 |
|
|
|||||
Welcome to AI Decoded, Fast Companys weekly newsletter that breaks down the most important news in the world of AI. You can sign up to receive this newsletter every week here. A cooler-headed look at Apples troubled AI story Apple said at its WWDC developer conference Monday that the much-hyped Apple Intelligence features it announced at last years event are still not ready to shipand likely wont be until 2026. Now, the company is living with a media narrative that its failed, fallen behind, stuck in neutral, or in retreat on AI. To be clear, Apples various operating systems already include AI-powered features, from photo cleanup to call screening. And the company did deliver some of the Apple Intelligence tools it promised last year. But those were first-generation large language model functions like text message summarization and writing assistanceunambitious by generative AI standards and not especially useful. What got everyone excited at WWDC 2024 were the advanced, personalized AI features that allow Siri to leverage a users personal data and carry out actions on their behalf. Some examples: A user could say, Hey Siri, when does Moms flight land? and Sirihaving access to the users email, contacts, calendar, and fileswould know the answer. Siri would also be able to see whats on the users screen, so it could, for example, read an address from a text message and add it to a contact card. A user could tell Siri to take action within an app, such as: Siri, take the red-eye out of this photo, or Send the email I drafted to Judy and Carmen. Cool stuff, useful, and things Apple is uniquely positioned to offer. So what happened? Apple software chief Craig Federighi and head of marketing Greg “Joz” Joswiak offered a fuller explanation during a round of media interviews Tuesday. According to them, Apple had a working version one of a new Siri architecture before announcing the personal AI features at WWDC 2024. (In other words: No, the demos werent vaporware.) The Siri team believed they could continue polishing the product until it performed reliably enough for a late 2024 release. But Federighi explained in an interview with TechRadar that the system continued to generate too many unreliable results in internal testing. When a year-end release became unrealistic, the team aimed for spring 2025. But by then, the system was still inconsistent. Thats when the Siri team concluded it would need to create a version two of the new Siri architectureone that extends across the entire Siri experience. Until that reworked Siri is performing reliably in-house, Federighi said, the company wont speculate on a public release date. Federighi also argued that Apple is trying to do something with personalized AI that nobody else has really done. Theres some truth to thatespecially if we limit the field to mobile-based AI assistants, and when were talking about doing it at Apples scale. Apple may look like an AI laggard now, but it would look far worse if it shipped something that didnt work or wasnt useful. Still, its not accurate to say Apple is the only one in the game. In May, Google announced a suite of personalized AI features that will begin rolling out this summer. Its Gemini assistant can use personal data in email to offer summaries, draft messages, provide reminders, and deliver context-aware insights. Gemini Live can analyze objects in the camera view and on the screen. And the agent mode in the Gemini app, built on Project Mariner, can perform multi-step tasks for users both online and within apps. Apples big mistake, of course, was announcing personal AI features in 2024 that it couldnt deliver in 2024or even 2025. That misstep might not have happened if this were just another set of traditional software updates. But were not in that era anymore. Apple is now operating in the realm of probabilistic computingthe domain of AI models. The usual rules for predicting software maturity and release timelines may no longer apply. Meta to rebrand itself MetaGPT (just kidding) Okay, forgive the Onion-esque headline, but it holds a seed of truth. Once again, Mark Zuckerberg is spending billions to push his company to the front of the line in the next big thing in consumer techthis time, its generative AI. (A few years ago, Facebook renamed itself Meta when it bet that the metaverse was the next big thing.) Now, The New York Times and Bloomberg report that the social media giantafter seeing its Llama models lag behind those from OpenAI, Google, and Anthropicis launching a new generative AI lab with ambitions of achieving superintelligence. (OpenAI, Anthropic, and others are pursuing artificial general intelligence, where AI matches or exceeds human abilities across a wide range of tasks. Superintelligence refers to AI that vastly outperforms humans at many tasks.) Zuckerberg has reportedly agreed to buy a 49% stake in Alexander Wangs Scale AIa company that specializes in human-labeled and synthetic training data for AI modelsfor nearly $15 billion. This appears to be another acquihire, where Metas primary interest is Wang himself. The 28-year-old will lead the new superintelligence group at Meta. (Microsoft made a similar move when it acquired Inflection AI and brought on its founder and CEO, Mustafa Suleyman.) Meta is also offering seven to nine