• How to protect yourself from phishing attacks in Chrome and Firefox
    www.zdnet.com
    Phishing is a form of social engineering where attackers try to get you to reveal your sensitive information through malicious links, SMS, QR codes, and more. Here's how to protect yourself in Chrome and Firefox.
    0 Комментарии ·0 Поделились ·127 Просмотры
  • Reolink debuts next-gen home security options with 24/7 battery recording
    www.zdnet.com
    The tech company's new releases assure high-res details, 180-degree coverage, and around-the-clock surveillance whether on or off the grid.
    0 Комментарии ·0 Поделились ·132 Просмотры
  • Nvidia Project Digits: A Linux-powered desktop for AI developers
    www.zdnet.com
    The most powerful desktop of all time will be arriving soon - and it will be running Linux.
    0 Комментарии ·0 Поделились ·130 Просмотры
  • Is The Party Over For Plastic Foam Foodware Packaging?
    www.forbes.com
    Foam cups for customers are pictured at the coffee counter of a donut shop in Monterey Park, ... [+] California on December 19, 2024. California is expected to ban the sale of plastic foam foodware (commonly known as Styrofoam) beginning January 1, 2025 , unless there is clear evidence submitted to California's Department of Resources Recycling and Recovery (CalRecycle) that it meets a 25% recycling rate in the state. This ban is part of the Plastic Pollution Prevention and Packaging Producer Responsibility Act (SB54) that Ocean Conservancy helped negotiate. (Photo by Frederic J. BROWN / AFP) (Photo by FREDERIC J. BROWN/AFP via Getty Images)AFP via Getty ImagesThere was a time when plastic foam foodware packaging - commonly known as Styrofoam - was a regular sight at music festivals, coffee shops and takeaway outlets.But as concerns grow about its environmental impact, a growing number of U.S. states are banning or restricting its use.In California, new regulations under the states SB54 Plastic Pollution Prevention and Packaging Producer Responsibility Act came into force at the beginning of this year, which effectively bans plastic foam foodware from being sold.A spokesperson for CalRecycle said the regulations stipulate in order to continue selling expanded polystyrene food service ware in the state, producers must demonstrate it meets a 25% recycling rate as of January 1, 2025 in an email.If this requirement is not met, SB54 prohibits EPS producers from selling, offering for sale, distributing, or importing EPS food service ware into California.In Oregon another set of regulations (SB 543) also came into effect on the same date (1 January), which bans plastic foam foodware, single-use coolers, and packing peanuts.Plastic foam is one of the most common forms of plastic pollution both in California and around the world.Since 1988, volunteers with Ocean Conservancys International Coastal Cleanup have removed over 750,000 plastic foam items and 700,000 tiny foam pieces from beaches and waterways in California alone.MORE FOR YOUScientists at Ocean Conservancy estimate the new regulations could eliminate up to 3.9 billion pieces of foam foodware in California every year.Ocean Conservancys director of plastic policy, Dr. Anja Brandon, said foam foodware might have up a relatively small fraction of all the plastic currently being produced, but is has a disproportionate impact on the environment in an interview.Dr. Brandon said this is because the material is super lightweight and breaks down into small foam beads, which can travel far and wide, due to tide currents and strong winds.Like other types of plastic, foam foodware is made from petrochemicals and it does not easily break down, she told me. Once it is out in the environment, it is stuck there forever.We've already seen 10 other states and the district of Columbia already pass legislation to get rid of plastic foam foodware. There are already plenty of alternatives out there. Its a no brainer to start getting rid of this material.Dr. Brandon added Ocean Conservancy would like to see more states pass legislation to ban plastic foam foodware, as well as action on a federal level.We would also hope industry would be more willing to get on board, given that the market for this material is just going to continue to shrink, she added.We really do anticipate the changes we will see in California will impact the rest of the country, and the rest of the world.You do not make one type of packaging for California and a different type for New York. That is not how producers want to do business, said Dr. Brandon.Oceanas plastics campaign director, Christy Leavitt said she expects to see additional state legislatures consider bills to phase out single-use plastic foam products in 2025 in an email.Leavitt added in most states, legislators are in the process of finalizing the bills they will introduce.According to 2023 polling data released by Oceana, she said 72% of American voters support policies to reduce the use of plastic foam foodware, packing peanuts, and coolers.Leavit said four states - California, Illinois, New York, and Washington - have also passed laws to phase out small plastic bottles for shampoo, conditioner, and other personal care products.She added Californias ban is already in place, while the New York law came into effect on January 1, 2025. Illinois will start its ban on July 1 this year, and Washington will follow in 2027.
