A widespread SMS scam is targeting thousands of smartphone users in the US here's how to stay safe.

Bestselling author Malcolm Gladwell and award-winning creator Kenya Barris sat down with Fast Companys KC Ifeanyi at the FC Grill at SXSW. The trio discussed The Unusual Suspects, Gladwell and Barriss latest podcast series, as well as their business partnership and creative process. Stay tuned for key takeaways from their conversation and why The Unusual Suspects is a must-listen.https://www.audible.com/pd/The-Unusual-Suspects-with-Kenya-Barris-and-Malcolm-Gladwell-Audiobook/B0DNLMFCX7

Stools often serve as unassuming supporting characters: practical pieces meant for brief moments of rest rather than extended comfort. Their straightforward purpose has traditionally led to equally straightforward designs, particularly in the flat-packed furniture market where simplicity reigns supreme. Yet within these constraints lies an opportunity for innovation that the Walle stool brilliantly seizes, transforming a basic seating solution into something far more functional.The genius of the Walle stool lies in its recognition of how we actually use these transitional seating spaces. Those moments when we pause to catch our breath or briefly observe our surroundings arent merely empty times but are perfect opportunities for a few pages of reading. This thoughtfully designed metal stool accommodates this natural behavior by incorporating clever book storage directly into its structure, keeping 2 to 3 books within arms reach whenever you need a literary escape during your short rest.Designer: Luis GimenoEngineering elegance defines the Walles construction, with just four components that assemble with remarkable ease. The assembly requires only four screws, while the bottom shelf cleverly slots into place using notches in the opposing legs no additional hardware needed. This approach not only simplifies assembly but allows the entire stool to be packaged in an exceptionally compact form, making it ideal for shipping, storage, and transport in our increasingly space-conscious world.Visually, the Walle makes a subtle yet distinctive statement that draws inspiration from architectural forms. Its silhouette evokes the elegant simplicity of a miniature tower or traditional Torii gate, with seat and shelf edges that extend slightly beyond the frame. This intentional overhang creates visual interest through balanced asymmetry, demonstrating how thoughtful proportions can elevate even the most minimalist designs without requiring ornate detailing or excessive materials.The stools storage solutions are ingeniously integrated into its structural elements rather than added as afterthoughts. The side gutters accommodate slim reading materials like magazines or tablets, while the narrow space directly beneath the seat provides perfect housing for thicker books. Meanwhile, the bottom shelf offers versatile storage for additional reading material or personal items like bags and accessories, ensuring nothing needs to touch the floor.The stools metal construction offers durability that matches its timeless design approach. Unlike many contemporary furniture pieces that prioritize trends over longevity, the Walles clean lines and practical functionality ensure it remains relevant regardless of shifting design fashions. What makes the Walle truly remarkable is how it transforms a typically single-purpose piece into a multifunctional design object without compromising its primary function or aesthetic integrity. It represents a perfect marriage of pragmatism and thoughtful design, acknowledging that even our briefest moments of rest can be enhanced through access to reading materials.The post Flat-packed Metal Stool Concept Combines Minimalism and Literary Leisure first appeared on Yanko Design.

Sonos chief innovation officer Nick Millington says in an update published yesterday that he and his team are 100% focused on understanding and addressing issues with the Sonos app. Its the latest in a series of Millingtons posts sharing his teams progress as it tries to restore features that were left out of the apps controversial redesign in May last year.Heres what Millington, the original Sonos system architect who is tasked with fixing the app, has to say about the effort:The team and I remain 100% focused on two important priorities:1. understanding the root cause of every single customer issue, whether big or small, whether common or rare, and making sure the technical performance of the app meets or exceeds what you have come to expect from Sonos.2. closing gaps in the functionality and usability of the new app relative to what you enjoyed before, in a priority order that is as responsive as possible to the feedback we receive from you.Millington says users who have shared diagnostic information have sharpened the teams focus, leading to several improvements. In previous updates on the page going back to October, he has promised several features were on their way, such as alarm snoozing, better queue management, and playlist editing. He points out that you can once again snooze alarms, and says users can check the battery of portable Sonos Move and Roam speakers when theyre in standby. But the work has been slow-going and Sonos still hasnt added playlist editing back, nor the ability for a user to queue an entire folder from their local music library. According to Millington, both are on their way in upcoming releases.

