Future of the Job Market — Impact of AI on Various Roles in 2025 0 like April 16, 2025 Share this post Author(s): Nishtha Nagar Originally published on Towards AI. The job market in 2025 is undergoing a fundamental shift, influenced heavily by the rapid adoption of artificial intelligence (AI). As AI technologies, particularly generative AI, become more integrated into workplace systems, their impact across industries and roles is becoming increasingly evident. According to the U.S. Bureau of Labor Statistics (BLS), AI will influence occupations in various ways between 2023 and 2033, depending on how easily core tasks can be automated or enhanced. While some roles face disruption, others are projected to grow significantly due to AI integration. Source: U.S. Bureau of Labor Statistics The graph above highlights projected employment changes for select occupations that are considered susceptible to AI-related impacts. Computer-related roles are expected to see strong growth, not despite AI, but because of it. Rather than replacing these jobs, AI is increasingly being used as a tool to boost productivity in tasks like coding, testing, data management, and system maintenance. As a result: Computer-related roles like software developers (17.9% growth), database architects (10.8%), and hardware engineers (7.2%) will thrive, as AI boosts productivity in coding and system maintenance.Financial services will see growth despite automation, with personal financial advisors projected to grow by 17.1%, as human expertise remains crucial.Engineering roles will grow, such as… Read the full blog for free on Medium. Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor. Published via Towards AI Towards AI - Medium Share this post

Last Epoch 1.2 Tombs of the Erased brings a sentinel rework, Weavers endgame Faction, revamped Monolith of Fate experience, and much more. This is Last Epoch’s biggest update yet and IGN has teamed up with the ARPG experts at Maxroll to prepare detailed guides on all of the new content along with excellent builds for leveling and endgame alike. Before we get to the builds, here’s all the new stuff coming in 1.2:New 1.2 Tombs of the Erased ContentWoven FactionOne of the biggest changes in 1.2 is the new Woven Faction. They provide a deeper endgame experience by adding Woven Echoes to your Monolith of Fate. Begin interacting with them by finding a Cemetery of the Erased on your Echo Web, after completing the mini-boss at the end you can enter the Haven of Silk and speak to Masque to join the faction. Gain ranks with the Woven Faction by completing Woven Echoes to add points to your Weaver Tree then use these points to customize your endgame experience.Read more with Maxroll’s guide on the Woven Faction!Champions & Mini-bossesAs you explore the Echo Web, you may also find some fun new surprises in the form of Champions, and new weaver mini-bosses. The Champions are larger, more powerful versions of rares, immediately visible due to their angry red glow. Defeating them is tricky but if you persevere you’re sure to be duly rewarded. There are also several new mini-bosses you can encounter in 1.2, so keep an eye out for a scary-looking boss as you explore the Monolith of Fate.Learn more about Champions with Maxroll’s Champions guide.Set Item Crafting and LP ChangesLast Epoch has an in-depth Crafting System where you use Affix Shards, Glyphs, and Runes to heavily customize your items. This popular system is straighforward to understand and has a fair bit of depth to master. In 1.2, crafting in Last Epoch gets even better. You can break Set items down to obtain crafting shards and then use those to craft set bonuses on your items. This gives you the best of both worlds where you have the power of a set bonus and the flexibility to craft your own gear with the affixes you need. EHG has also made a big change to how you use Unique items with Legendary Potential (LP) to craft Legendary items. In the past, when interacting with the Eternity Cache the modifiers added to your Legendary Items were chosen at random. So if you used a 1LP Unique, you get a Legendary with 1 random affix, if you use a 2LP Unique, you get a Legendary with 2 random affixes.Now, if you complete a Temporal Sanctum of level 2 or higher you can choose 1 affix to transfer to your Legendary! This means you get to pick what you want when using a 1 LP Unique, if you use an item with 2-4 LP, then you get to pick 1 affix and the remaining modifiers are chosen at random.Learn more about Set Item Crafting and how to make your own Legendary Items with these guides from Maxroll.Uber AberrothLast Epoch’s ultimate boss is getting even tougher with 1.2. In addition to the normal version of this encounter, there is a special Uber Aberroth encounter. Very little is known about this foe beyond the fact that he is the most dangerous being in Eterra. Challenge him at your own risk!1.2 Tombs of the Erased Leveling BuildsUse Maxroll’s leveling builds to cruise through the campaign with ease. Season 2 introduces the ability to switch your Mastery so you can level with the most efficient build for the campaign, then respec your mastery upon reaching the endgame. Here are the best ways to level each of the game’s 5 classes, straight from Last Epoch’s top speed