Economic growth is often driven by breakthroughs, but also requires slow and steady progress

antitrust again FTC to launch investigation into Microsofts cloud business Microsoft is accused of using punitive licensing terms for Azure. Arash Massoudi, James Fontanella-Khan, Stephen Morris, and Stefania Palma, Financial Times Nov 15, 2024 9:44 am | 31 A Microsoft office (not to be confused with Microsoft Office). Credit: Julien GONG Min / Flickr A Microsoft office (not to be confused with Microsoft Office). Credit: Julien GONG Min / Flickr Story textSizeSmallStandardLargeWidth *StandardWideLinksStandardOrange* Subscribers only Learn moreThe Federal Trade Commission is preparing to launch an investigation into anti-competitive practices at Microsofts cloud computing business, as the US regulator continues to pursue Big Tech in the final weeks of Joe Bidens presidency.The FTC is examining allegations that Microsoft is abusing its market power in productivity software by imposing punitive licensing terms to prevent customers from moving their data from its Azure cloud service to competitors platforms, according to people with direct knowledge of the matter.Tactics being examined include substantially increasing subscription fees for those that leave, charging steep exit fees, and allegedly making its Office 365 products incompatible with rival clouds, they added.The FTC is yet to formally request documents or other information from Microsoft as part of the inquiry, the people said.A move to challenge Microsofts cloud business practices would mark the latest broadside against Big Tech by the FTCs chair, Lina Khan, who has centered her tenure on aggressively curbing the monopolistic powers of the likes of Meta and Amazon.Khan, who has become the public enemy for most of Wall Streets dealmaking community, is set to be replaced after president-elect Donald Trump enters the White House next year.While any successor to Khan may not adopt as tough a stance, potential contenders are expected to continue targeting Big Tech companies that have attracted bipartisan ire in Washington. The Republican Party has accused online platforms of allegedly censoring conservative voices.The decision to launch a formal probe would come after the FTC sought feedback from industry participants and the public on cloud computing providers business practices. The results in November last year revealed that most responses raised concerns around competition, the agency said at the time, including software licensing practices that curb the ability to use some software in other cloud providers ecosystems.The FTC also highlighted fees charged on users transferring data out of certain cloud systems and minimum spend contracts, which offer discounts to companies in return for a set level of spending.Microsoft has also attracted scrutiny from international regulators over similar matters. The UKs Competition and Markets Authority is investigating Microsoft and Amazon after its fellow watchdog Ofcom found that customers complained about being locked in to a single provider, which offers discounts for exclusivity and charge high egress fees to leave.In the EU, Microsoft has avoided a formal probe into its cloud business after agreeing to a multimillion-dollar deal with a group of rival cloud providers in July.The FTC in 2022 sued to block Microsofts $75 billion acquisition of video game maker Activision Blizzard over concerns the deal would harm competitors to its Xbox consoles and cloud-gaming business. A federal court shot down an attempt by the FTC to block it, which is being appealed. A revised version of the deal in the meantime closed last year following its clearance by the UKs CMA.Since its inception 20 years ago, cloud infrastructure and services has grown to become one of the most lucrative business lines for Big Tech as companies outsource their data storage and computing online. More recently, this has been turbocharged by demand for processing power to train and run artificial intelligence models.Spending on cloud services soared to $561 billion in 2023 with market researcher Gartner forecasting it will grow to $675 billion this year and $825 billion in 2025. Microsoft has about a 20 percent market share over the global cloud market, trailing leader Amazon Web Services that has 31 percent, but almost double the size of Google Cloud at 12 percent.There is fierce rivalry between the trio and smaller providers. Last month, Microsoft accused Google of running shadow campaigns seeking to undermine its position with regulators by secretly bankrolling hostile lobbying groups.Microsoft also alleged that Google tried to derail its settlement with EU cloud providers by offering them $500 million in cash and credit to reject its deal and continue pursuing litigation.The FTC and Microsoft declined to comment. 2024 The Financial Times Ltd. All rights reserved. Not to be redistributed, copied, or modified in any way.Arash Massoudi, James Fontanella-Khan, Stephen Morris, and Stefania Palma, Financial TimesArash Massoudi, James Fontanella-Khan, Stephen Morris, and Stefania Palma, Financial Times 31 Comments

Half-Life 2 Week I played Half-Life 2 for the first time this yearheres how it went Wake up and smell the ashes, Ms. Washenko. Anna Washenko Nov 15, 2024 9:00 am | 109 This article is part of our 20th anniversary of Half-Life 2 series. Credit: Aurich Lawson This article is part of our 20th anniversary of Half-Life 2 series. Credit: Aurich Lawson Story textSizeSmallStandardLargeWidth *StandardWideLinksStandardOrange* Subscribers only Learn moreIt'sHalf-Life 2week at Ars Technica! This Saturday, November 16, is the 20th anniversary of the release ofHalf-Life 2a game of historical importance for the artistic medium and technology of computer games. Each day up through the 16th, we'll be running a new article looking back at the game and its impact.The time has finally come to close one of the most notable gaps in my gaming history. Despite more than a decade of writing about video games and even more years enjoying them, I never got around to playing Half-Life 2.Not only have I not played it, but I've managed to keep myself in the dark about pretty much everything to do with it. I always assumed that one day I would get around to playing this classic, and I wanted the experience to be as close as possible to it would have been back in 2004. So my only knowledge about Half-Life 2 before starting