Multi-modal Large Language Models (MLLMs) have demonstrated remarkable capabilities across various domains, propelling their evolution into multi-modal agents for human assistance. GUI automation agents for PCs face particularly daunting challenges compared to smartphone counterparts. PC environments present significantly more complex interactive elements with dense, diverse icons and widgets often lacking textual labels, leading to perception difficulties. Even advanced models like Claude-3.5 achieve only 24.0% accuracy in GUI grounding tasks. Also, PC productivity tasks involve intricate workflows spanning multiple applications with lengthy operation sequences and inter-subtask dependencies, causing dramatic performance declines where GPT-4os success rate drops from 41.8% at subtask level to just 8% for complete instructions.Previous approaches have developed frameworks to address PC task complexity with varying strategies. UFO implements a dual-agent architecture separating application selection from specific control interactions. Meanwhile, AgentS augments planning capabilities by combining online search with local memory. However, these methods demonstrate significant limitations in fine-grained perception and operation of on-screen texta critical requirement for productivity scenarios like document editing. In addition, they generally fail to address the complex dependencies between subtasks, resulting in poor performance when handling realistic intra- and inter-app workflows that characterize everyday PC usage.Researchers from MAIS, Institute of Automation, Chinese Academy of Sciences, China, School of Artificial Intelligence, University of Chinese Academy of Sciences, Alibaba Group, Beijing Jiaotong University, and School of Information Science and Technology, ShanghaiTech University introduce PC-Agent framework to address complex PC scenarios through three innovative designs. First, the Active Perception Module enhances fine-grained interaction by extracting locations and meanings of interactive elements via accessibility trees, while using MLLM-driven intention understanding and OCR for precise text localization. Second, Hierarchical Multi-agent Collaboration implements a three-level decision process (Instruction-Subtask-Action) where a Manager Agent decomposes instructions into parameterized subtasks and manages dependencies, a Progress Agent tracks operation history, and a Decision Agent executes steps with perception and progress information. Third, Reflection-based Dynamic Decision-making introduces a Reflection Agent that assesses execution correctness and provides feedback, enabling top-down task decomposition with bottom-up precision feedback across all four collaborating agents.PC-Agents architecture addresses GUI interaction through a formalized approach where an agent processes user instructions I, observations O, and history H to determine actions A. The Active Perception Module enhances element recognition using pywinauto to extract accessibility trees for interactive elements while employing MLLM-driven intention understanding with OCR for precise text localization. For complex workflows, PC-Agent implements Hierarchical Multi-agent Collaboration across three levels: the Manager Agent decomposes instructions into parameterized subtasks and manages dependencies; the Progress Agent tracks operation progress within subtasks; and the Decision Agent executes step-by-step actions based on environmental perception and progress information. This hierarchical division effectively reduces decision-making complexity by breaking complex tasks into manageable components with clear interdependencies.Experimental results demonstrate PC-Agents superior performance compared to both single and multi-agent alternatives. Single MLLM-based agents (GPT-4o, Gemini-2.0, Claude3.5, Qwen2.5-VL) consistently fail on complex instructions, with even the best performer achieving only 12% success rate, confirming that single-agent approaches struggle with lengthy operational sequences and complex dependencies. Multi-agent frameworks like UFO and AgentS show modest improvements but remain limited by perception deficiencies and dependency management issues. They struggle with fine-grained operations such as text editing in Word or proper data entry in Excel, and often fail to utilize information from previous subtasks. In contrast, PC-Agent significantly outperforms all previous methods, surpassing UFO by 44% and AgentS by 32% in success rate through its Active Perception Module and hierarchical multi-agent collaboration.This study introduces PC-Agent framework, a significant advancement in handling complex PC-based tasks through three key innovations. The Active Perception Module provides refined perception and operation capabilities, enabling precise interaction with GUI elements and text. The hierarchical multi-agent collaboration architecture effectively decomposes decision-making across instruction, subtask, and action levels, while reflection-based dynamic decision-making allows for real-time error detection and correction. Validation through the newly created PC-Eval benchmark with realistic, complex instructions confirms PC-Agents superior performance compared to previous methods, demonstrating its effectiveness in navigating the intricate workflows and interactive environments characteristic of PC productivity