    0 Комментарии ·0 Поделились ·131 Просмотры
  • UFC Releases Hall-Of-Famer And 3 Other Fighters From Roster
    www.forbes.com
    NEWARK, NJ - MAY 04: Dana White (Photo by Louis Grasse/PxImages/Icon Sportswire via Getty Images)Icon Sportswire via Getty ImagesThe UFCs official roster shrunk a bit more on Monday, January 7, per UFC Roster Tracker. According to the account that has accurately tracked the comings and goings of UFC fighters for nearly four years, the No. 1 MMA promotion in the world officially said goodbye to four fightersincluding Hall-of-Famer Urijah Faber.While Faber hasnt fought in the UFC since 2019, when he was finished by a head kick from Petr Yan, he has been active on the grappling scene. Faber scored a win over Jeff Glover in Combat Jiu-Jitsu in May 2024 and another win over ONE Championship legend Bibiano Fernandes in August 2024.ANAHEIM, CA - FEBRUARY 20: Urijah Faber poses for a portrait on February 20, 2013 in Anaheim, ... [+] California. (Photo by Jim Kemper/Zuffa LLC/Zuffa LLC via Getty Images)Zuffa LLC via Getty ImagesFaber also unretired in March 2024 but didnt get a fight in the UFC. At 45, Faber probably doesnt have much of a chance to compete at a high level in the UFC. However, he does seem like a fit for the newly formed Global Fight League (GFL).It would not be a surprise to see him added to the GFLs upcoming draft pool. The draft is scheduled to take place on January 24. If Fabers name is dropped into the pool, the GFL would then have another UFC Hall-of-Famer on its roster, joining Wanderlei Silva.MORE FOR YOUFaber was inducted into the UFCs Hall of Fame in 2017 following a historic career.The UFC also cut ties with Danny Hot Chocolate Roberts. Always a lock for an exciting, all-action fight, Roberts, 37, never really found his stride in the UFC. He made his promotional debut in 2015 with a submission win over Nathan Coy.LAS VEGAS, NEVADA - OCTOBER 16: Danny Roberts of England poses for a portrait after his victory ... [+] during the UFC Fight Night event at UFC APEX on October 16, 2021 in Las Vegas, Nevada. (Photo by Mike Roach/Zuffa LLC)Zuffa LLCHowever, Roberts could only muster a 7-7 record in the UFC. Six of his seven losses in the promotion came by stoppage. Roberts lost his last three fights and hadnt competed since July 2023.The UFC also parted ways with Victor Hugo. The Brazilian, only 32 years old, has struggled mightily with making weight in his brief UFC career. After earning a spot on the roster with a win on the Contender Series in October 2023, Hugo had a fight with Daniel Marcos canceled in November after he missed weight.SAO PAULO, BRAZIL - NOVEMBER 01: Victor Hugo poses for a portrait during a UFC photo session on ... [+] November 1, 2023 in Sao Paulo, Brazil. (Photo by Mike Roach/Zuffa LLC via Getty Images)Zuffa LLC via Getty ImagesIn April 2024, Hugo won a unanimous decision over Pedro Falcao. However, in October at UFC 308, Hugo was going to badly miss weight for his clash with Farid Basharat. The fight was changed to a featherweight bout, where Hugo outweighed Basharat by 7.5 pounds.Basharat still prevailed via unanimous decision. It appears the UFC chose to go in another direction after Hugos repeated weight issues.RIYADH, SAUDI ARABIA - JUNE 21: Kyung Ho Kang of South Korea poses on the scale during the UFC Fight ... [+] Night ceremonial weigh-in at the Crowne Plaza Riyadh on June 21, 2024 in Riyadh, Saudi Arabia. (Photo by Chris Unger/Zuffa LLC via Getty Images)Zuffa LLC via Getty ImagesKyung Ho Kang, 37, had a long and successful career in the UFC. However, he is on a two-fight losing streak and is nowhere near contention in the UFCs mens bantamweight division.Next for the UFC - UFC Apex 101Here is a look at the card coming up on January 11. Its the first UFC card in three weeks and is headlined by a rematch between Mackenzie Dern and Amanda Ribas. The two women fought in 2019, with Ribas winning by unanimous decision.Main Card:Mackenzie Dern vs. Amanda Ribas - StrawweightSantiago Ponzinibbio vs. Carlston Harris - WelterweightAbdul Razak Alhassan vs. Csar Almeida - MiddleweightChris Curtis vs. Roman Kopylov - MiddleweightPrelims:Austin Bashi vs. Christian Rodriguez - FeatherweightPunahele Soriano vs. Uro Medi - WelterweightJose Johnson vs. Felipe Bunes - FlyweightMarco Tlio vs. Ihor Potieria - MiddleweightThiago Moiss vs. Trey Ogden - LightweightAndreas Gustafsson vs. Preston Parsons - WelterweightErnesta Kareckaite vs. Nicolle Caliari - FlyweightMagomed Gadzhiyasulov vs. Bruno Lopes - Light HeavyweightViktoriia Dudakova vs. Fatima Kline - StrawweightNurullo Aliev vs. Yanal Ashmouz - Lightweight
    0 Комментарии ·0 Поделились ·140 Просмотры
  • UFC Champion Wants 3 Fights In 6 Months, Names Opponents
    www.forbes.com