this project was 1) the game is set in the same universe as Portal, a game I love, 2) the protagonist is named Gordon Freeman, and he looks uncannily like a silent, spectacled young Hugh Laurie, and 3) there's something called the Gravity Gun.That's it. I didn't even know exactly what the Gravity Gun did, only that it existed.So, the time has come for me to learn what the fuss is all about. I've cataloged my off-the-cuff reactions as well as my more analytical thoughts about Half-Life 2, both as a standalone project and as a catalyst for setting new standards in design. But if you're looking for the TL;DR of whether I think the game holds up, my answer is: it depends.Beginning a classic with a clunkA red letter day indeed! Time to experience this iconic piece of video game history. I spend most of the intro sequence in the train station soaking in the atmosphere of the dystopian City 17. A few minutes in, though, I think I'm supposed to sneak past a guard. Because I'm a fugitive trying to escape this freaky Big Brother building, and I swear Barney told me to avoid detection. Instead, the guard immediately sees me and whomps me on the head for not putting a bottle into the trash. Not an auspicious beginning.I make it to Dr. Kleiner's lab for a little bit of story exposition. I like this rag-tag group of geniuses and the whole vibe of a secret scientific rebellion. I also appreciate that it's not a static cutscene, so I can poke around the lab while I listen or observe the characters interacting.After a failed teleport and getting a crowbar from Barney, I then spend a long time getting shot and dying in a train yard. Like, an embarrassingly long time. Perhaps I was assuming at this early stage that Half-Life 2 would be like Portal with real guns, because I figured this area had to be a puzzle. I'm not sure how I missed the one portion of the environment that I could slip through, but I convinced myself that I was supposed to leap across the tops of the train cars, Frogger style. And Gordon might have many skills, but his jumping leaves something to be desired.Finally, I realize that there's a gap in the cars, and I move along. This canal setting is striking, but I keep being unsure which areas of the map I can access. I've heard that the level design is one of the most lauded parts of Half-Life 2, but this is proving to be a genuine struggle with the game.When I played Portal, I sometimes was unsure how to progress, but because that game is presented in the austere confines of a science experiment, I felt like I was supposed to be challenged. In Half-Life 2, though, where there are higher stakes and I'm running for my life, getting stuck just makes me feel dumb and annoyed. And I'm doubly annoyed because this escape sequence would probably feel amazing if I didn't keep getting lost. Again, not the thrilling start I was hoping for.Killing a barnacle by feeding it an explosive barrel is a definite high point. I may have cackled. This is the sort of clever environmental interaction I expected to see from the minds that later made Portal.Headcrabs, on the other hand, are just obnoxious. My dinky little pea shooter pistol doesn't feel like great protection. What's a rogue physicist gotta do to get a shotgun?From airboats to zombiesAfter a break, I return armed with a renewed determination to grok this game and, more importantly, with an airboat. For 90 percent of the Water Hazard chapter, I am feeling like a badass. I'm cruising in my watery ride, flying over ramps, and watching a silo collapse overhead. Especially in those rare moments when the 2000s electro jams punctuate my fights, I feel like a true action hero. The airboat sequence was divisive in 2004, but this writer enjoyed it. Credit: Anna Washenko Next I reach the Black Mesa East chapter, which is a perfect interlude. The game's approach to world-building is probably the area where my feelings align most closely with those of Half-Life 2 veterans. It is spectacular. Heading down into the lab may be the best elevator ride I've taken in a game. Judith is talking science, and outside the shaft, I see humans and vortigaunts conducting fascinating experiments. Small vignettes like those are a perfect way to introduce more information about the rebellion. They give subtle context to a game that doesn't do much to explain itself and doesn't need to.Also, Dog is the best boy. Seriously, I've seen modern games where the animations didn't have as much personality as when Alyx treats her robot protector like an actual dog, and he shakes in delight. My only sadness is that Dog doesn't accompany me to Ravenholm. Dog is, in fact, the best boy. Credit: Valve I do wish Dog had come with me to Ravenholm. I learned after the fact that this chapter is one of the most iconic and beloved, but I had the opposite reaction. Survival horror is not my jam. These whirling death traps are sweet, but I hate jump scares, and I don't love any of my weapons for the encounters.That brings me to something I don't want to say, but in the spirit of journalistic honesty, I must: I don't adore the Gravity Gun. Obviously it was the game's signature creation here and probably what most of you recall most fondly, but I did not fully grasp its potential immediately. Based on the tutorial in Black Mesa East, I assumed it would mostly be a component to puzzle-solving and traversal rather than a key part of combat. I only started using it as a weapon in Ravenholm because I ran out of ammo for everything else.It's not that I don't get the appeal. Slicing zombies up with a saw blade or bashing them with paint cans is satisfyingno complaints there. But I found the tool inconsistent, which discouraged me from experimenting as much as the developers may have hoped. I'm pretty sure I do as much damage to myself as to enemies trying to lob exploding barrels. I want to be able to fling corpses around and can't (for reasons that became apparent later, but in the moment felt limiting). Later chapters reinforced my uncertainty, when I couldn't pull a car to me, yet a push blast had enough power to overturn the vehicle.And once again, I had a rough time with navigation. Maybe I was missing what other people would have seen as obvious cues, the way I'm attuned to finding climbing paths marked by color in modern gamescontroversial as that yellow marking convention may be, its absence is noted when you're struggling to read the environment with a visual