runner Terek!PrimalistThe best way to level a Primalist is by playing Druid. This build has good damage, it’s very cozy as you can use transformations to spam your skills during the campaign, and gets quite tanky thanks to Werebear form.Maxroll’s Primalist Leveling Build Guide.MageThe Spellblade mastery seamlessly weaves blades and spells to decimate foes using Shatter Strike. You are reliant on spells in the early game but once you get Shatter Strike most of your time is spent rotating between Shatter Strike for damage and Mana Strike to recover your resources to continue dealing damage. Don’t be fooled by the Mage’s health, you can easily generate tons of Ward just by hitting enemies.SentinelHot off the rework, Sentinel is stronger than ever. The most efficient way to level is by playing Paladin to get easy access to Resistances and Holy Aura for both offensive and defensive bonuses. Crusade your way through the campaign burning enemies in the holy fire of your Judgment!Maxroll’s Sentinel Leveling Build Guide.AcolyteAs an Acolyte you are a master of the undead and the profane. Embrace your inner darkness and become a Warlock to create rifts in reality which unleash a volley of Chaos Bolt. The Chthonic Fissure skill provides great persistent damage for clearing while your Summoned Infernal Shades melt bosses with their hellish assault.Maxroll’s Acolyte Leveling Build Guide.RogueThe Rogue levels with her trusty Falcon Companion, flying through levels with ease and grace. This Falconer build is easy to scale with great sustain thanks to the Melee Leech affix on gear. With a bit of investment into Health, Regen and Physical Resistance, this Rogue becomes a well-rounded beast of a build.1.2 Tombs of the Erased Endgame BuildsAfter you complete the campaign, you enter Last Epoch’s endgame. Maxroll’s Last Epoch team has been extensively testing a wide variety of endgame builds to dominate in the Monolith of Fate. While all of these builds are powerful, you can use Maxroll’s tier lists to see how they compare across different kinds of content.Heartseeker MarksmanThe Heartseeker Marksman uses a new skill which seeks enemy targets and triggers a spectacular display of Fire and Cold Projectiles that cover the entire screen. Watch your enemies evaporate as they get blasted by your Attacks and Spells!The build revolves around dealing Damage with Heartseeker and proccing an insane amount of Dragonfire from Dragonsong. Heartseeker also triggers Burning Dagger for even more craziness. This Marksman has incredible mobility and utility thanks to Shift, Smoke Bomb, Shurikens, and unspecialized Decoy.Pros✔️ Amazing DPS✔️ Fast Clear Speed✔️ Deals Damage While Moving✔️ Insane Scaling with Gear and UniquesCons❌ Active Playstyle❌ Needs Uniques to Shine❌ Survivability Relies on Gear❌ Defensive Tools Require TimingJudgement PaladinThe Judgement Paladin creates Consecrated Ground with Judgement, to burn enemies with holy fire. Become the Judge and cleanse your foes of impurity, as they hopelessly try to break through your immense defenses. This build is a natural continuation of the Paladin Leveling Build, making it extremely easy to play as your first character of the season.The Judgement Paladin deals Fire Spell Damage Over Time with Consecrated Ground and generates Mana with Vengeance. Consecrated Ground also grants massive Healing, since multiple instances can stack on top of each other. Use Holy Aura and Symbols of Hope for great passive buffs. Healing Hands grants Ward and Healing, while acting as your Movement Skill.Pros✔️ Insane Tankiness✔️ Exceptional Sustain✔️ High Single Target DPS✔️ Amazing Scaling With Gear and UniquesCons❌ Limited AoE❌ Average Clear Speed❌ Needs Legendaries To Shine❌ Generator/Spender PlaystyleWarpath Void KnightThe Warpath Void Knight uses Warpath to wipe the battlefield clean of enemies. Stay spinning to increase your Damage and AoE until Warpath almost covers the whole screen. Unleash the Spin-to-Win playstyle with this build!The Warpath Void Knight specializes into Warrior's Fury to reach amazing levels of Void Melee Damage and Area of Effect. Warpath costs more Mana every second, but thanks to the Void Well Passive and Mana stacking, you can stay spinning for a long time. The Void Well Passive is mandatory when you have Warrior's Fury allocated. Devouring Orb, Anomaly, Symbols of Hope, and Abyssal Echoes grant powerful buffs.Pros✔️ Amazing AoE✔️ Extreme Clear Speed✔️ Very Simple Gameplay✔️ Excellent Single Target DPSCons❌ Gear Dependent❌ Average Tankiness❌ Needs Legendaries To Shine❌ Requires Mana InvestmentTornado ShamanSummon the fury of the storm with this Tornado Shaman Guide as you strike down foes from afar. This build invokes Tornados onto the battlefield that rapidly calls Storm Bolts down on enemies in an enormous area. Speed through monoliths and destroy bosses with this outstanding build.Place down long-lasting Tornados on the battlefield that rapidly cast Storm Bolts thanks to Charged Storm for huge Lightning damage. Gathering Storm allows you to drastically increase the damage of those storm bolts while Summon Spriggan, Warcry, and Maelstrom offer you some amazing offensive and defensive bonuses. Since you are not casting Gathering Storm manually, you can equip an unspecialized Fury Leap for