    Dana White, UFC President (Photo by Yifan Ding/Getty Images)Getty ImagesUFC mens bantamweight champion Merab Dvalishvili wanted to take five or six months off after winning his first title in September 2024 at Noche UFC, but he plans to hit the ground running in 2025.Dvalishvili will defend his title against Umar Nurmagomedov on January 18 in Los Angeles in the co-main event of UFC 311. If Dvalishvili has his way, that will be the first of three fights he plans to have over a six-month period.LAS VEGAS, NEVADA - SEPTEMBER 14: Merab Dvalishvili of Georgia poses with his team after defeating ... [+] Sean OMalley of the United States to win the bantamweight title during UFC 306: Riyadh Season Noche at Sphere on September 14, 2024 in Las Vegas, Nevada. (Photo by Christian Petersen/Getty Images)Getty ImagesDvalishvili was a guest on Michael Bispings Believe You Me podcast, where he discussed his immediate future.Didnt you say you want three fights in six months? Bisping asked.I do. I do, Dvalishvili replied.MORE FOR YOUAnd would one of those fights be against Sean OMalley? Bisping asked. He wants a rematch, of course. Are you happy to give him a rematch?Yes, absolutely, Dvalishvili replied. First, I have to focus on Umar fight. God willing, Im going to win this fight. After my fight, I wanted to take five or six months. Five months would be in February, and six months would be in March. If OMalley isnt ready in March, then I fight Petr Yan in March. If OMalley is ready to fight in April in Miamihe says he is undefeated in Miamithen we can do it there.Here is a look at the episode:It would be nearly impossible for Dvalishvili to pull off what hes describing. You hve to wonder if Dana White and Co. would even allow him to attempt to complete this feat.All three of the opponents Dvalishvili named are extremely tough. The battle with Nurmagomedov could be a fight that goes from grappling to stand-up fighting across all five rounds.If Dvalishvili wins at UFC 311, the prospect of fighting someone as tough as Yan in March is staggering. Dvalishvili has already defeated Yan, but the rematch figures to be even more challenging.Lastly, Dvalishvili is talking about a one-month turnaround to take on OMalley in a rematch the latter will be very prepared for, and the champion could be banged up from injuries.In any case, Dvalishvili has to take things one fight at a time, and it all begins with a tough title defense against Nurmagomedov on January 18.Next Up in the UFC - UFC Vegas 101On January 11, the UFC comes back from a three-week absence with a card at the Apex in Las Vegas.The main event is Mackenzie Dern vs. Amanda Ribas in a rematch of their 2019 meeting. Ribas won the first fight, but Dern is now ranked one slot above Ribas.Here is a look at the complete card.Mackenzie Dern vs. Amanda Ribas - StrawweightSantiago Ponzinibbio vs. Carlston Harris - WelterweightAbdul Razak Alhassan vs. Csar Almeida - MiddleweightChris Curtis vs. Roman Kopylov - MiddleweightAustin Bashi vs. Christian Rodriguez - FeatherweightPunahele Soriano vs. Uro Medi - WelterweightJose Johnson vs. Felipe Bunes - FlyweightMarco Tlio vs. Ihor Potieria - MiddleweightThiago Moiss vs. Trey Ogden - LightweightAndreas Gustafsson vs. Preston Parsons - WelterweightErnesta Kareckaite vs. Nicolle Caliari - FlyweightMagomed Gadzhiyasulov vs. Bruno Lopes - Light HeavyweightViktoriia Dudakova vs. Fatima Kline - StrawweightNurullo Aliev vs. Yanal Ashmouz - Lightweight
    0 Комментарии ·0 Поделились ·139 Просмотры
  • AI Agents Explained: The Next Evolution in Artificial Intelligence
    www.techspot.com
    Interacting with AI chatbots like ChatGPT can be fun and sometimes useful, but the next level of everyday AI goes beyond answering questions: AI agents carry out tasks for you.Major technology companies, including OpenAI, Microsoft, Google and Salesforce, have recently released or announced plans to develop and release AI agents. They claim these innovations will bring newfound efficiency to technical and administrative processes underlying systems used in health care, robotics, gaming and other businesses.Simple AI agents can be taught to reply to standard questions sent over email. More advanced ones can book airline and hotel tickets for transcontinental business trips. Google recently demonstrated Project Mariner to reporters, a browser extension for Chrome that can reason about the text and images on your screen.Nvidia CEO Jensen Huang was clearly trying to tell us something as he outlined the evolution of AI technologies, from perception AI to generative AI, agentic AI, and the rise of physical AI, during his keynote at CES 2025.In the demonstration, the agent helped plan a meal by adding items to a shopping cart on a grocery chain's website, even finding substitutes when certain ingredients were not available. A person still needs to be involved to finalize the purchase, but the agent can be instructed to take all of the necessary steps up to that point.Editor's Note:Guest author Dr. Brian O'Neill is an Associate Professor of Computer Science at Quinnipiac University. Dr. O'Neill's