language for the game that emphasizes realism over readability. Or maybe I've gotten over-reliant on the tools of the sprawling RPGs I favor these days, where you have a mini-map and quest markers to help you manage all the threads. But for an agonizingly long time, I stared at an electrified fence and wires that seemed to lead to nowhere before realizing that I was supposed to enter the building where Father Grigori first appeared on the balcony. A giant bonfire of corpses out front seemed like a clear 'do not enter' sign, so it didn't occur to me that I could go inside. Alas.Speaking of which, Father Grigori is the best part of the section. He's a total bro, giving me a shotgun at long last. I feel kind of bad when I just abandon him to his murderous flock at the end of the chapter. I hope he survives?Familiarity and finding my footingThe new weapons are coming fast and furious now. I'm impressed at how good the combat feel is. I like the pulse rifle a lot, and that has become my go-to for most long-distance enemies. I wish I could aim down sights, but at least this feels impactful at range. Although I don't usually favor the slow cadence of a revolver in other games, I also enjoy the magnum. The SMG serves well as a workhorse, while the rocket launcher and crossbow are satisfying tools when the right situation arises.But my favorite weapon, far more than the Gravity Gun, is the shotgun. Especially at point-blank range and into a fast zombie's head. Chef's kiss. Maybe it's my love of Doom (2016) peeking through, but any time I can go charging into a crowd with my shotgun, I'm a happy camper.While the worldbuilding in Half-Life 2 is stellar, I don't think the writing matches that high. Just about every brief encounter with allies starts with someone breathlessly gasping, "Gordon? Gordon Freeman?" It's the sort of repetition that would make for an effective and dangerous drinking game.I was surprised when I entered another vehicle section. I liked the airboat, even though the chapter ran a touch long, but this dune buggy feels a lot jankier. At least it starts with a gun attached.I love the idea of this magnet crane puzzle. I wish it didn't control like something from Octodad, but I do get my buggy up out of the sand. The "floor is lava" sequence involves placing objects with the Gravity Gun to avoid disturbing an army of angry antlions by stepping on the sand. Credit: Anna Washenko Things start turning around for me once I reach the sandy version of 'the floor is lava.' That's a cute idea. Although I keep wanting to rotate objects and have a more controlled placement with the Gravity Gun like I could when I did these kinds of tasks with the Ultrahand ability during The Legend of Zelda: Tears of the Kingdom. I understand that Half-Life 2 crawled so TotK could run, but that knowledge doesn't mean I have a better time using the mechanic. Toward the end of the sequence, I got bored by the slow pace of creating a bridge and just barrelled ahead, willing to face a firefight just to move things along.At this point, however, things take a decided turn for the better when I get my other favorite weapon of the game: My own antlion army! Commanding them is so fabulously ridiculous. The scene where hordes of antlions leap over high walls to attack gunmen on the towers leading into Nova Prospekt may be my favorite moment so far in the entire game. If this is how all of you felt flinging radiators around with the Gravity Gun back in the day, then I get why you love it so. I'm sure I won't be able to keep this glorious power indefinitely, but I would happily finish the rest of the story with just antlions and a shotgun if they'd let me.The Nova Prospekt area is the first time I really see a clear line connecting Half-Life 2 to Portal. None of the puzzles or characters thus far gave me Portal vibes, but I definitely get them here, especially once turrets come into play. By this point, I'm finally navigating the space with some confidence. That might be the result of logging enough hours or maybe it was just the sense that GLaDOS could start talking to me at any moment. Whatever the reason, I think I'm finding my groove at last. Nova Prospekt is one of the first areas Valve made when it developed Half-Life 2, so it's not surprising it bears a lot of similarity to environments and vibes in both the original Half-Life and in Portal. Credit: Anna Washenko Somehow I am not surprised by Judith's sudden but inevitable betrayal in this chapter. Alyx not getting along with her in the Black Mesa East chapter felt pretty telling. But then she's just going to let Judith enter teleport coordinates unsupervised? Alyx, you're supposed to be smart!What do you knownow Judith has re-kidnapped Eli. Color me shocked.Onward and upward to the endIt's nice having human minions. They're no antlions, but I like how the world has shifted to a real uprising. It reminds me of the big charge at the end of Mass Effect 3, running and gunning through a bombed-out city with bug-like baddies overhead.Snipers are not a welcome addition to the enemy roster. Not sure why Barney's whining so much. You could throw some grenades, too, my dude.Barney tags along with my minions as we reach the Overwatch Nexus. Destroying floor turrets is probably the first time I've struggled with combat. These are the least precise grenades of all time. Once we make it through the interior sequence, it's time to face down the striders. I can't imagine how you'd play this section bringing down the swarm of them on a harder difficulty. My health takes a beating as I run around the wreckage desperately looking for ammo reloads and medkits. In theory, this is probably a great setpiece, but I'm just stressed out. Things go a little better once the combat is paired with traversal, and the final showdown on the roof does feel like a gratifying close to a boss fight.On to the Citadel. Why on earth would I get into one of these pods? That's a terrible idea. But apparently that's what I'm going to do. I hope I'm not supposed to be navigating this pod in any way, because I'm just taking in the vibes. It's another transit moment with glimpses into what the enemies have been getting up to while the rebellion rages outside. It's eerie; I like it. Battling the striders as the game moves toward its finale. Credit: Anna Washenko The Gravity Gun is the core of Half-Life 2, so it makes sense that a supercharged version is all I have for the