mobility.Pros✔️ Insane Tankiness✔️ Exceptional Sustain✔️ High Single Target DPS✔️ Amazing Scaling With Gear and UniquesCons❌ Limited AoE❌ Weaker Defensively While Moving❌ Optimal Play Requires Standing Still❌ Summon Spriggan Dies Often And Needs To Be ResummonedStygian Coal LichThe Stygian Coal Lich Guide harnesses the power of undeath in the form of Stygian Beams. This Lich transforms into its horrific Reaper Form to unleash a barrage of deadly Necrotic Stygian Beams. Hit your enemies with Rip Blood to empower yourself, and generate Mana and obliterate everything around you.Turn Drain Life into a DPS powerhouse thanks to the Stygian Coal Unique Catalyst. Death Seal provides powerful buffs and releases pulsing Waves of Death around you while it's active that kill any enemies near you. The build has great mobility thanks to Reap and Transplant.Pros✔️ Good Tankiness✔️ Exceptional Sustain✔️ Insane Single Target DPS✔️ Deals Damage While MovingCons❌ Limited AoE❌ Mana Management❌ Unique Heavy Build❌ Generator/Spender PlaystyleUmbral Blades FalconerThe Umbral Blades Falconer calls down its Falcon from the skies to spawn Umbral Blades, which are Throwing Attacks that detonate into giant Cold explosions and decimate your enemies. Watch your foes die as you swiftly cruise through the battlefield with this amazing build that focuses on maximizing the destructible power of the Twilight Strike Node inside the Aerial Assault Skill.Thanks to the Cooldown Reduction from Free Lofting Bird, you can spam Explosive Trap to cycle through Dive Bomb and Aerial Assault almost non-stop. Smoke Bomb is your Movement Skill and Shadow generator.Pros✔️ Very Fast✔️ Amazing AoE✔️ Insane Burst Damage✔️ Multiple Defensive LayersCons❌ Cooldown Based DPS❌ Mana Management❌ Can't Leech From Crits❌ Requires Uniques to ShineFrostbite Frost Claw SorcererThe Frostbite Frost Claw Sorcerer strides into battle wielding the potent Frost Claw as its Skill of choice. With each swift cast, Frost Claw releases a barrage of projectiles, all capable of converging upon a single target. This relentless assault not only deals Damage but also has the power to Freeze foes in their tracks. Thanks to its high Cast Speed, it can maintain a relentless offensive assault, while also accumulating Ward to fortify its Defenses.This is an Ailment-based build that scales Frostbite and Cold Damage Over Time. Its Damage primarily comes from Frost Claw, Elemental Nova, and Ice Barrage. The build also has amazing CC thanks to the ability to Freeze enemies reliably.Pros✔️ Easy to Play✔️ Incredible CC✔️ Flashy Visuals✔️ High Scaling PotentialCons❌ Visual Clutter❌ Average Survivability❌ Requires Uniques to Shine❌ Mana Intensive Without Proper SetupLast Epoch 1.2 Tier ListsIf you want to see the best performing builds of the patch, here are Maxroll’s Tierlists to help you decide what to play depending on your goals and taste! The best ones stand in S-Tier and are expected to outperform others. Choose wisely, read the guides, and defeat all content!CorruptionMaxroll’s Corruption Tier List categorizes the performance of all builds against any high-end challenges, taking into account all their strengths and weaknesses. Builds are compared based on their Endgame Planner Setup.BossingUse Maxroll's Bossing Tier List to see which builds excel at taking down Monolith Bosses, Dungeons Bosses, and Pinnacle Content. Maxroll considers many factors when placing the builds but mainly focuses on Overall reliable DPS. Builds are compared based on their Endgame Planner Setup.HardcoreMaxroll's Hardcore Tier List ranks a build's suitability for the optional permadeath mode “Hardcore”. Therefore the builds are primarily ranked based on their raw survivability and reliability, even on their Starting Planner Setup. Unlike the other tier lists, the early game gearing is given higher weight here, because a single death means losing your character when playing on Hardcore.Speed FarmingMaxroll's Speed Farming Tier List ranks a build's ability to speed through content. Naturally, speed is the main factor, which means the ability to AoE down mobs, move quickly, and maintain this speed even at high corruption. Builds are compared based on their Endgame Planner Setup.This article was created by IGN Staff with help from the Maxroll Last Epoch Team.

This article contains spoilers for Daredevil: Born Again season 1 episode 9. While some feared the day would never come, Netflix’s Daredevil has been ‘born again’ in the Marvel Cinematic Universe. Although it took nearly seven years and a massive creative overhaul, Daredevil: Born Again finally brings the Man Without Fear into the MCU full-time. As things came to a head in a bone-breaking finale that included some surprise character returns and plenty of teases about the future of Charlie Cox’s Daredevil, it’s good news that the show was renewed for season 2 before it even aired its first episode.  