primary research area is artificial intelligence, with a particular emphasis on the role that AI can play in supporting creativity. This article is republished from The Conversation under a Creative Commons license. In a sense, you are an agent. You take actions in your world every day in response to things that you see, hear and feel. But what exactly is an AI agent? As a computer scientist, I offer this definition: AI agents are technological tools that can learn a lot about a given environment, and then - with a few simple prompts from a human - work to solve problems or perform specific tasks in that environment.Rules and goalsA smart thermostat is an example of a very simple agent. Its ability to perceive its environment is limited to a thermometer that tells it the temperature. When the temperature in a room dips below a certain level, the smart thermostat responds by turning up the heat.A familiar predecessor to today's AI agents is the Roomba. The robot vacuum cleaner learns the shape of a carpeted living room, for instance, and how much dirt is on the carpet. Then it takes action based on that information. After a few minutes, the carpet is clean.The smart thermostat is an example of what AI researchers call a simple reflex agent. It makes decisions, but those decisions are simple and based only on what the agent perceives in that moment. The robot vacuum is a goal-based agent with a singular goal: clean all of the floor that it can access. The decisions it makes - when to turn, when to raise or lower brushes, when to return to its charging base - are all in service of that goal.A goal-based agent is successful merely by achieving its goal through whatever means are required. Goals can be achieved in a variety of ways, however, some of which could be more or less desirable than others.Many of today's AI agents are utility based, meaning they give more consideration to how to achieve their goals. They weigh the risks and benefits of each possible approach before deciding how to proceed. They are also capable of considering goals that conflict with each other and deciding which one is more important to achieve. They go beyond goal-based agents by selecting actions that consider their users' unique preferences.Making decisions, taking actionWhen technology companies refer to AI agents, they aren't talking about chatbots or large language models like ChatGPT. Though chatbots that provide basic customer service on a website technically are AI agents, their perceptions and actions are limited. Chatbot agents can perceive the words that a user types, but the only action they can take is to reply with text that hopefully offers the user a correct or informative response.The AI agents that AI companies refer to are significant advances over large language models like ChatGPT because they possess the ability to take actions on behalf of the people and companies who use them.OpenAI says agents will soon become tools that people or businesses will leave running independently for days or weeks at a time, with no need to check on their progress or results. Researchers at OpenAI and Google DeepMind say agents are another step on the path to artificial general intelligence or "strong" AI - that is, AI that exceeds human capabilities in a wide variety of domains and tasks.The AI systems that people use today are considered narrow AI or "weak" AI. A system might be skilled in one domain - chess, perhaps - but if thrown into a game of checkers, the same AI would have no idea how to function because its skills wouldn't translate. An artificial general intelligence system would be better able to transfer its skills from one domain to another, even if it had never seen the new domain before.Worth the risks?Are AI agents poised to revolutionize the way humans work? This will depend on whether technology companies can prove that agents are equipped not only to perform the tasks assigned to them, but also to work through new challenges and unexpected obstacles when they arise.Uptake of AI agents will also depend on people's willingness to give them access to potentially sensitive data: Depending on what your agent is meant to do, it might need access to your internet browser, your email, your calendar and other apps or systems that are relevant for a given assignment. As these tools become more common, people will need to consider how much of their data they want to share with them.A breach of an AI agent's system could cause private information about your life and finances to fall into the wrong hands. Are you OK taking these risks if it means that agents can save you some work?What happens when AI agents make a poor choice, or a choice that its user would disagree with? Currently, developers of AI agents are keeping humans in the loop, making sure people have an opportunity to check an agent's work before any final decisions are made. In the Project Mariner