final push. I appreciate that I can use it to fling bodies, but my reaction is a little muted since this was an idea I'd had from the start. But I do find the new angle of sucking up energy orbs to be pretty rad.I arrive in Dr. Breen's office, and it looks grim for our heroes. Judith redeeming herself surprises me more than her betrayal, which is nice. When he runs off, I'm mentally preparing myself as I chase him for a final boss showdown. Surely, something extra bonkers with the Gravity Gun awaits me. I climb the teleportation tower, I pelt Breen's device with energy orbs, I'm waiting for the other shoe to drop, andHuh?Context is everythingIn the moment, I was torn between feeling that the opaque ending was genius and that it was an absolute cop-out. It was certainly not how I expected the game to end.But on reflection, that wound up being a fitting final thought as the credits rolled, because I think 'expectations' were at the heart of my conflicted reactions to finally playing Half-Life 2. I've rarely felt so much pressure to have a particular response. I wanted to love this game. I wanted to share the awe that so many players feel for it. I wanted to have an epic experience that matched the epic legacy Half-Life 2 has in gaming history.I didn't.Instead, I had whiplash, swinging between moments of delight and stretches of being stymied or even downright pissed off. I was tense, often dreading rather than eagerly awaiting each next twist. Aside from a handful of high points, I'm not sure I'd say playing Half-Life 2 was fun.As I mentioned at the start, the big question I felt I had to answer was whether Half-Life 2 felt relevant today or whether it only holds up under the rosy glow of nostalgia. And my answer is, "It depends." As an enigmatic person once said, "The right man in the wrong place can make all the difference in the world." It's all in the context.Moments that were jaw-dropping in 2004 have less impact for someone like me who's played the many titles that copied, standardized, and perfected Half-Life 2's revelations. Intellectually, I understand that the Gravity Gun was a literal game-changer and that a physics engine deployed at this scale was unheard of. But funnily enough, a modern player is even less likely to see those innovations as so, well, innovative when a game has as much polish as Half-Life 2 does. Half-Life 2 has almost no rough edges in the execution. Everything works the way it was intended.Since that polish means the new ideas don't feel like experiments, and since I've seen them in other games in the intervening years, they don't register as notable.Just as you don't need to be a fan of Aristotle's Poetics to appreciate drama, you don't need to love Half-Life 2 to appreciate its legacy. As a fun game to play, whether it holds up will come down to you and your context. However, as a showcase of the technology of the time and a masterclass in world-building, yes, Half-Life 2 holds up today. 109 Comments

Companies are keeping their central data centers, but theyre also moving more IT to the enterprise edge. The result is a re-imagined concept of data center that includes the data center but also subsumes cloud and other edge-computing operations.In this expanded data center model, ITs role hasnt fundamentally changed. It must still implement, monitor and maintain data center operations, no matter where they occur.But since IT staff cant be at all remote locations at once, software and hardware technologies are being called upon to do the job of facilitating end-to-end data center management, no matter where that management is.Technologies to Facilitate Remote Data Center ManagementTo assist IT in managing the expanded data center, tools and technology solutions must do two key things: monitor and manage IT operations, functions and events; and automate IT operations.Here are five technologies that help:System on a chip (SOC). First conceived in the 1970s, system on a chip embeds processing, memory and, today, even security and artificial intelligence on a single chip. The chip powers a device or network endpoint.SOC can appear in a router, sensor, smartphone, wearable, or any other Internet of Things (IoT) device. The original selling point of SOCs was their ability to offload processing from the central data center and reduce latency when processing can be done locally.Related:Now, these SOC routers, devices, and access points come with embedded security that is WPA2/3 compliant and can encrypt data and block DNS attacks or suspicious websites. That security is complemented with AI that aids in threat detection and in some cases, threat mitigation, such as being able to automatically shut down and isolate a detected threat.To use SOC threat detection and mitigation at the edge, IT must:Ensure that the security ruleset on edge devices is in concordance with corporate-wide data center security policies; andEmploy an overarching network monitoring solution that can integrate the SOC-based security with central data center security and monitoring so every security action can be observed, analyzed, and mitigated from a single pane of glass in the central data center.Zero-trust networks. Zero-trust networks trust no one with unlimited access to all network segments, systems, and applications. In the zero-trust scheme, employees only gain access to the IT resources they are authorized for.Users, applications, devices, endpoints, and the network itself can be managed from a central point. Internal network boundaries can be set to allow only certain subsets of users access. An example is a central data center in Pittsburgh with a remote manufacturing plant in Phoenix. A micro network can be defined for the Phoenix plant that can only be used by the employees in Phoenix. Meanwhile, central IT has full network management, monitoring, and maintenance capability without having to leave the central data center in Pittsburgh.Related:Automated operations. Data and system backups can be automated for servers deployed at remote points, whether these backups are ultimately rerouted to the central data center or a cloud service. Other IT functions that can be automated with guidance from an IT ruleset include IT resource provisioning and de-provisioning, resource optimization, and security updates that are automatically pushed out for multiple devices.Its also possible to use remote access software that allows IT to gain control of a users