scenarios.Check outthe Paper and GitHub Page.All credit for this research goes to the researchers of this project. Also,feel free to follow us onTwitterand dont forget to join our80k+ ML SubReddit. Mohammad AsjadAsjad is an intern consultant at Marktechpost. He is persuing B.Tech in mechanical engineering at the Indian Institute of Technology, Kharagpur. Asjad is a Machine learning and deep learning enthusiast who is always researching the applications of machine learning in healthcare.Mohammad Asjadhttps://www.marktechpost.com/author/mohammad_asjad/Implementing Text-to-Speech TTS with BARK Using Hugging Faces Transformers library in a Google Colab environmentMohammad Asjadhttps://www.marktechpost.com/author/mohammad_asjad/Salesforce AI Releases Text2Data: A Training Framework for Low-Resource Data GenerationMohammad Asjadhttps://www.marktechpost.com/author/mohammad_asjad/Q-Filters: A Training-Free AI Method for Efficient KV Cache CompressionMohammad Asjadhttps://www.marktechpost.com/author/mohammad_asjad/Starter Guide For Running Large Language Models LLMs Parlant: Build Reliable AI Customer Facing Agents with LLMs (Promoted)

Knoxville has plenty of reliable internet options, but one provider stands out from the rest.

The owner of a data brokerage business recently put out a creepy-ass video in which he bragged about the degree to which his industry could collect and analyze data on the habits of billions of people. In 2019, the data broker Epsilon was acquired by French advertising conglomerate Publicis Groupe. Then, earlier this month, Publicis also acquired Lotame, another data and advertising firm, and announced it plans to integrate it with Epsilons business. At the time, Publicis CEO Arthur Sadoun said that the new corporate integration would allow his company to deliver personalized messaging at scale to some 91 percent of the internets adult web users. To deliver that kind of personalized messaging (i.e., advertising), Publicis must gather an extraordinary amount of information on the people it serves ads to. In a newly released video, Sadoun breaks down what kind of information his company claims to have its hands on. The video, which concerns a software program, CoreAI, shows the degree to which the company can analyze and predict the habits and behavior of individual consumers. To demonstrate this, Sadoun introduces the audience to Lola, a hypothetical young woman who represents the typical web user that Publicis now has data about. At a base level, we know who she is, what she watches, what she reads, and who she lives with, Sadoun says. Through the power of connected identity, we also know who she follows on social media, what she buys online and offline, where she buys, when she buys, and more importantly, why she buys. It gets worse. We know that Lola has two children and that her kids drink lots of premium fruit juice. We can see that the price of the SKU she buys has been steadily rising on her local retailers shelf. We can also see that Lolas income has not been keeping pace with inflation. With CoreAI, we can predict that Lola has a high propensity to trade down to private label, Sadoun says, meaning that the algorithm apprehends whether Lola is likely to start buying a cheaper brand of juice. If the software decides this is the case, the CoreAI algo can automatically start showing Lola ads for those reduced price juice brands, Sadoun says. Poor Lola. The big problem is that its not just Lola that is subjected to this all-encompassing corporate gazeit is, apparently, almost everybody on the internet.Thanks to CoreAI, we can do that with 91 percent of adults all around the world, the CEO brags. That amounts to nearly four billion people. Lena Cohen, a technologist with the Electronic Frontier Foundation, said that data brokers like Publicis collect as much information as they can about web users. The data broker industry is under-regulated, opaque, and dangerous, because as you saw in the video, brokers have detailed information on billions of people, but we know relatively little about them, Cohen said. You dont know what information a data broker has on you, who theyre selling it to, and what the people who buy your data are doing with it. Theres a real power/knowledge asymmetry. Meaningful regulation is notably absent from the data broker industry, Cohen said, allowing for companies to operate without much transparency or oversight. The U.S. needs comprehensive federal privacy law, Cohen said, while noting: I dont know how likely that is right now.Even when state-level privacy regulations are passed (such as the California Consumer Privacy Law), those cases are often not given enough focus or resources for the laws to be enforced effectively. Most government agencies dont have the resources to enforce privacy laws at the scale that theyre being broken, Cohen said. Cohen added that she felt online behavioral advertisingthat is, advertising that is based on an individual web users specific browsing activityshould be illegal. Thats not data that advertisers should have, she said, noting that contextual advertising (which is based on the content in a particular webpage) should be effective enough for companies. Banning behavioral ads would fundamentally change the financial incentive for online actors to constantly surveil web users and share their data with brokers, Cohen said.