example, Google won't let the agent carry out the final purchase or accept the site's terms of service agreement. By keeping you in the loop, the systems give you the opportunity to back out of any choices made by the agent that you don't approve.Like any other AI system, an AI agent is subject to biases. These biases can come from the data that the agent is initially trained on, the algorithm itself, or in how the output of the agent is used. Keeping humans in the loop is one method to reduce bias by ensuring that decisions are reviewed by people before being carried out.The answers to these questions will likely determine how popular AI agents become, and depend on how much AI companies can improve their agents once people begin to use them.
    0 Комментарии ·0 Поделились ·136 Просмотры
  • Tencent Makes Biggest Buyback Since 2006 After Addition to Pentagon List
    www.wsj.com
    Chinas largest public company by market capitalization spent the equivalent of about $193 million to buy back 3.93 million shares on Tuesday.
    0 Комментарии ·0 Поделились ·116 Просмотры
  • When is the best time to exercise to get the most from your workout?
    www.newscientist.com
    Your muscles, fat and other cells respond differently to exercise depending on the time of dayPanoramic Images/AlamyUsain Bolt smashed the 100-metre sprint record at the 2009 World Championships in Athletics in Berlin in a floodlit stadium under an inky night sky.This article is part of a series on fitness that answers eight questions about exercise and its influence on our bodies and minds. Read more here.This was no coincidence: when it comes tosporting prowess, timing matters. For activities relying on sheer muscle power andendurance, evening or late afternoon arewhenmost world records have been set, probably due to assistance frompeaks in thedaily rhythms ofanumber of the bodys key physiological outputs.But what about everyday exercisers? Is therean optimal timeof day, or month, togetthe most out ofworkouts and reduce therisk of injury?AdvertisementWhichever aspect of sport you look at whether its sports medicine or exercise response time of day matters, says Qing-Jun Meng at the University of Manchester, UK. Later afternoon or evening is when our bodytemperature peaks, resulting in fastermetabolic reactions and nerve signal transmission compared with the early morning. Connective tissue is also more flexible in the afternoon, while our reserves of glycogen the biochemical energy source our muscles draw on during moderate to intense exercise have had time to be replenished.Circadian rhythmsOther physiological parameters also varyover the 24-hour period, which could berelevant for exercise: testosterone secretion peaks at around 9am; coordination tends tobebest at around 2.30pm; reaction times arefastest at around 3.30pm; cardiovascular efficiency, muscle and grip strength peak at 5to5.30pm.Circadian rhythms exist in almost all cellsof the body and regulate key processes relating to exercise and metabolism, says Meng. Depending on what time you exercise, your muscle, fat and other cells will also be inadifferent state, and they will respond to exercise differently.Indeed, a recent study by Renske Lok at Stanford University in California and her colleagues found that, on average, Olympic swimmers are more than athird of a second faster if they compete in theevening rather than in the morning. In40per cent of [swimming] races, the time-of-day effect isbigger than the difference between finishingfirst or second, they wrote.For sports involving more technical skills, such as tennis or soccer, peak performance tends to arrive a little earlier possibly because our cognitive abilities usually peak in the late morning or early afternoon. Soccer players juggle and chip the ball with the greatest precision at around 4pm; tennis players servestend to be faster in the evening, but more accurate in the morning. Of course, thesetimings are based on averages in reality,larks who tend to wake early and bemore active in the mornings will be at theirbest earlier, while owls, who naturally come to life in the evenings, will peak later.What are the implications of all this for howthe rest of us time our exercise? Last year, Fabienne Bruggisser at the University of Basel in Switzerland and her colleagues pored over the evidence from 26 previous studies and found little to support or refute the idea that training at a specific time leads to better performance or improved health outcomes.They did, however, find some evidence tosupport training at the same time of day asarace or competition to improve physical performance at this time. In other words, morning training improves morning performance more than evening training does,and vice versa. However, given that thestudies only included young male participants, it remains to be seen whether such conclusions apply to the general population, the authors said.Monthly