remote workstation to fix a software issue.Edge data centers.Savings in communications can be achieved, and low-latency transactions can be realized if mini-data centers containing servers, storage and other edge equipment are located proximate to where users work. Industrial manufacturing is a prime example. In this case, a single server can run entire assembly lines and robotics without the need to tap into the central data center. Data that is relevant to the central data center can be sent later in a batch transaction at the end of a shift.Related:Organizations are also choosing to co-locate IT in the cloud. This can reduce the cost of on-site hardware and software, although it does increase the cost of processing transactions and may introduce some latency into the transactions being processed.In both cases, there are overarching network management tools that enable IT to see, monitor and maintain network assets, data, and applications no matter where they are. The catch is that there are still many sites that manage their IT with a hodgepodge of different types of management software.A single pane of glass. At some point, those IT departments with multiple network monitoring software packages will have to invest in a single, umbrella management system for their end-to-end IT. This will be necessary because the expanding data center is not only central, but that could be in places like Albuquerque, Paris, Singapore and Miami, too.ITs end goal should be to create a unified network architecture that can observe everything from a central point, facilities automation, and uses a standard set of tools that everybody learns.Are We There Yet?Most IT departments are not at a point where they have all of their IT under a central management system, with the ability to see, tune, monitor and/or mitigate any event or activity anywhere. However, we are at a point where most CIOs recognize the necessity of funding and building a roadmap to this uber management network concept.The rise of remote work and the challenge of managing geographically dispersed networks have driven the demand for network management system (NMS) solutions with robust remote capabilities, reports Global Market Insights, adding, As enterprises increasingly seek remote network management, the industry is poised for substantial growth.

Marco Addino, Managing Director, AccentureNovember 15, 20244 Min ReadPanther Media GmbH via Alamy StockWithout doubt, todays society relies on the semiconductor industry. After all, can you imagine a world without smartphones, cars, power stations, and televisions? We, as people, and the global economy more broadly, rely on continued innovation from the chips the industry produces. But there are challenges facing these companies across the board -- design, manufacturing and demand. Talent is in increasingly short supply, and on top of that geopolitical tensions and onshore manufacturing add another layer of complexity. The industry keeps having hurdles to cross, one after another it seems. Only recently, another problem for the industry made the headlines when Hurricane Helene hit Spruce Pine, one of the worlds most important locations for semiconductor, raising questions about the impact it would have.It's already tough enough for semiconductor companies to deal with and resolve these issues, but they are appearing while generative AI has made the need for innovation a must do now, not a must do at some point. The question is whether the semiconductor industry can reinvent itself quickly enough for this new generative AI moment. Accenture analysis found reinventors (those companies that have already built the capability for continuous reinvention) increased revenues by 15 percentage points over other companies between 2019 and 2022. We expect that gap in revenue growth between reinventors and the rest to increase by 2.4 times to 37 percentage points by 2026, so theres a clear opportunity for them. Yet our survey of global semiconductor executives found that 71% believe it will take at least three years for the semiconductor industry to deploy generative AI at scale. The industry could do with that timeframe accelerating somewhat.Related:The Challenges AheadIts not going to be easy, however, of course it wont. But semiconductor companies need to use generative AI across the entire spectrum -- spanning design and manufacturing, through sales and marketing, to customer service --to seize opportunities for innovation in both the short- and long-term. Adapting that broad view across the value chain is a must to reinvent the value chain, however daunting that may initially seem.There are other concerns too, such as IP. In fact, 73% of executives cite IP concerns as the biggest barrier to generative AI deployments. Then theres of course the cost issue and the need to balance technical debt with investments for the future, both of which, are necessary.Once leaders grapple with how those challenges can be overcome, theres another pressing challenge and thats having the right talent in place to deploy these applications successfully.Related:Most semiconductor companies are already fully aware of that and are doing everything they can to accelerate gaining new talent and reskilling their existing workforce. However, the speed with which generative AI is changing the way businesses work means they must also get support from across their ecosystem to ensure they have all critical skills in place.It's Time for Leaders to Place Their BetsThe industry needs to move forward with two workstreams running in parallel. First, CEOs and other business leaders must make no regrets moves; those use cases with the lowest risk, shortest time to show results and therefore, value. For example, Generative AI-enabled field service assistants would allow field service engineers to perform root cause analyses faster and recommend repair methods based on machine data, therefore reducing downtime and accelerating production. It also provides immediate access to information that helps technicians increase their knowledge, therefore helping with the skills gap. Generative AI can also be used in other areas, such as sales and marketing where it can improve the quality and level of personalization of the content to drive more personalized campaigns.Related:At the same time, strategic bets need to be decided upon to support the long-term goals of the business. An example of this is in process