Wang added. 'Some of the most exciting features of this microscope include its wide-field imaging capability, offering an image area of a few square centimetres, an order of magnitude larger than conventional microscopes,' per Mishra. 'To the best of our knowledge, our work is the first ever to achieve the feat of single-shot 2D molecular sizing.' Playing jigsaw They tested their microscope using a molecule called fluorescein-dextran. Fluorescein is a food colouring dye. Fluorescein-dextran is used to monitor blood flow, drug delivery, and tissue and cell labelling. These fluorescent molecules come in the form of powders. The scientists blended them with water and used clean pipettes to pour drops of these samples into cuvettes (clear, short, rectangular tubes for holding liquid samples). Then they turned to ultrashort pulses from a laser. These lasers aren't unlike those used in LASIK and cataract surgeries. The laser sheet slices through the sample in the cuvette. As it does, the sample emits light that falls on an array of small square mirrors making up a digital micromirror device (DMD). The DMD's job is to shape the light beam. Researchers use software code to tilt each individual mirror in this light-crafter depending on the corresponding pixel in the input image. 'Imagine you're trying to solve a jigsaw puzzle, but instead of having all the pieces, you only have a few of them — and surprisingly, you can still figure out what the full picture looks like,' Wang said. This idea underpins the team's technique, which can reconstruct the full picture from very few measurements provided the structure is repetitive. The DMD converts the transient scene into a random jigsaw pattern from which researchers can extract information about the full picture. The light finally passes through a streak tube that converts the photons in light to electrons. A phosphor screen collects these electrons as they sweep across it and creates a pattern of streaks. The streak pattern reveals the pulse duration from which scientists can infer the sizes of the molecules. Ensemble of molecules 'It is an interesting piece of work. The key in this work is the use of the streak camera to detect dynamics in nanoseconds. This is within the actual lifetimes of the molecules and wouldn't be possible with slow detectors or photodetectors,' Basudev Roy, an associate professor at IIT Madras who works on super-resolution microscopy and wasn't involved in the recent study, said. The size of molecules measured using their technique concurred with previous estimates. 'It still sees an ensemble of molecules inside a detection region — it still doesn't see a single molecule yet. But the dynamics indicate chemical compositions and also chemical reactions,' Roy said. 'Surprisingly, we found out that the technique also works in gas phases. … Initially, we assumed it would be challenging to apply [it] in turbulent environments, such as within a flame,' said study co-lead Peng Wang of Caltech. The team observed black carbon nanoparticles in flames through the microscope. 'Our data in the gas phase turned out to work excellently and the molecule size matches … experimental observation well,' Peng said. This new imaging technique could help better visualise processes and transform biomedical research, disease detection, drug design, and nanomaterial fabrication, among others. Unnati Ashar is a freelance science journalist.