After a troubled production behind the cameras, showrunner Dario Scardapane has already described Born Again season 2 as a “better-oiled machine,” with the next eight episodes expected to further cement this more grounded corner of the MCU. Like season 1 felt like a (very) loose adaptation of 2021’s “Devil’s Reign” comic arc, the finale “Straight to Hell” seemingly sets up another iconic Daredevil arc from Marvel Comics. Daredevil: Born Again episode 9 expanded on the unsurprising reveal that Vanessa Fisk (Ayelet Zurer) was behind Bullseye’s (Wilson Bethel) hit on Foggy Nelson, with Matt Murdock’s dearly departed BFF threatening to blow the lid off Kingpin’s empire. Things came to a head when Wilson Fisk (Vincent D’Onofrio) used Bullseye’s apparent hit on him to trigger a city-wide lockdown and send his Anti-Vigilante Task Force into action to round up the masked menaces who’ve been trying to stop his criminal enterprises. It was here that Born Again kicked things up a gear, and after a brief cameo earlier in the season, there were cheers that Jon Bernthal’s Punisher was back to give his former friend/foe a hand.  We get one of those classic Daredevil fight scenes (although without a hallway moment), where it’s also revealed that Cole North (Jeremy Earl) was behind the death of Hector Ayala/White Tiger (played by the late Kamar de los Reyes). Despite Daredevil and Punisher getting the better of the AVTF, Matt’s apartment is blown to smithereens, and with it, his impressive collection of suits goes with it. Sadly, this means all the hard work of She-Hulk’s Luke Jacobson (Griffin Matthews) is wasted, including that red and gold suit. Set photos from Born Again season 2 have already confirmed that ‘Red’ will be getting a new look, with his black and red suit from the “Shadowland” comic arc making an appearance.  Fans have been eagerly anticipating the debut of this legendary comic book suit, with Variety asking D’Onofrio if we can expect season 2 to dive into the events of “Shadowland.” The man behind Big Willy told the outlet: “​​That’s a very good question. I don’t know if I can answer that question. But it’s a good one. You’re obviously following the story in a good way.” Similar to how the MCU has previously used the likes of “Civil War” as background reading material, and Born Again season 1 only used elements of “Devil’s Reign” like Fisk being Mayor of New York, don’t expect the next run of episodes to be a totally comic-accurate adaptation of “Shadowland.” After all, we’re not sure the MCU will include the part about Daredevil being possessed by a demon.  2010’s “Shadowland” has DD return from a trip to Japan as the new head of the Hand Ninja Clan and construct a prison/temple in the ruins of Hell’s Kitchen. With Matt’s building being reduced to rubble, it could become his remodeled base of operations. The big hint that the MCU’s take on “Shadowland” is coming is the cliffhanger of Bethel’s Bullseye. One of the comics’ major plot points includes Daredevil brutally killing Bullseye with his own sai to mirror how the villain killed Elektra. Swap in Born Again’s shooting of Foggy Nelson (Elden Henson), and the pieces neatly fall into place.  Although Shadowland Daredevil’s brutal methods put him at conflict with street-level heroes, including Luke Cage and Iron Fist, buzz about the Netflix-era Defenders making a return means the likes of Mike Colter and Finn Jones could join Daredevil’s Army that’s forming at Josie’s Bar in the season finale. It just so happens that “Shadowland” Daredevil is joined by Angela del Toro’s White Tiger, which would bring Camila Rodriguez’s underused Born Again alumni back into the fold. It feels like the first season had a distinct lack of Defenders, but with constant whispers of Krysten Ritter reprising her role as Jessica Jones, the prospect of more Netflix-era characters returning for a “Shadowland” arc is a promising one.  In terms of what else we can expect from Born Again season 2, the return of Deborah Ann Woll’s Karen Page has teased an interesting dynamic between Murdock and Frank Castle after it seems they’re almost on best buddy terms. Matt’s relationship with Heather Glenn (Margarita Levieva) is likely coming to an end after he woke up and asked for Page when in the hospital, while her new role as Fisk’s admin commissioner of mental health seemingly puts her on Team Kingpin moving forward. Even though Karen’s return might be setting Daredevil’s pulse racing, one key moment in the finale saw Matt take note of both Castle and Page’s racing hearts when they were with each other. Murdock and Page did briefly date in Daredevil season 2, but the return of Elektra Natchios (Élodie Yung) saw them stick to the ‘friends code.’ Although the trio might be busy trying to recruit an Anti-ATVF team in Born Again season 2, a twisted love triangle could be a well-placed spanner in the works.  We know Daredevil: Born Again season 2 is heading in a darker direction, with Cox praising Scardapane to Collider and explaining: “What he’s done so far with this new season is so cool and so different and so sinister, and I’m really, really excited about it.” Given this and D’Onofrio’s coy response, the fact that “Shadowland” is considered one of Daredevil’s darkest storylines (complete with a brooding suit) makes it a popular choice for season 2’s light reading. Either way, with Born Again hopefully back next year, the prospect that the Defenders could soon be back in action is an exciting one.  All nine episodes of Daredevil: Born Again season 1 are available to stream on Disney+ now. Join our mailing list Get the best of Den of Geek delivered right to your inbox!