engineering. Generative AI-enabled applications that incorporate historical process parameter data to create more efficient designs for semiconductor equipment and wafer development. These tools can use drawings, text, images and more to create customized outputs that engineers can use to augment experiments, allowing for a more objective approach to experimental design. These strategic bets will be the things that will offer the highest value. They may well take some time to roll out, but they could pave the way for total reinvention and therefore, competitive advantage.Whether the no regret moves or strategic bets, the guiding principle is choosing the right use cases, at the right point and at the right time. Every semiconductor companys generative AI journey is different, but the approaches will be similar. All companies must establish a solid data foundation, have the necessary skills in place, and importantly, have the right ecosystem in position. Those that come out on top wont just be the best player, but the businesses that put the right connections in place.About the AuthorMarco AddinoManaging Director, AccentureMarco Addino is a managing director in Accentures high tech industry practice leading the companys semiconductor business in EMEA, and is the client account lead for Italy, Central Europe and Greece, responsible for building and growing strategic relationships in the region. He is experienced in high complexity products engineering, supply chain and operations, large scale digital and technology transformations, organizational design, post-merger integration, and the design and implementation of platform business models.See more from Marco AddinoNever Miss a Beat: Get a snapshot of the issues affecting the IT industry straight to your inbox.SIGN-UPYou May Also LikeWebinarsMore WebinarsReportsMore Reports

It has been 50 years since archaeologists discovered Lucy, perhaps the most famous ancient hominin ever found. But the scientists who have studied her say that this fossil gave us a misleading image of the nature of her species

A classic study found that people can fail to notice a gorilla when they are focusing on something else, but new experiments suggest this "inattentional blindness" might not tell the whole story

We are on the cusp of a voice AI boom, with tech companies such as Apple and OpenAI rolling out the next generation of artificial-intelligence-powered assistants. But the default voices for these assistants are often white AmericanBritish, if youre luckyand most definitely speak English. They represent only a tiny proportion of the many dialects and accents in the English language, which spans many regions and cultures. And if youre one of the billions of people who dont speak English, bad luck: These tools dont sound nearly as good in other languages. This is because the data that has gone into training these models is limited. In AI research, most data used to train models is extracted from the English-language internet, which reflects Anglo-American culture. But there is a massive grassroots effort underway to change this status quo and bring more transparency and diversity to what AI sounds like: Mozillas Common Voice initiative. The data set Common Voice has created over the past seven years is one of the most useful resources for people wanting to build voice AI. It has seen a massive spike in downloads, partly thanks to the current AI boom; it recently hit the 5 million mark, up from 38,500 in 2020. Creating this data set has not been easy, mainly because the data collection relies on an army of volunteers. Their numbers have also jumped, from just under 500,000 in 2020 to over 900,000 in 2024. But by giving its data away, some members of this community argue, Mozilla is encouraging volunteers to effectively do free labor for Big Tech. Since 2017, volunteers for the Common Voice project have collected a total of 31,000 hours of voice data in around 180 languages as diverse as Russian, Catalan, and Marathi. If youve used a service that uses audio AI, its likely been trained at least partly on Common Voice. Mozillas cause is a noble one. As AI is integrated increasingly into our lives and the ways we communicate, it becomes more important that the tools we interact with sound like us. The technology could break down communication barriers and help convey information in a compelling way to, for example, people who cant read. But instead, an intense focus on English risks entrenching a new colonial world order and wiping out languages entirely. It would be such an own goal if, rather than finally creating truly multimodal, multilingual, high-performance translation models and making a more multilingual world, we actually ended up forcing everybody to operate in, like, English or French, says EM Lewis-Jong, a director for Common Voice. Common Voice is open source, which means anyone can see what has gone into the data set, and users can do whatever they want with it for free. This kind of transparency is unusual in AI data governance. Most large audio data sets simply arent publicly available, and many consist of data that has been scraped from sites like YouTube, according to research conducted by a team from the University of Washington, and Carnegie Mellon andNorthwestern universities. The vast majority of language data is collected by volunteers such as Blent zden, a researcher from Turkey. Since 2020, he has been not only donating his voice but also raising awareness around the project to get more people to donate. He recently spent two full-time months correcting data and checking for typos in Turkish. For him, improving AI models is not the only motivation to do this work. Im doing it to preserve cultures, especially low-resource [languages], zden says. He tells me he has recently started collecting samples of Turkeys smaller languages, such as Circassian and Zaza. However, as I dug into the data set, I noticed that the coverage of languages and accents is very uneven. There are only 22 hours of Finnish voices from 231 people. In comparison, the data set contains 3,554 hours of English from 94,665 speakers. Some languages, such as Korean and Punjabi, are even less well represented. Even though they have tens of millions of speakers, they account for only a couple of hours of recorded data. This imbalance has emerged because data collection efforts are started from the bottom up by language communities