GTC, Nvidias biggest conference of the year, begins Monday and runs till Friday in San Jose. TechCrunch will be on the ground covering the news as it happens and were expecting a healthy dose of announcements.CEO Jensen Huang will give a keynote address at the SAP Center on Tuesday at 10 a.m. Pacific, focusing on what else? AI and accelerating computing technologies, according to Nvidia. The company is also teasing reveals related to robotics, sovereign AI, AI agents, and automotive plus 1,000 sessions with 2,000 speakers and close to 400 exhibitors.Heres how to watch the Nvidia GTC 2025 keynote online, along with many other sessions, talks, and panels.So what do we expect to see at GTC? Well, Nvidia typically reserves a big chunk of the conference for GPU-related debuts. A new, upgraded iteration of the companys Blackwell chip lineup seems likely. During Nvidias most recent earnings call, Huang confirmed that the upcoming Blackwell B300 series, codenamed Blackwell Ultra, is slated for release in the second half of this year. In addition to higher computing performance, Blackwell Ultra cards pack more memory (288GB), an attractive feature for customers looking to run and train memory-hungry AI models. Rubin, Nvidias next-gen GPU series, is almost certain to get a mention at GTC alongside Blackwell Ultra. Due out in 2026, Rubin promises to deliver what Huang has described as a big, big, huge step up in computing power.Huang said during the aforementioned Nvidia earnings call that hed talk about post-Rubin products at GTC, as well. That could be Rubin Ultra GPUs, or perhaps the GPU architecture thatll come after the Rubin family. (The chips are named after Vera Rubin, the astronomer who discovered dark matter.)Beyond GPUs, Nvidia may illuminate its approach to recent quantum computing advancements. The company has scheduled a quantum day for GTC, during which itll host execs from prominent companies in the space to [map] the path toward useful quantum applications.One things for sure: Nvidia could use a win. Early Blackwell cards reportedly suffered from severe overheating issues, causing customers to cut their orders. U.S. export controls and fears of tariffs have massively depressed Nvidias stock price in recent months. At the same time, the success of Chinese AI lab DeepSeek, which developed efficient models competitive with models from leading AI labs, has prompted investors to worry about the demand for powerful GPUs like Blackwell.Huang has asserted that DeepSeeks rise to prominence will in fact be a net positive for Nvidia because itll accelerate the broader adoption of AI technology. He has also pointed to the growth of power-hungry so-called reasoning models like OpenAIs o1 as Nvidias next mountain to climb. To be clear, Nvidia isnt exactly hurting. The company reported a record-breaking quarter in February, notching $39.3 billion in revenue and projecting $43 billion in revenue for the subsequent quarter. While rivals such as AMD have begun to encroach on the companys territory, Nvidia still commands an estimated 82% of the GPU market. But Huang is reportedly feeling impatient to see AI applications that matter beyond the tech industry.