American Airlines is the most popular airline in the US, and an especially great option for Apple fans thanks to free Apple Music and Apple TV+ offers. But unlike some of its competitors, it has always charged for Wi-Fi connectivity. Fortunately, that practice is about to end: free Wi-Fi is coming to American Airlines soon. American Airlines free Wi-Fi coming January 2026 In a new press release, American Airlines has announced that free connectivity is on the way. Starting next year in January 2026, the vast majority of American Airlines flights will get what it’s calling the “best-in-class high-speed Wi-Fi.” From the press release: Beginning in January 2026, American will offer complimentary Wi-Fi on all aircraft equipped with Viasat and Intelsat high-speed satellite connectivity, accounting for roughly 90% of its fleet. To prepare for the launch, American conducted a limited-time complimentary Wi-Fi test on select routes, allowing the airline to gauge the strength of its service, which surpassed performance expectations. American has partnered with AT&T to provide the new connectivity service. You do have to be an AAdvantage program member to get the Wi-Fi, but signing up is entirely free so that’s little hindrance to connectivity. American notes that once its new complimentary Wi-Fi launches, it will offer “free inflight connectivity on more planes than any other domestic carrier.” Are you excited for this new American Airlines perk? Let us know in the comments. Best iPhone accessories Add 9to5Mac to your Google News feed.  FTC: We use income earning auto affiliate links. More.You’re reading 9to5Mac — experts who break news about Apple and its surrounding ecosystem, day after day. Be sure to check out our homepage for all the latest news, and follow 9to5Mac on Twitter, Facebook, and LinkedIn to stay in the loop. Don’t know where to start? Check out our exclusive stories, reviews, how-tos, and subscribe to our YouTube channel

Tesla has been accused of manipulating the odometers in its cars to avoid repair responsibilities and warranty agreements.As The Street reports, a class-action lawsuit filed in February claims that Tesla has been trying to dodge warranty-related obligations by intentionally overstating the distances its vehicles travel.The plaintiff, a man in California who bought a used 2020 Model Y with 36,772 miles, noticed an "abnormal spike in average daily miles driven" — despite a "consistent driving routine" — while he was taking it to the shop multiple times for suspension repairs.As a result, his warranty expired after reaching the 50,000-mile threshold "well ahead of schedule," according to the lawsuit. However, after the expiry, the "reported average daily miles" suddenly dropped off, becoming "more closely aligned with his historical data on his other vehicles," making him suspicious that the EV maker was fudging the numbers."Tesla Inc. employs an odometer system that utilizes predictive algorithms, energy consumption metrics, and driver behavior multipliers that manipulate and misrepresent the actual mileage traveled by Tesla vehicles," the lawsuit reads. "In so doing, Defendants can, and do, accelerate the rate of depreciation of the value of Tesla vehicles and also the expiration of Tesla vehicle warranties to reduce or avoid responsibility for contractually required repairs as well as increase the purchase of its extended warranty policy."The news couldn't come at a worse time for the carmaker. Sales have tanked worldwide in large part due to CEO Elon Musk's abhorrent behavior, embrace of far-right extremism, and gutting of the federal government.Investors have also had enough. Tesla's share price formed a "death cross" this week, a chart pattern when a stock's short-term moving averages cross below its long-term averages.Tesla stock is down over 34 percent year to date, indicating widespread disillusionment and uncertainty over the company's long-term future.Rapid depreciation has become a huge problem, with the value of used Teslas tanking three times faster than any other brand.In short, a class-action lawsuit accusing it of fudging the odometer numbers is the last thing the carmaker needs — especially as it braces for a likely disastrous first-quarter earnings report.In many ways, what the plaintiff is alleging in the suit is nothing particularly new. Tesla owners have long noticed an unusual discrepancy between the mileage indicated on the odometer and the actual miles driven."Is Tesla inflating odometer to show more range, ding lessors with over miles and duck repairs under warranty?" one Reddit user wrote in a post last year. "As I return my leased Model Y I noticed the odometer is off by 20 percent vs my Lexus RZ."And it's not just "dodgy" mileage numbers.In a 2023 investigation, Reuters found that Tesla had "rigged the dashboard readouts in its electric cars to provide 'rosy' projections of how far owners can drive before needing to recharge."Drivers noticed they were getting far fewer miles out of their cars' batteries than what the EV maker had claimed.According to the investigation, Tesla had set up an entire "diversion team" to divert calls from complaining customers and deceive them into running a "remote diagnostics" scan, which more often than not determined that the battery was fine.While it remains unclear how the latest lawsuit will play out, it's certainly not a good look for a company whose brand has already been severely tarnished by its CEO's antics.Share This Article

Apr 16, 2025Ravie LakshmananEndpoint Security / Vulnerability Cybersecurity researchers have detailed four different vulnerabilities in a core component of the Windows task scheduling service that could be exploited by local attackers to achieve privilege escalation and erase logs to cover up evidence of malicious activities. The issues have been uncovered in a binary named "schtasks.exe," which enables an administrator to create, delete, query, change, run, and end scheduled tasks on a local or remote computer. "A [User Account Control] bypass vulnerability has been found in Microsoft Windows, enabling attackers to bypass the User Account Control prompt, allowing them to execute high-privilege (SYSTEM) commands without user approval," Cymulate security researcher Ruben Enkaoua said in a report shared with The Hacker News. "By exploiting this weakness, attackers can elevate their privileges and run malicious payloads with Administrators' rights, leading to unauthorized access, data theft, or further system compromise." The problem, the cybersecurity company said, occurs when an attacker creates a scheduled task using Batch