themselves, says Lewis-Jong. Were trying to give communities what they need to create their own AI training data sets. We have a particular focus on doing this for language communities where there isnt any data, or where maybe larger tech organizations might not be that interested in creating those data sets, Lewis-Jong says. They hope that with the help of volunteers and various bits of grant funding, the Common Voice data set will have close to 200 languages by the end of the year. Common Voices permissive license means that many companies rely on itfor example, the Swedish startup Mabel AI, which builds translation tools for health-care providers. One of the first languages the company used was Ukrainian; it built a translation tool to help Ukrainian refugees interact with Swedish social services, says Karolina Sjberg, Mabel AIs founder and CEO. The team has since expanded to other languages, such as Arabic and Russian. The problem with a lot of other audio data is that it consists of people reading from books or texts. The result is very different from how people really speak, especially when they are distressed or in pain, Sjberg says. Because anyone can submit sentences to Common Voice for others to read aloud, Mozillas data set also includes sentences that are more colloquial and feel more natural, she says. Not that it is perfectly representative. The Mabel AI team soon found out that most voice data in the languages it needed was donated by younger men, which is fairly typical for the data set. The refugees that we intended to use the app with were really anything but younger men, Sjberg says. So that meant that the voice data that we needed did not quite match the voice data that we had. The team started collecting its own voice data from Ukrainian women, as well as from elderly people. Unlike other data sets, Common Voice asks participants to share their gender and details about their accent. Making sure different genders are represented is important to fight bias in AI models, says Rebecca Ryakitimbo, a Common Voice fellow who created the project's gender action plan. More diversity leads not only to better representation but also to better models. Systems that are trained on narrow and homogenous data tend to spew stereotyped and harmful results. We dont want a case where we have a chatbot that is named after a woman but does not give the same response to a woman as it would a man, she says. Ryakitimbo has collected voice data in Kiswahili in Tanzania, Kenya, and the Democratic Republic of Congo. She tells me she wanted to collect voices from a socioeconomically diverse set of Kiswahili speakers and has reached out to women young and old living in rural areas, who might not always be literate or even have access to devices. This kind of data collection is challenging. The importance of collecting AI voice data can feel abstract to many people, especially if they arent familiar with the technologies. Ryakitimbo and volunteers would approach women in settings where they felt safe to begin with, such as presentations on menstrual hygiene, and explain how the technology could, for example, help disseminate information about menstruation. For women who did not know how to read, the team read out sentences that they would repeat for the recording. The Common Voice project is bolstered by the belief that languages form a really important part of identity. We think its not just about language, but about transmitting culture and heritage and treasuring peoples particular cultural context, says Lewis-Jong. There are all kinds of idioms and cultural catchphrases that just dont translate, they add. Common Voice is the only audio data set where English doesnt dominate, says Willie Agnew, a researcher at Carnegie Mellon University who has studied audio data sets. Im very impressed with how well they've done that and how well they've made this data set that is actually pretty diverse, Agnew says. It feels like theyre way far ahead of almost all the other projects we looked at. I spent some time verifying the recordings of other Finnish speakers on the Common Voice platform. As their voices echoed in my study, I felt surprisingly touched. We had all gathered around the same cause: making AI data more inclusive, and making sure our culture and language was properly represented in the next generation of AI tools. But I had some big questions about what would happen to my voice if I donated it. Once it was in the data set, I would have no control about how it might be used afterwards. The tech sector isnt exactly known for giving people proper credit, and the data is available for anyones use. As much as we want it to benefit the local communities, theres a possibility that also Big Tech could make use of the same data and build something that then comes out as the commercial product, says Ryakitimbo. Though Mozilla does not share who has downloaded Common Voice, Lewis-Jong tells me Meta and Nvidia have said that they have used it. Open access to this hard-won and rare language data is not something all minority groups want, says Harry H. Jiang, a researcher at Carnegie Mellon University, who was part of the team doing audit research. For example, Indigenous groups have raised concerns. Extractivism is something that Mozilla has been thinking about a lot over the past 18 months, says Lewis-Jong. Later this year the project will work with communities to pilot alternative licenses including Nwulite Obodo Open Data License, which was created by researchers at the University of Pretoria for sharing African data sets more equitably. For example, people who want to download the data might be asked to write a request with details on how they plan to use it, and they might be allowed to license it only for certain products or for a limited time. Users might also be asked to contribute to community projects that support poverty reduction, says Lewis-Jong. Lewis-Jong says the pilot is a learning exercise to explore whether people will want data with alternative licenses, and whether they are sustainable for communities managing them. The hope is that it could lead to something resembling open source 2.0. In the end, I decided to donate my voice. I received a list of phrases to say, sat in front of my computer, and hit Record. One day, I hope, my effort will help a company or researcher build voice AI that sounds less generic, and more like me. This story has been updated.