cyclesWomen may have a further layer of complexity to consider. In recent years, several womens soccer teams, including Chelsea FC Women, have started tailoring their players training programmes around their menstrual cycles, claiming that doing so boosts their performance and reduces their risk of injury.The theory is that when oestrogen is high and progesterone is low thats an anabolic environment; its a good [time] to work hard, says Stuart Phillips at McMaster University in Canada. Yet when he and his colleagues recently reviewed the evidence for an influenceof menstrual cycle phase on exercise performance, they found it was remarkably thin. The evidence that we do have suggests that theres no merit to it, says Phillips.Even so, he doesnt dismiss recording symptoms and using them as a guide to scheduling training. I know some women aregenuinely adversely affected at certain phases of their cycle with menstrual-related symptoms: cramps, backaches, a lack of motivation, fatigue, etc. And for women athletes that do experience symptoms, tryingto manage them and making their coachaware of them is a great thing, he says.But as for a blanket, this-is-the-way-you-do-things-type approach, theres no consistent pattern to performance when its studied systematically, and we know gold medals andworld records have been set at different [menstrual] phases, and on and off the contraceptive pill.Meng believes there may be yet other factors to consider. Generally, he advocates exercising in the morning especially outdoors as this exposes people to bright light, which helps to synchronise our biological clocks with the time of day. Our bodies work best when the clocks inall our cells and tissues are aligned with oneanother, and with the time of day.And even if there isnt a best time of daytowork out, there may be a time to avoidexercise. Recent research by Meng andhis colleagues has suggested that a key mechanism for keeping the biological clocks inour bones and joints synchronised with those in other tissues is exercise, and that if mice are encouraged to exercise when they would usually be sleeping, this causes their skeletal clocks to desynchronise from their brain clocks a phenomenon Meng has namedskeletal jet lag.Though the implications for injury and physical performance in humans are unclear, further experiments in mice suggest that consistently exercising during their equivalent of night-time leads to the activation of genes that are associated with osteoarthritis. We suspect that if you did this over a long period of time, it could be really detrimental, says Meng.Topics:
    0 Комментарии ·0 Поделились ·113 Просмотры
  • Whats next for AI in 2025
    www.technologyreview.com
    MIT Technology Reviews Whats Next series looks across industries, trends, and technologies to give you a first look at the future. You can read the rest of themhere. For the last couple of years weve had a go at predicting whats coming next in AI. A fools game given how fast this industry moves. But were on a roll, and were doing it again. How did we score last time round? Our four hot trends to watch out for in 2024 included what we called customized chatbotsinteractive helper apps powered by multimodal large language models (check: we didnt know it yet, but we were talking about what everyone now calls agents, the hottest thing in AI right now); generative video (check: few technologies have improved so fast in the last 12 months, with OpenAI and Google DeepMind releasing their flagship video generation models, Sora and Veo, within a week of each other this December); and more general-purpose robots that can do a wider range of tasks (check: the payoffs from large language models continue to trickle down to other parts of the tech industry, and robotics is top of the list). We also said that AI-generated election disinformation would be everywhere, but herehappilywe got it wrong. There were many things to wring our hands over this year, but political deepfakes were thin on the ground. So whats coming in 2025? Were going to ignore the obvious here: You can bet that agents and smaller, more efficient, language models will continue to shape the industry. Instead, here are five alternative picks from our AI team. 1. Generative virtual playgrounds If 2023 was the year of generative images and 2024 was the year of generative videowhat comes next? If you guessed generative virtual worlds (a.k.a. video games), high fives all round. We got a tiny glimpse of this technology in February, when Google DeepMind revealed a generative model called Genie that could take a still image and turn it into a side-scrolling 2D platform game that players could interact with. In December, the firm revealed Genie 2, a model that can spin a starter image into an entire virtual world. Other companies are building similar tech. In October, the AI startups Decart and Etched revealed an unofficial Minecraft hack in