AUSTIN, TEXAS - MARCH 10: Mark Cuban speaks onstage during the 2025 SXSW Conference and Festival at ... [+] Hilton Austin on March 10, 2025 in Austin, Texas. (Photo by Julia Beverly/WireImage)WireImageSXSW is the ultimate FOMO conferenceyou cant see everything, no matter how hard you try. The XR Experience Exhibition, curated for over a decade by Blake Kammerdiener, once again stood out as one of the festivals highlights. With 31 projects15 in competition and 16 in the XR Experience Spotlightthis years showcase underscored the continued evolution of immersive storytelling, from mixed reality (MR) experiences in real-world settings to dome projections and large-scale, multi-user installations.In the Current of Being, by Cameron Kostopoulous, won the top prize in the 2025 XR competition. SXSWThe XR Experience Competition featured 15 world premieres, with Cameron Kostopoulos In the Current of Being taking home the top prize. This haptic VR experience, co-produced by Onassis ONX, placed users in the shoes of Carolyn Mercer, a survivor of electroshock conversion therapy. Through wearable haptic vests, sleeves, and gloves, participants felt Mercers heartbeat, breath, and the tremors she endured, making for one of the most visceral and empathy-driven experiences at SXSW.Kostopoulos, a previous SXSW Jury Award winner for Body of Mine (2023), continues to promote VR as a tool for advocacy. He is so into that first-person, walk-in-somebody-elses-shoes style of storytelling, Kammerdiener said. And you can really see his growth as an artist.The Jury Award went to Chlo Lees Reflections of a Little Red Dot, an open-world documentary exploring a country in transition. Judges praised its subtle yet deeply resonant storytelling, calling it a universal sense of place.Outside the competition, one of the most talked-about VR projects was Zhuzmos All I Know About Teacher Li, a documentary on censorship and digital activism in China. Winner of the Venice Immersive Special Jury Prize, it let participants throw origami planes onto a digital grid, unlocking banned reports from exiled activist Teacher Li. The experience of Chinese under draconian lock down, where residents starved and committed suicide, are hidden from the Chinese people. For example, violence in Shenzhen is unknown in Beijing due to strict censorship. Teacher Li switched to Twitter (now X) to avoid censorship and enough people knew about VPNs to avoid the heavy hand of the government. This inspired the blank paper protests which played a role in the lifting of quarantine in China.Eliza McNitt, Blake Kammerdiner, and AI/XR Podcast co-hosts Charlie Fink and Ted Schilowitz at SXSW ... [+] 2024.Charlie FinkThe immersive exhibition and accompanying XR conference track are overseen by Blake Kammerdiener, who spends the year talking to hundreds of creators, searching for projects that push boundaries. I was super pleased to bring in more multi-user, non-headset-based installations, he told The AI/XR Podcast before theSXSW XR exhibit, which ran three days, from March 9 - 11. This shift allowed more attendees to engage without the throughput limitations of individual VR headsets. Competition winners in 2023 and 2024 were non-headset experiences.In addition to making a wonderful concert experience, Resolution also won for best film poster. ResolutionOne of the most anticipated debuts was Resolution: A Cinephonic Rhapsody for the Soul, SXSWs first-ever dome theater experience. A custom-built dome hosted this psychedelic visual album, created by Scott Berman and Tim DeLaughter of The Polyphonic Spree. Reclining in camp chairs, audiences were enveloped by panoramic visuals and sound, redefining what an album-listening experience could be. Had it been released in 1977, I would have seen it ten times. It will be seen on a much larger scale at Cosm venus around the country. Resolution won the SXSWs film poster award.Another standout was Proof As If Proof Were Needed, a multi-user projection experience by Blast Theory and Ting-Tong Chang. Participants navigated a virtual Taiwanese home, triggering different scenes as they moved. Theyre tracking where people move on a grid, and what entices them to shift determines whats shown on-screen, Kammerdiener explained. The project won the Special Jury Award for its innovative non-linear storytelling.Ancestors is a 70-minute AI-driven experience by The Smartphone Orchestra, whose speciality is technology-driven experiences (the real world is immersive). Participants (there were around sixty of us) use an app, which asks you to take a selfie. A picture of a young woman is presented. This is my future daughter. I need to find my future wife, who has the same picture on her phone. The daughter is a blend of our selfies. I couldnt tell. We then were matched with another couple the same way, with a picture on our phones. At the end of the show, everyone in the audience realized we are all related, an exponentially large multi generation family. I wish I could give these people some kind of award. It surely was a contender.Writer and editor Jesse Max Barber is played by Bryonn Bain.Emblematic GroupOne of the most compelling experiences in the expo was The 1906 Atlanta Race Massacre, a mobile AR reconstruction of a tragic and largely forgotten event nearly erased from history. They used mobile AR on a tablet to show the horrific events of September, 1906, in Atlanta, where, for four days a white mob, incited by the race-baiting