Logon (i.e., a password) as opposed to an Interactive Token, causing the task scheduler service to grant the running process the maximum allowed rights. However, for this attack to work, it hinges on the threat actor acquiring the password through some other means, such as cracking an NTLMv2 hash after authenticating against an SMB server or exploiting flaws such as CVE-2023-21726. A net result of this issue is that a low-privileged user can leverage the schtasks.exe binary and impersonate a member of groups such as Administrators, Backup Operators, and Performance Log Users with a known password to obtain the maximum allowed privileges. The registration of a scheduled task using a Batch Logon authentication method with an XML file can also pave the way for two defense evasion techniques that make it possible to overwrite Task Event Log, effectively erasing audit trails of prior activity, as well as overflow Security Logs. Specifically, this involves registering a task with an author with the name, say, where the letter A is repeated 3,500 times, in the XML file, causing the entire XML task log description to be overwritten. This behavior could then be extended further to overwrite the whole "C:\Windows\System32\winevt\logs\Security.evtx" database. "The Task Scheduler is a very interesting component. Accessible by anyone willing to create a task, initiated by a SYSTEM running service, juggling between the privileges, the process integrities and user impersonations," Enkaoua said. "The first reported vulnerability is not only a UAC Bypass. It is far more than that: it is essentially a way to impersonate any user with its password from CLI and to obtain the maximum granted privileges on the task execution session, with the /ru and /rp flags." Found this article interesting? Follow us on Twitter  and LinkedIn to read more exclusive content we post. SHARE    

It’s easy to watch Star Wars. It is widely available on several different physical formats, and it’s readily available to rent digitally or stream on Disney+. But that assumes you’re willing to watch the version of Star Wars that Disney and Lucasfilm want you to watch.They only make available one of the “Special Edition” prints of Star Wars; a version of the film with updated digital effects and some additional scenes, which were first created by George Lucas in the late 1990s as part of the original Star Wars trilogy’s re-release in theaters ahead of the debut of the Star Wars Prequels. Lucas further tinkered with the Special Editions multiple times; the most recent change debuted with the film’s addition to Disney+, when suddenly Greedo suddenly yelled “Maclunkey!” for some reason before being shot by Han Solo in the Mos Eisley Cantina.Even before Lucas made the Special Editions he had tweaked the original theatrical print of Star Wars — for example adding “Episode IV —  A New Hope” to the opening text crawl around the time of The Empire Strikes Back. (Originally the top of the crawl simply read “Star Wars.”) So it has been almost impossible to see the version of Star Wars that played in theaters in 1977 — at least legally.Until this year. Per The Telegraph, an original theatrical print of the film will play at the BFI in a few months. The print was struck in London at a Technicolor lab there, and is apparently in terrific condition thanks to a special process that ensures its colors don’t fade over time.And according to their report, these screenings were permitted by the powers that be at Lucasfilm. (“Permission from Lucasfilm and Disney was required in order for us to screen the original cut at the BFI Film on Film Festival,” a BFI spokeswoman told The Telegraph, adding “We have a long-standing partnership with Lucasfilm and Disney and this permission was sought and obtained.”)LucasfilmLucasfilmloading...READ MORE: 10 Actors Who Turned Down Star Wars RolesRight now, the print is only screening twice at this one venue in England. Whether this event is a prelude to similar screenings elsewhere remains to be seen. But to me, this is the one big way out there that Disney could make a lot of money off a re-release of Star Wars. (They continue to re-release the various films in theaters; Revenge of the Sith is returning to theaters next month, for example.) After decades out of circulation, there is a lot of curiosity among longtime fans, and even some casual Star Wars viewers, see these versions.A wide re-release of the original prints today could have the same impact on Star Wars that the Special Editions had in 1997. (At the time, those cuts were not controversial. They were greeted gleefully by fans who got the chance to see Star Wars on the big screen. It was only later, as it became clear that those were now the only versions of the trilogy that you could watch, that people turned on them.)The rare original Star Wars print will screen twice on opening night of the BFI’s Film on Film Festival on June 12. For more information, check out the BFI’s website.Get our free mobile app20 Actors You Forgot Were in Star Wars

Anyone who crammed for exams in college knows that an impressive ability to regurgitate information is not synonymous with critical thinking. The large language models (LLMs) first publicly released in 2022 were impressive but limited—like talented students who excel at multiple-choice exams but stumble when asked to defend their logic. Today's advanced reasoning models are more akin to seasoned graduate students who can navigate ambiguity and backtrack when necessary, carefully working through problems with a methodical approach. As AI systems thatlearn by mimicking the mechanisms of the human brain continue to advance, we're witnessing an evolution in models from rote regurgitation to genuine reasoning. This capability marks a new chapter in the evolution of AI—and what enterprises can gain from it. But in order to tap into this enormous potential, organizations will need to ensure they have the right infrastructure and computational resources to support the advancing technology. The reasoning revolution "Reasoning models are qualitatively different than earlier LLMs," says Prabhat Ram, partner AI/HPC architect at Microsoft, noting that these models can explore different hypotheses, assess if answers are consistently correct, and adjust their approach accordingly. "They essentially create an internal representation of a decision tree based on the training data they've been exposed to, and explore which solution might be the best." This adaptive approach to problem-solving isn’t without trade-offs. Earlier LLMs delivered outputs in milliseconds based on statistical