which every frame of the game gets generated on the fly as you play. And World Labs, a startup cofounded by Fei-Fei Licreator of ImageNet, the vast data set of photos that kick-started the deep-learning boomis building what it calls large world models, or LWMs. One obvious application is video games. Theres a playful tone to these early experiments, and generative 3D simulations could be used to explore design concepts for new games, turning a sketch into a playable environment on the fly. This could lead to entirely new types of games. But they could also be used to train robots. World Labs wants to develop so-called spatial intelligencethe ability for machines to interpret and interact with the everyday world. But robotics researchers lack good data about real-world scenarios with which to train such technology. Spinning up countless virtual worlds and dropping virtual robots into them to learn by trial and error could help make up for that. Will Douglas Heaven 2. Large language models that reason The buzz was justified. When OpenAI revealed o1 in September, it introduced a new paradigm in how large language models work. Two months later, the firm pushed that paradigm forward in almost every way with o3a model that just might reshape this technology for good. Most models, including OpenAIs flagship GPT-4, spit out the first response they come up with. Sometimes its correct; sometimes its not. But the firms new models are trained to work through their answers step by step, breaking down tricky problems into a series of simpler ones. When one approach isnt working, they try another. This technique, known as reasoning (yeswe know exactly how loaded that term is), can make this technology more accurate, especially for math, physics, and logic problems. Its also crucial for agents. In December, Google DeepMind revealed an experimental new web-browsing agent called Mariner. In the middle of a preview demo that the company gave to MIT Technology Review, Mariner seemed to get stuck. Megha Goel, a product manager at the company, had asked the agent to find her a recipe for Christmas cookies that looked like the ones in a photo shed given it. Mariner found a recipe on the web and started adding the ingredients to Goels online grocery basket. Then it stalled; it couldnt figure out what type of flour to pick. Goel watched as Mariner explained its steps in a chat window: It says, I will use the browsers Back button to return to the recipe. It was a remarkable moment. Instead of hitting a wall, the agent had broken the task down into separate actions and picked one that might resolve the problem. Figuring out you need to click the Back button may sound basic, but for a mindless bot its akin to rocket science. And it worked: Mariner went back to the recipe, confirmed the type of flour, and carried on filling Goels basket. Google DeepMind is also building an experimental version of Gemini 2.0, its latest large language model, that uses this step-by-step approach to problem solving, called Gemini 2.0 Flash Thinking. But OpenAI and Google are just the tip of the iceberg. Many companies are building large language models that use similar techniques, making them better at a whole range of tasks, from cooking to coding. Expect a lot more buzz about reasoning (we know, we know) this year. Will Douglas Heaven 3. Its boom time for AI in science One of the most exciting uses for AI is speeding up discovery in the natural sciences. Perhaps the greatest vindication of AIs potential on this front came last October, when the Royal Swedish Academy of Sciences awarded the Nobel Prize for chemistry to Demis Hassabis and John M. Jumper from Google DeepMind for building the AlphaFold tool, which can solve protein folding, and to David Baker for building tools to help design new proteins. Expect this trend to continue next year, and to see more data sets and models that are aimed specifically at scientific discovery. Proteins were the perfect target for AI, because the field had excellent existing data sets that AI models could be trained on. The hunt is on to find the next big thing. One potential area is materials science. Meta has released massive data sets and models that could help scientists use AI to discover new materials much faster, and in December, Hugging Face, together with the startup Entalpic, launched LeMaterial, an open-source project that aims to simplify and accelerate materials research. Their first project is a data set that unifies, cleans, and standardizes the most prominent material data sets. AI model makers are also keen to pitch their generative products as research tools for scientists. OpenAI let scientists test its latest o1 model and see how it might support them in research. The results were encouraging. Having an AI tool that can operate in a similar way to a scientist is one of the fantasies of the tech sector. In a manifesto published in October last