politicians and local newspapers, violently attacked the black community across the city. Using volumetric performances, hand-drawn Quill image and AI animations, this augmented reality project allows audiences to follow writer and editor Jesse Max Barber (Bryonn Bain) as he tries to tell the world this buried truth. This piece was made in a collaboration with the National Center for Civil and Human Rights and award-winning Emblematic Groups director Nonny de la Pea, who has more than a decade of AR/VR experience, alongside a consultancy with director Alton Glass and journalist Retha Hill.Bats thrive under Austin's river bridges, which inspired this unique made by a local artist. EchoVisionIn EchoVision, created by Jiabao Li, participants become bats and navigate a path in VR, using the echolocation of their voice to navigate a maze. AI translated the sound waves into light and movement, allowing users to experience the world as a bat does.Another innovative entry was theater company Tender Claws Face Jumping, which used eye-tracking on the Meta Quest Pro to let participants swap perspectives with what looked like ghosts (like the ones in Caspar), by locking gazes with them. You literally are seeing the world from another perspective. The ghosts are interactive, but they are not live performers featured in co-director Samantha Gormans previous VR theater productions The Under Presents and The Tempest. Pictured from l to r Gh0st and Atom, in Uncanny AlleyScreaming ColorUncanny Alley: A New Day is a live performance in VR by the Ferryman Collective, an ensemble of cross-media creators using performance to explore role-playing and drama, which has previously brought Gumball Dreams and Welcome to Respite to SXSWs XR Exhibition. The new show, Uncanny Alley, features two actors who seamlessly bring multiple characters to life. Audiences will converse with the actors, other audience members and interact with their virtual surroundings. "Our international cast is truly exceptional, featuring talent from film, television, gaming, and Broadway. They bring a diverse range of skills, and are some of the most experienced VR actors in the world today" says Whitton Frank, one of the project's producers.GatewayCharlie FinkThe Secret Life of Monsters The Gateway Experience, is casts you as a recruit in the GEU, a team tasked with exploring parallel realities from a mysterious station known as Gateway. Your first mission takes you into SOLUM, a world full of unique inhabitants and wonderful environments. Not a lot of straight up narrative entertainment at a show like this, so it was a welcome respite from the heavier topics.Traces: The Grief Processor used the Vision Pro. Its an interactive VR experience that explores loss, which invites users to upload a picture of someone they miss from their phone before the experience starts. It then becomes a hyer-personalized part of the show. The Apple Vision Pro was also used by The Last Practice and Currents. Now that its been in the market for a while, developers have had a chance to explore its potential as a storytelling device. It has more computing power than a Quest 3, which is an Android-based device, but the experiences so far dont demonstrate technology perceptibly different.Ways of Knowing: A Navajo Nuclear History was the only straight up 360 live-action documentary I saw at the show. The story of how uranium mining poisoned the Navajo Nation is a well-told tragic tale of epic injustice. A good filmmaker is going to tell a solid and effective story with this material. I have no idea why they would limit their audience using 360. Millions of people need to see this.SXSW PosterRistbandTraditionally the format of a concert is an artist on stage and the audience looking forward, however the Pivots show, with XR programmed by Venice Bleach, created a mixed reality concert and AV installation that plunges audiences into 360 digital worlds with live music in spatial audio envelopes the audience in a cocoon of sound.This isnt just a new way to present musicits a new revenue model for artists and a way for venues to expand their audience, says Ristband Director Anne McKinnon. Pivots and Venice Bleach have been booked for festivals from Shanghai to Vancouver even before releasing their first EPs, also touring art galleries, playing the opening night for film and innovation festivals, and working with companies like Music Team- a platform for independent artists to release and manage their music.AUSTIN, TEXAS - MARCH 10: Matthew McConaughey attends the "The Rivals Of Amziah King" world premiere ... [+] during the 2025 SXSW Conference and Festival at The Paramount Theatre on March 10, 2025 in Austin, Texas. (Photo by Tibrina Hobson/Getty Images)Getty ImagesWith an upcoming renovation of the Austin Convention Center, SXSW faces many challenges if it is to remain anchored in downtown Austin. There is the storied 6th street music scene, but it is buried in luxury hotels and high rises now. The media business, of which SXSW is a part, is struggling with high costs and shrinking audiences. The large international audience now has competing SXSW editions in Sydney and London to choose from. Kammerdiener is confident the show will go on. Its about discovery. Its about the people who come, the speakers, the networking. The beauty of SXSW is that its so big you can find your own thread through it. I hope hes right. XR needs the annual showcase SXSW has provided for the past ten years.