pattern-matching and probabilistic analysis. This was—and still is—efficient for many applications, but it doesn’t allow the AI sufficient time to thoroughly evaluate multiple solution paths. In newer models, extended computation time during inference—seconds, minutes, or even longer—allows the AI to employ more sophisticated internal reinforcement learning. This opens the door for multi-step problem-solving and more nuanced decision-making. To illustrate future use cases for reasoning-capable AI, Ram offers the example of a NASA rover sent to explore the surface of Mars. "Decisions need to be made at every moment around which path to take, what to explore, and there has to be a risk-reward trade-off. The AI has to be able to assess, 'Am I about to jump off a cliff? Or, if I study this rock and I have a limited amount of time and budget, is this really the one that's scientifically more worthwhile?'" Making these assessments successfully could result in groundbreaking scientific discoveries at previously unthinkable speed and scale. Reasoning capabilities are also a milestone in the proliferation of agentic AI systems: autonomous applications that perform tasks on behalf of users, such as scheduling appointments or booking travel itineraries. "Whether you're asking AI to make a reservation, provide a literature summary, fold a towel, or pick up a piece of rock, it needs to first be able to understand the environment—what we call perception—comprehend the instructions and then move into a planning and decision-making phase," Ram explains. Enterprise applications of reasoning-capable AI systems The enterprise applications for reasoning-capable AI are far-reaching. In health care, reasoning AI systems could analyze patient data, medical literature, and treatment protocols to support diagnostic or treatment decisions. In scientific research, reasoning models could formulate hypotheses, design experimental protocols, and interpret complex results—potentially accelerating discoveries across fields from materials science to pharmaceuticals. In financial analysis, reasoning AI could help evaluate investment opportunities or market expansion strategies, as well as develop risk profiles or economic forecasts. Armed with these insights, their own experience, and emotional intelligence, human doctors, researchers, and financial analysts could make more informed decisions, faster. But before setting these systems loose in the wild, safeguards and governance frameworks will need to be ironclad, particularly in high-stakes contexts like health care or autonomous vehicles. "For a self-driving car, there are real-time decisions that need to be made vis-a-vis whether it turns the steering wheel to the left or the right, whether it hits the gas pedal or the brake—you absolutely do not want to hit a pedestrian or get into an accident," says Ram. "Being able to reason through situations and make an ‘optimal’ decision is something that reasoning models will have to do going forward." The infrastructure underpinning AI reasoning To operate optimally, reasoning models require significantly more computational resources for inference. This creates distinct scaling challenges. Specifically, because the inference durations of reasoning models can vary widely—from just a few seconds to many minutes—load balancing across these diverse tasks can be challenging. Overcoming these hurdles requires tight collaboration between infrastructure providers and hardware manufacturers, says Ram, speaking of Microsoft’s collaboration with NVIDIA, which brings its accelerated computing platform to Microsoft products, including Azure AI. "When we think about Azure, and when we think about deploying systems for AI training and inference, we really have to think about the entire system as a whole," Ram explains. "What are you going to do differently in the data center? What are you going to do about multiple data centers? How are you going to connect them?" These considerations extend into reliability challenges at all scales: from memory errors at the silicon level, to transmission errors within and across servers, thermal anomalies, and even data center-level issues like power fluctuations—all of which require sophisticated monitoring and rapid response systems. By creating a holistic system architecture designed to handle fluctuating AI demands, Microsoft and NVIDIA’s collaboration allows companies to harness the power of reasoning models without needing to manage the underlying complexity. In addition to performance benefits, these types of collaborations allow companies to keep pace with a tech landscape evolving at breakneck speed. "Velocity is a unique challenge in this space," says Ram. "Every three months, there is a new foundation model. The hardware is also evolving very fast—in the last four years, we've deployed each generation of NVIDIA GPUs and now NVIDIA GB200NVL72. Leading the field really does require a very close collaboration between Microsoft and NVIDIA to share roadmaps, timelines, and designs on the hardware engineering side, qualifications and validation suites, issues that arise in production, and so on." Advancements in AI infrastructure designed specifically for reasoning and agentic models are critical for bringing reasoning-capable AI to a broader range of organizations. Without robust, accessible infrastructure, the benefits of reasoning models will remain relegated to companies with massive computing resources. Looking ahead, the evolution of reasoning-capable AI systems and the infrastructure that supports them promises even greater gains. For Ram, the frontier extends beyond enterprise applications to scientific discovery and breakthroughs that propel humanity forward: "The day when these agentic systems can power scientific research and propose new hypotheses that can lead to a Nobel Prize, I think that's the day when we can say that this evolution is complete.” To learn more, please read Microsoft and NVIDIA accelerate AI development and performance, watch the NVIDIA GTC AI Conference sessions on demand, and explore the topic areas of Azure AI solutions and Azure AI infrastructure. This content was produced by Insights, the custom content arm of MIT Technology Review. It was not written by MIT Technology Review’s editorial staff. This content was researched, designed, and written entirely by human writers, editors, analysts, and illustrators. This includes the writing of surveys and collection of data for surveys. AI tools that may have been used were limited to secondary production processes that passed thorough human review.

As recommended by a Brit.