year, Anthropic founder Dario Amodei highlighted science, especially biology, as one of the key areas where powerful AI could help. Amodei speculates that in the future, AI could be not only a method of data analysis but a virtual biologist who performs all the tasks biologists do. Were still a long way away from this scenario. But next year, we might see important steps toward it. Melissa Heikkil 4. AI companies get cozier with national security There is a lot of money to be made by AI companies willing to lend their tools to border surveillance, intelligence gathering, and other national security tasks. The US military has launched a number of initiatives that show its eager to adopt AI, from the Replicator programwhich, inspired by the war in Ukraine, promises to spend $1 billion on small dronesto the Artificial Intelligence Rapid Capabilities Cell, a unit bringing AI into everything from battlefield decision-making to logistics. European militaries are under pressure to up their tech investment, triggered by concerns that Donald Trumps administration will cut spending to Ukraine. Rising tensions between Taiwan and China weigh heavily on the minds of military planners, too. In 2025, these trends will continue to be a boon for defense-tech companies like Palantir, Anduril, and others, which are now capitalizing on classified military data to train AI models. The defense industrys deep pockets will tempt mainstream AI companies into the fold too. OpenAI in December announced it is partnering with Anduril on a program to take down drones, completing a year-long pivot away from its policy of not working with the military. It joins the ranks of Microsoft, Amazon, and Google, which have worked with the Pentagon for years. Other AI competitors, which are spending billions to train and develop new models, will face more pressure in 2025 to think seriously about revenue. Its possible that theyll find enough non-defense customers who will pay handsomely for AI agents that can handle complex tasks, or creative industries willing to spend on image and video generators. But theyll also be increasingly tempted to throw their hats in the ring for lucrative Pentagon contracts. Expect to see companies wrestle with whether working on defense projects will be seen as a contradiction to their values. OpenAIs rationale for changing its stance was that democracies should continue to take the lead in AI development, the company wrote, reasoning that lending its models to the military would advance that goal. In 2025, well be watching others follow its lead. James ODonnell 5. Nvidia sees legitimate competition For much of the current AI boom, if you were a tech startup looking to try your hand at making an AI model, Jensen Huang was your man. As CEO of Nvidia, the worlds most valuable corporation, Huang helped the company become the undisputed leader of chips used both to train AI models and to ping a model when anyone uses it, called inferencing. A number of forces could change that in 2025. For one, behemoth competitors like Amazon, Broadcom, AMD, and others have been investing heavily in new chips, and there are early indications that these could compete closely with Nvidiasparticularly for inference, where Nvidias lead is less solid. A growing number of startups are also attacking Nvidia from a different angle. Rather than trying to marginally improve on Nvidias designs, startups like Groq are making riskier bets on entirely new chip architectures that, with enough time, promise to provide more efficient or effective training. In 2025 these experiments will still be in their early stages, but its possible that a standout competitor will change the assumption that top AI models rely exclusively on Nvidia chips. Underpinning this competition, the geopolitical chip war will continue. That war thus far has relied on two strategies. On one hand, the West seeks to limit exports to China of top chips and the technologies to make them. On the other, efforts like the US CHIPS Act aim to boost domestic production of semiconductors. Donald Trump may escalate those export controls and has promised massive tariffs on any goods imported from China. In 2025, such tariffs would put Taiwanon which the US relies heavily because of the chip manufacturer TSMCat the center of the trade wars. Thats because Taiwan has said it will help Chinese firms relocate to the island to help them avoid the proposed tariffs. That could draw further criticism from Trump, who has expressed frustration with US spending to defend Taiwan from China. Its unclear how these forces will play out, but it will only further incentivize chipmakers to reduce reliance on Taiwan, which is the entire purpose of the CHIPS Act. As spending from the bill begins to circulate, next year could bring the first evidence of whether its materially boosting domestic chip production. James ODonnell
    0 Комментарии ·0 Поделились ·131 Просмотры