DAVOS, SWITZERLAND - JANUARY 23: CEO of the software company Salesforce, Marc Benioff attends the ... [+] 55th annual meeting of the World Economic Forum (WEF) in Davos, Switzerland on January 23, 2025. (Photo by Halil Sagirkaya/Anadolu via Getty Images)Anadolu via Getty ImagesAs 2025 unfolds, AI agents are emerging as the years defining business technology trend, with tech giants racing to develop tools that promise to transform how companies operate. However, amidst the excitement, a fundamental question remains: what exactly is an AI agent?The Business World's Next FrontierImagine technology that negotiates contracts, predicts supply chain collapses, and resolves customer complaintsall without human oversight. This isnt speculative fiction. AI agents are already doing it, rewriting business playbooks at a pace thats startling even Silicon Valleys optimists.OpenAIs latest API and SDK release isnt just an upgradeits checkmate on inefficiency. This new tool stack allows companies to deploy agents that independently scour internal databases, crawl websites, and execute workflows with surgical precision, effectively retiring the era of scripted chatbots and rigid automation.The business impact could be substantial. Companies that leveraged AI in 2024 saw 1.5 times higher revenue growth, and AI agents promise to extend these advantages. For instance, in customer service, conversational agents equipped with natural language processing are handling complex customer inquiries, with Bank of America's virtual assistant Erica efficiently managing over 1.5 billion customer interactions.But this breakneck progress conceals a fault line. The same boardrooms pouring billions into AI agents cant agree on what the term means. Are they autonomous decision-makers? Sophisticated workflow tools? Or something entirely new? The answer could determine which companies dominate the next decadeand which get left behind.The Definition DilemmaDespite the enthusiasm, the industry faces a significant challenge: no one can agree on what an AI agent actually is. Tech leaders including OpenAIs Sam Altman, Microsofts Satya Nadella, and Salesforces Marc Benioff have made bold predictions about AI agents joining the workforce and becoming the number one provider of digital labor in the world. To the same point, in an interview on occasion of the World Economic Forum 2025, Benioff claimed, Todays chief executives are the last generation to manage all-human workforces as companies increasingly adopt artificial intelligence.The push for AI adoption from top tech CEOs is loud. Their financial stake in it is obvious. The question is, however, how much of it is the organic result of underlying dynamics versus the forced message of business propaganda.As one expert noted, The concepts of AI agents and agentic workflows used to have a technical meaning, but about a year ago, marketers and a few big companies got a hold of them. This ambiguity creates both opportunities and challenges for businesses. While it allows for flexibility in customizing agents to specific needs, it also leads to misaligned expectations and difficulties in measuring ROI.The tech industry's track record suggests it's unlikely companies will coalesce around one definition anytime soon if ever. This definitional crisis echoes previous technology buzzwords that became diluted through overuse and marketing hype.Practical Applications Amid the HypeDespite the definitional chaos, businesses are finding concrete ways to implement these technologies. AI agents can execute time-consuming data-related tasks in minutes, creating high-quality content and extracting more value from meetings. They operate in three stages: perceiving their environment through models like natural language processing, applying algorithms to reason and make decisions, and continuously learning to refine their performance over time.Some companies are already seeing results. ServiceNow reports that 80% of its customer support cases are now handled without human intervention, utilizing analytical and generative AI to address common questions. When human workers remain involved, the company found that agentic AI shrank the time to handle complex cases by 52% in a two-week period.As businesses navigate this evolving landscape, the key challenge will be separating practical implementations from marketing hype. While 2025 may not deliver the fully autonomous AI workforce that some tech leaders have promised, organizations that develop clear internal definitions and strategic implementation plans for AI agents stand to gain significant competitive advantages as the technology matures.