April 16, 20255 min readAlan Turing’s Lost Work Could Reveal How Tigers Got Their StripesThe world-famous mathematician Alan Turing found a possible mechanism for iconic animal patterns thanks to differential equationsBy Manon Bischoff edited by Daisy YuhasSumatran Tiger (Panthera tigris sumatrae). Juniors Bildarchiv GmbH/Alamy Stock PhotoToday mathematician Alan Turing is world-famous because he helped the Allies achieve victory against the Axis powers by deciphering an encryption that was considered unbreakable. That story inspired the 2014film The Imitation Game. Turing’s cryptographic work remained under wraps until the 1970s, however, so his incredible achievements only became known after his death.During his lifetime, Turing was known among certain experts. He developed the mathematical model of a computer and explained which mathematical quantities it could calculate—and which tasks would exceed even the most sophisticated algorithms. He is also well known for a test that he developed, later named after him, that assesses how “human” artificial intelligence appears to be. For instance, if people cannot tell whether they are chatting to a real person or an AI, then the machine has passed the Turing test.The list of Turing’s scientific contributions is long. But one area of his research is rarely mentioned: his work on mathematical biology that dealt with the formation of patterns. He was interested in the question of how animals develop their impressive stripes and spots, and he was convinced that there must be a mechanism by which pigments in skin cells arrange themselves into these patterns.On supporting science journalismIf you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.How Does the Tiger Get Its Stripes?When I first heard about this, I was puzzled. One of my physics professors mentioned a link between abstract mathematical operators and a tiger’s stripes in a first-semester lecture, a connection that made me and my fellow students laugh rather than think. After all, what could the pattern of a tiger’s skin have to do with abstract mathematics? Until then, I had assumed that some complex biochemical processes led to the tiger’s impressive patterns of dots and stripes—not something that could be represented by a tensor (a kind of high-dimensional table).I now realize that I lacked Turing’s imagination. According to his mother, even as a child, he was a dreamer who marveled at the natural world around him. He wanted to understand his surroundings. Mathematics lent itself as a language to reduce even the most complex relationships to the essentials. And so Turing found a very simple mechanism that could explain nature’s patterns.To understand Turing’s ideas, you first need a little biological background. A tiger’s coat pattern is already determined before it is born. In the embryo, pigment-producing cells emerge at the point where the spinal column will later develop. From there, they migrate through the entire body. Although research into these cells was lacking in Turing’s time, he recognized that there was a developmental process that formed skin patterns, and he wanted to find out how this occurred.It was impossible to model all the interacting molecules of an animal embryo. Moreover, Turing was not an expert in biochemistry. Therefore, as is usual for mathematicians, he started with a very simple model. He investigated how two different pigment-producing molecules, which he generally called morphogens, spread from cell to cell.A Story of Two MorphogensLet’s assume that one morphogen is responsible for the color black and another for orange. The more black or orange morphogens there are, the more of these molecules are generally produced. In addition, these two substances influence each other: the orange morphogens can inhibit the production of the black ones.Such an interplay is often found in ecology. For example, the black morphogens can be thought of as hares, which reproduce rapidly and thus attract foxes (the orange morphogens). The foxes, however, eat the hares and thus limit their population.This complex interaction can lead to a variety of situations. Sometimes small colonies of hares are kept in check by different foxes in the area. Translating that example to morphogens, one can imagine how a dotlike pattern similar to a cheetah’s fur could reflect that one morphogen has limited the spread of another.Turing did not use the fox-and-hare visualization when he described how morphogens could move through an embryo’s cells, but he did take into account the phenomenon of diffusion. If a cell harbors many black morphogens, for example, and a neighboring cell harbors few of them, then the molecules strive to move such that they are distributed as evenly as possible.All of these processes can be described by so-called differential equations. These equations contain one or more derivatives and can indicate how the number of morphogens per cell changes. Turing used the equations to investigate how two morphogens spread in the cells and what distribution occurs in the end. In doing so, he was able to adjust several parameters. To put this in terms of our animal analogy: How many foxes and hares are there at the beginning? How quickly do the hares reproduce, and how many foxes do they attract? How quickly do they spread? How are the cells arranged through which the molecules migrate? All these factors influence the pattern that emerges at the end.When Turing investigated this problem, he did not have a powerful computer at his disposal and had to carry out many of the calculations by hand. He solved the differential equations, and he recorded how the two morphogens were ultimately distributed in the cells and observed how patterns emerged. In some cases, there were stripes; in others, he found dots or sometimes spots similar to those on cows. (If you would also like to experiment with the Turing mechanism but don’t feel like doing lengthy calculations, here is a simulator.)As Turing discovered, the type of pattern depends on the arrangement of cells. Stripes tend to form in smaller, elongated structures, while dots form on large surfaces. Many years later British mathematical biologist James Murray applied Turing’s idea to big cats. A grown tiger is not a small animal. But, per Turing’s theory, the cat’s patterning indicates that the distribution of morphogens takes place at a time when the tiger embryo is still very small. The situation is different in leopards. And both effects can be seen in cheetahs: their body is spotted, but their tail is striped.Nice Theory, but What Happens in Practice?Unfortunately, this work by Turing attracted little attention during his lifetime. Shortly after he published his research in 1952, the discovery of the DNA’s double-helix structure overshadowed everything else. It took about 20 years for biologists to rediscover Turing’s work, inspiring a new generation to test whether the Turing mechanism really does occur in nature. But the technologies required have only been available since the 2000s.To prove Turing’s hypothesis, the corresponding morphogens must be identified in animals. Although this is not easy, a few cases are now known. For example, scientists identified two proteins in mice that produce the striped structure of their palate, and the arrangement of the animals’ hair follicles appears to be influenced by two other proteins. In the same way, the coloration of zebra fish is probably caused by the Turing mechanism, but it involves a complex interaction of three rather than two morphogens. These examples show that Turing’s ideas are not limited to color patterns but also apply to other structures.And what about the tiger and its stripes? To date, morphogens have been most clearly detected in mice, which are much easier to study in the lab. But hopefully a future scientific strategy can be found to prove Turing’s mechanism beyond a doubt in big cats as well.This article originally appeared in Spektrum der Wissenschaft and was reproduced with permission.