figure salaries to attract top researchers to the new lab. The company already has a major AI figure in Yann LeCun, who leads its core AI research group. But The Times reports that Meta has experienced significant internal friction over how to approach the development and deployment of AI models. The companys focused AI efforts began in 2013 after it failed in a bid to acquire DeepMind, which ultimately went to Google. Unlike OpenAI and Anthropic, Meta doesnt rely on selling AI models or apps. Its primary business is advertising. Thanks to its massive ad revenue, Meta can afford to open-source its AI models, effectively giving developers free access in hopes of flooding the ecosystem and becoming dominant through ubiquity. But if Zuckerberg succeeds in getting Meta to the front of the superintelligence race, that strategy could shift. The company might begin locking down itsmodels and charging for accessjust like Google, OpenAI, and Anthropic. Hollywood is going after Midjourney, and it could be just the start Over the past couple of years, a range of content ownersnews organizations, authors, artists, comics, and record labelshave filed lawsuits against generative AI companies. Their claim: these companies used copyrighted content, usually scraped from the internet, to train AI models without permission or compensation. In response, AI firms often argue that their use of such material falls under the fair use provision of the Copyright Act. OpenAI and Microsoft, for example, have made that argument in an ongoing copyright case brought by The New York Times. Until now, the big Hollywood studios had mostly stayed on the sidelines. That changed today when the two largest, Universal and Disney, filed a joint copyright infringement lawsuit against Midjourney, one of the original AI image generators. (Midjourney is expected to roll out a new image-to-video feature this month, which may have prompted the legal action.) The studios have been under growing pressure from artists and writers to challenge AI labs over how training data is sourced. The Disney and Universal suit may be the opening salvo in a much broader conflict between Hollywood and the AI industry. Why start with Midjourney, a relatively small player, when far larger and wealthier AI companies have used the same kinds of scraped, copyrighted data to train their models? As The New York Times notes, the lawsuit appears to set the stage for something bigger. The language goes beyond a simple dispute between two companies. The plaintiffs argue that the use of copyrighted training data threatens to upend the bedrock incentives of U.S. copyright law that drive American leadership in movies, television and other creative arts. Stay tuned. More AI coverage from Fast Company: Databricks new One dashboard brings AI to the business class This corny conservative credit card ad signals a very scary future for AI OpenAI and Anthropic are getting cozier with government Teaching AI isnt enoughwe need to teach wisdom, too Want exclusive reporting and trend analysis on technology, business innovation, future of work, and design? Sign up for Fast Company Premium.
Category:
E-Commerce
Many developers find that AI programming assistants have made writing code easier than ever. But maintaining the infrastructure that actually runs that code remains a challenge, requiring engineers to have detailed knowledge of complex cloud systems and how their companies use them. A startup called Antimetal is working to harness AI to guide engineers through resolving issues with software infrastructure in much the same way existing tools help them write code. “We’re going to investigate that by piecing this data together, telling you this really thorough narrative of what happened and why, and leading you to prescriptive actions about what you should do next,” says Antimetal cofounder and CTO Shreyas Iyer. The company began about two-and-a-half years ago, initially focusing on helping businesses optimize their spending on Amazon Web Services. The cloud platform is known for offering a vast range of infrastructure for startups and enterprises to run software online, but it’s also notoriously difficult to optimize for cost. Unexpected eventssuch as a surge in app popularity or a simple coding or configuration issuecan trigger sudden spikes in costs as metered resources are consumed more rapidly than anticipated. Antimetal’s software can automatically recommend ways to reduce monthly bills and quickly alert engineers to unexpected cost spikes before they cause financial strain. But, says cofounder and CEO Matthew Parkhurst, cost management was only the starting point for building a broader AI-driven infrastructure management platform. “Cost is still very important to us,” he says. “But now we’re going a bit more broad in terms of more holistic management.” [Image: Courtesy of Antimetal] Antimetal, which recently secured a $20 million Series A funding round, is now beta testing technology that helps engineers address infrastructure issues as they arise. The AI can offer suggestions or automate tasks such as rolling back code or restarting systems to fix problems. It also integrates with a wide range of platforms, including code repositories like GitHub, logging and monitoring tools, and major cloud providers like AWS, Google Cloud