Drinking alcohol in moderation was thought to be healthy, but research now suggests there is no safe level.Alcohol harm expert Richard Piper used to drink heavily.His 'dry by default' rule helped him reduce the risks of drinking without going sober.At 52, Richard Piper runs often, feels full of energy, weighs less, and enjoys concerts and holidays more than he did at 42. He puts this down to being what he calls "dry by default."Piper told Business Insider that he drank heavily daily for years. But after becoming the CEO of harm reduction charity Alcohol Change UK in 2017, he realized he needed to address his drinking habits.He joined the growing number of people who are drinking less, particularly Gen Zers and millennials. A survey of US adults carried out by Gallop between 2021 and 2023 found that 62% of respondents aged 18 to 34 drank alcohol, down from 72% between 2001 and 2003. A separate Gallup poll published earlier this year found that 86% of the US adults surveyed thought that alcohol was "very" (33%) or "somewhat" (53%) harmful.Being "dry by default" means Piper will drink on rare occasions rather than most days, for example when non-alcoholic alternative to a beer he likes isn't available. He prefers being dry by default to sober because he doesn't have to follow "permanent lifelong rules," he said, but he can lower his risk of alcohol-related harm."There is no such thing as healthy drinking," Piper said. "But there's more risky drinking, or there's less risky drinking. And the less you drink, the healthier it is." Piper mostly drinks alcohol-free beers, but will have an alcoholic one from time to time. Niall Carson/PA Images via Getty Images Research suggests all drinking is harmful even in moderationSome evidence suggests that moderate drinking could be better for youNow, growing evidence suggests that no amount of alcohol is safe. A 2021 study published in the European Journal of Public Health found that in 2017, light to moderate drinking caused 13.3% of cancers related to alcohol in the European Union. That was defined as two shots of spirits, just under two bottles of beer (20.2 ounces), or a little more than a glass (6.76 ounces) of wine a day.Drinking doesn't only bring health risks. "Alcohol can lead to poor decision-making about getting home safely, about who you talk to, and what you say to them and makes many of us do things that we wouldn't do otherwise," Piper said.You don't have to quit drinking to lower the health risksThe US Department of Health and Human Services recommends adults don't drink, or drink moderately, meaning fewer than one to two drinks a day.Piper said that one to two units of alcohol a week is genuinely low risk for dying from alcohol and alcohol-related issues, in line with Canada's guidelines.People should make sure that the benefits of drinking outweigh the risks, and generally, most people would see benefits from drinking less, he said. Benefits might include losing weight, saving money, sleeping better, easing depression and anxiety, having more energy, and things you can't see, such as lower cancer risks."Basically the less you drink, the better for your health, your mental health, and your well-being," he said, even if you don't give up completely."Make sure you have a really good reason for having a drink," he said. "And it's always good to have a few days off a week. Alcohol should be on the periphery of our lives, not at the center."

As America's grocery king, Walmart sources a high share of its products from the US.Target, by contrast, relies more on merchandise that is often imported, such as apparel and housewares.The difference puts more of Target's business at risk of being impacted if Trump follows through on his tariff promises.For all their similarities, Walmart and Target have key distinctions that could spell big differences in how each would be affected by rising import costs if Donald Trump follows through with his promised tariffs."Target is actually much more exposed than Walmart because Walmart is grocery-heavy and groceries are predominantly domestic," Jason Miller, supply chain professor at Michigan State University, told Business Insider.As America's grocery king, Walmart US makes nearly 60% of its revenue from the grocery category and only about a quarter of sales from general merchandise.In addition, grocery as a share of sales has been increasing in recent years as the general merchandise share has declined, according to Walmart's annual report.Target, by contrast, relies much more heavily on merchandise that is often imported, such as apparel, housewares, and beauty. Food and beverage sales accounted for less than a quarter of Target's sales last year.TD Cowen retail analyst Oliver Chen told BI that Target's apparel segment presents another potential complication, as fashion is more sensitive to seasonality. That could make it more difficult to pull orders forward ahead of possible tariffs or line up new suppliers in time to still be in style."When you miss apparel timeline, you don't get it back, and Target has more apparel exposure," he said.Beyond its grocery-to-merchandise ratio, Walmart has another key advantage over Target: scale."Walmart is much bigger," Kantar analyst Gina Logan told BI. "And I'm not just talking about sales."They have a much more advanced supply chain," she added. "They have a wider range of suppliers, they have more ability to pivot and predict demand, and can use their automation and forecasting in order to react to this in a much faster, more predictive way than Target."This is not the first time the Spark has found a competitive advantage over the Bullseye in the grocery aisle.When US shoppers began to cut spending back in 2023, prioritizing essentials like groceries in their weekly budgets, sales at Walmart chugged along while Target struggled.Target has since found success by taking a much more Walmart-like stance with price cuts and bargain brands, and its share of grocery sales has increased by 1-2 percentage points per year over the last three years.Logan says tariffs, if announced by the Trump administration, would likely push Target harder into the grocery game, especially with its portfolio of private-label food brands.Target and Walmart both report earnings next week. Walmart declined to comment for this story, citing the quiet period, and Target did not respond to BI's request.The last time the companies' executives discussed tariffs on earnings calls was in 2019, according to data from AlphaSense.At the time, Walmart said it would not raise prices on food impacted by tariffs and would instead look to offset the cost elsewhere."As a guest-focused retailer," Target CEO Brian Cornell said in May 2019, "we're concerned about tariffs because they lead to higher prices on everyday products for American families."He later said in a November call that then-President Trump's tariffs were amounting to $50 million to $60 million in added costs per quarter, adding that "obviously we're all facing the same tariff issues together."But as Target's and Walmart's financials show, not everyone is likely to be impacted by potential tariff increases to the same degree.