Platform, Microsoft Azure, and Oracle Cloud Infrastructure. This allows it to fetch relevant data and propose potential solutions. It can even tap into companies ticketing platforms to understand how previous issues were resolved. “What we’re doing to start is plugging into effectively every surface area of your infrastructure stack, your observability platform, your cloud, your ticketing system, your source code,” Iyer says. The software, expected to be publicly available later this year, is designed to learn how individual customers prefer to approach various issues, and which solutions tend to be most effective. Many customers begin by granting read-only access, allowing the software to analyze data and offer insights without making changes. As trust grows, they can assign more granular permissionssuch as letting it roll back problematic configuration changes on its own. “We’re not trying to jump to full automation out of the gate,” Iyer says. “We want to build this sort of incremental flow from being an autopilot and assisting you in your time of need to actually automating.” Looking ahead, Antimetal also aims to assist with new software deployments. The system could help teams anticipate potential issues and set up the necessary infrastructure to support new code from day one, Iyer says. The companys original pricing model for its cost optimization tool was based on a percentage of customers’ AWS spending. Parkhurst says pricing for the new AI capabilities will likely be usage-based, though specifics are still being finalized while the product is in beta. He sees strong demand for AI that can support the scaling and maintenance of software productsa task that for many companies is now more daunting than writing the code itself. “Writing code is not the hardest thing anymore,” Parkhurst says. “It’s maintaining it that is, by far, the biggest issue when building a company, or really anything, and serving it to the world.”
Category:
E-Commerce
China’s dominance over critical minerals in global supply chains was a powerful bargaining chip in trade talks between Beijing and Washington that concluded with both sides saying they have a framework to pursue a deal.China has spent decades building the world’s main industrial chain for mining and processing such materials, which are used in many industries such as electronics, advanced manufacturing, defense and health care.Mines and factories in and around Ganzhou, a key production hub for rare earths, underpin China’s control over the minerals. Many residents grew up collecting rocks containing the valuable minerals from the forested hills surrounding the southern city and today make a living from mining, trading or processing them. Critical minerals as a trade issue Responding to ever higher tariffs and other controls on advanced technology, China told exporters of certain key rare earths and other critical minerals to obtain licenses for every shipment abroad. Approvals can take weeks, leading to supply chain disruptions in the U.S. and other countries.President Donald Trump said Wednesday that China would make it easier for American industry to obtain much-needed needed magnets and rare earth minerals, clearing the way for talks to continue between the world’s two biggest economies. In return, Trump said, the U.S. will stop efforts to revoke the visas of Chinese nationals on U.S. college campuses.But details remain scarce. Beijing has not confirmed what the negotiators agreed to, and Chinese President Xi Jinping and Trump himself have yet to sign off on it.The Chinese Commerce Ministry said Saturday it had approved a “certain number” of export licenses for rare earth products, apparently acknowledging Trump’s personal request to Xi during a phone call last week. And on Wednesday, the Ganzhou-based rare-earth conglomerate JL MAG Rare-Earth Co. confirmed it had obtained some export licenses for shipments to destinations including the U.S., Europe and Southeast Asia.Experts say, however, Beijing is unlikely to do away with the permit system enabling it to control access to those valuable resources.“I think what the Chinese have proven is they have now created an entire export control regime for rare earths,” said Daniel Kritenbrink, a partner at The Asia Group consultancy. “They can turn that spigot on and off at will.”The only scenario in which China might deregulate its critical minerals export is if the U.S. fully removes tariffs imposed on Chinese goods as part of the trade war, said Wang Yiwei, a professor of international affairs at Renmin University, echoing the Chinese government’s earlier stance.“Without that,” he said, “it will be difficult to blame China for continuing to strengthen its export controls.” An industry built over decades with government support In 1992, Deng Xiaoping, the leader who launched China’s ascent as the world’s biggest manufacturing power, famously said “the Middle East has oil, China has rare earths,” signaling a desire to leverage access to the key minerals.Several generations later, Beijing has made its rich reserves of rare earths, a group of 17 minerals that are abundant in the earth’s crust but hard, expensive and environmentally polluting to process, a key element of China’s economic security. In 2019, during a visit to a rare earth processing plant in Ganzhou, Xi described rare earths as a “vital strategic resource.”China today has an essential monopoly over “heavy rare earths,” used for making powerful, heat-resistance magnets used in industries such as defense and electric vehicles.The country also produces around 80% of the world’s tungsten, gallium and antimony, and 60% of the world’s germaniumall minerals used in the making of semiconductors, among other advanced technologies.The risks of dependency on Chinese suppliers first came into focus in 2010, when Beijing suspended rare earths exports to Japan due to a territorial dispute. The ban was lifted after about two months, but as a precaution, Japan invested in rare earths processing plants in other countries and began stockpiling the materials.Beijing’s across-the-board requirement for export licenses for some critical minerals has put pressure on world electronics manufacturers and automakers.Some auto parts makers in Europe have shut down production lines due to delays in supply deliveries, according to the European Association of Automotive Suppliers. In the U.S., Tesla CEO Elon Musk said a shortage of rare earths is affecting his company’s work on humanoid robots. China’s critical minerals resources are dwindling In the drab industrial hub of Ganzhou, cradled by the scenic Dayu Mountains, the U.S.-China trade war is still a distant stressor. Miners and small mineral traders interviewed by The Associated Press said they are more concerned about depleting the mountains’ once-abundant resources.Zhong, a tungsten factory manager in Ganzhou who would only give his last name, worked his way up to manager from a miner, but he’s unsure there is a future for him and others in the industry.“I find growing difficulties to source tungsten these days,” he said, adding that smaller mines and trading companies are slowly disappearing as the resources are dwindling. Tungsten is an ultra-hard metal used in armor-piercing ammunition, nuclear reactors and semiconductors.At least five tungsten mines have closed in the area in recent years, according to state media. Remaining reserves are deeper and harder to extract and process after decades of exploitation, said Li Shangkui, chairman of the Ganzhou-based Jiangxi Yuean Advanced Materials Co., Ltd.Processing factories in Ganzhou now routinely source materials from other provinces or other countries. Zhong’s plant imports some raw materials from places like Africa and Cambodia.Major state-owned and private companies in Ganzhou are also ramping up investments abroad. Tungsten producer Ganzhou Haisheng, for instance, announced last year a $25 million investment in a new tungsten plant in Thailand.Whatever the challenges in procuring raw materials, China likely will seek to maintain its dominance in critical minerals, said Fabian Villalobos, an engineer and critical minerals expert at the RAND think tank. The U.S. lags far behind China on critical minerals Between 2020 and 2023, the U.S. imported at least 70% of the rare earth compounds it used from China, according to the U.S. Geological Survey. It has diversified its sources in recent years, but still mainly relies on China.Since beginning his second term in office, Trump has made improving access to critical minerals a matter of national security. Bu the U.S. has an incredibly long way to go to catch up with China, experts say.The sole operational U.S. rare earths mine, in Mountain Pass, California, is unable to separate heavy rare earths. It sends its ore to China for processing. The U.S. Defense Department has provided funding to the mine’s owner, MP Materials, to build new separation facilities. It will take months to build and still only produce a fraction of what is needed.Friction over the issue has opened the way for government-backed financing that was unavailable before, said Mark Smith, who ran the Mountain Pass mine in the early 2010s and now leads NioCorp. It’s seeking about $780 million in financing through the U.S. Export-Import Bank to build a processing facility in Nebraska for critical minerals including rare earths.The Defense Department has committed $439 million to building domestic rare earth supply chains, but building a complete mining and processing industrial chain like China’s could take decades.“There are going to be some real issues here unless we can figure out how to get along with China for a period of time while we’re developing our own resources and our mainstream processing,” Smith said.The spotlight on critical minerals also provides opportunities for smaller miners to invest in extracting and processing some critical minerals, such as tungsten, considered “niche” because they are needed in relatively small amounts in key industries, said Milo McBride, an expert on sustainability and geopolitics at the Carnegie Endowment for International Peace.“For many of these companies, the business strategy hedges on a scenario where the U.S. and China become more confrontational and where trade relations become more uncomfortable,” McBride said. “And all of a sudden, what was once an uneconomic project somewhere outside of China starts to make more sense.” Associated Press news researcher Shihuan Chen contributed to this story. Simina Mistreanu Associated Press
Category:
E-Commerce
All news |
||||||||||||||||||
|
||||||||||||||||||