 |
|
|
|||||
Hi again, and welcome back to Fast Companys Plugged In. On November 18, Google announced a new product. More precisely, it declared that it was ushering in a new erawhich is what tech companies do when they really want you to pay attention. The product in question is Gemini 3 Pro, the latest version of Googles LLM. Its not just the foundation of Googles ChatGPT-like chatbot, also called Gemini. It underlies vast quantities of features in flagship offerings such as Google Search, Gmail, and Android. It powers Antigravity, a new Google AI coding platform that debuted on the same day. And thanks to Google Cloud, the model is also available to third-party developers as an ingredient for their apps. In short, Gemini 3 Pro could hardly be more essential to Googles aspiration to be AIs most important player. As Google DeepMind CEO Demis Hassabis said in the announcement, the company sees it as a big step on the path toward AGIAI thats at least as capable as humans are at most cognitive tasks. Already, the announcement stated, Gemini 3 Pro demonstrates PhD-level reasoning. Google supported its claims with a table listing 20 AI benchmarks in which Gemini 3 Pro beatand often just plain trouncedGemini 2 Pro, OpenAIs GPT-5.1, and Anthropics Claude Sonnet 4.5. Humanitys Last Exam, for example, is a 2,500-question test covering mathematics, physics, the humanities, and other topics. Its designed to be remarkably difficult (hence the name) and there has been debate over whether its so nebulous that some of the theoretically correct answers are nuanced or wrong. According to Googles table, GPT-5.1 achieved a score of 26.5%, while Claude Sonnet 4.5 managed only 13.7%. By contrast, Gemini 3 Pro scored 37.5%, and did even better when allowed to do searches and run code, with a score of 45.8%. Outside the lab, Gemini 3 Pro has been received as enthusiastically as any new AI model I can remember. Ethan Mollick, one of my favorite providers of AI analysis based on hands-on usage, pronounced it very good. Others said it delivered on the great expectations that OpenAIs GPT-5 stoked but failed to satisfy. As I write, Ive been playing with the Gemini chatbot for just a few days. Much of that experience has been positive. Two writing assignments I gave it came out exceptionally well: an article on the future of the penny, and a detailed report on pricing for Digital Equipment Corp.s 1960s minicomputers. Its first pass at a simple vibe coding projectbuilding a search engine for Fast Companys Next Big Things in Techwas a bit of a mess, but when I explicitly put it into Build mode, it nailed the assignment in a few minutes. It also excelled at figuring out what was going on in an assortment of photos I uploaded. Yet for all thats gone right so far, I also encountered significant glitches with Gemini 3 Pro from almost the moment I tried it. They left me particularly wary of Googles blanket claims about the LLM being ready to help users learn anything and delivering responses that are smart, concise and direct, trading cliché and flattery for genuine insight. My interactions gone wrong were mostly about animation and comics, topics I turn to when fooling around with new AI because I know them well enough to spot mistakes. Asked about these subjects, Gemini repeatedly spewed hallucinations. For instance, when I asked if Walt Disney himself had ever worked on the Mickey Mouse comic strip, the LLM gave a correct answer (yes, though only briefly) but then volunteered a bunch of facts I hadnt asked for that werent actually factual. For example, it said that when the strips longtime artist retired, his final panel showed Mickey and Minnie gazing into a sunset, a subtle way of marking his departure. (No such strip appeared.) In a different chat, it manufactured an elaborate, entirely fictional backstory involving a different cartoonist also being a noted animation historian, which it told me was well-documented and recognized. It wasnt just that Gemini hallucinated. ChatGPT and Claude still do that, too. But more than other models, Gemini tended to compound its failures by gaslighting me. Helpfully pointing out its gaffes led to some of the strangest exchanges Ive had with AI since February 2023, when Microsofts Bing said it didnt want to talk to me anymore. (Full disclosure: I understand that AI is just stringing together a sequence of words it doesnt understand. All of its human-seeming qualities, be they impressive or annoying, are simulated. But its hard to write about them without slipping into a certain degree of anthropomorphizing!) Repeatedly, Gemini acknowledged its inaccuracies but insisted they were lore, common misconceptions, or examples of my own confusion. In one case, it eventually confessed: I have failed you in this conversation by fabricating details to cover up previous errors. In another instance, it continued to insist that it was right, providing citations that didnt even mention the topic at hand. Im not arguing that the fate of AI hangs on how much the technology knows about old cartoons. However, if any company is burdened with the responsibility of ensuring that its LLM is a trustworthy source of general information, its Google. That I tumbled into an abyss of AI-generated misinformation so quickly isnt an encouraging sign. Part of the problem lies in the fact that Gemini 3 Pro offers two modes, Fast and Thinking. The first is the default and was responsible for the prevarications I encountered, at least one of which involved it conflating two separate topics Id brought up. So far, Thinking mode has worked better in my experiments. But even the speediest of AI models should meet a baseline of accuracy and good behavior, at leastif theyre being presented as a way to learn anything. (Like many AI tools, the Gemini chatbot does carry a mistakes-are-possible disclaimer.) To repeat myself, Gemini 3 Pro is impressive in many ways. Still, its launch is yet another example of the AI industry presenting an overly rosy portrait of what it has achieved. It also underlines that benchmarks tell us only so much about a models real-world performance. When OpenAI introduced ChatGPT three years ago this month, it did so in a brief blog post that took pains to detail the bots limitations and avoid grand pronouncements about its future. Letting its breakthrough new product speak for itself turned out to be a pretty effective marketing strategy. Even as AIs giants jostle for bragging rights in what may be the most hypercompetitive tech category of all time, they should remember that lesson. Youve been reading Plugged In, Fast Companys weekly tech newsletter from me, global technology editor Harry McCracken. If a friend or colleague forwarded this edition to youor if you’re reading it on fastcompany.comyou can check out previous issues and sign up to get it yourself every Friday morning. I love hearing from you: Ping me at hmccracken@fastcompany.com with your feedback and ideas for future newsletters. I’m also on Bluesky, Mastodon, and Threads, and you can follow Plugged In on Flipboard. More top tech stories from Fast Company A battle against the AI oligarchy is brewing in this wealthy New York districtTwo congressional candidates have made AI a major issue in the campaign. Read More Crypto’s path to legitimacy depends on the industry itself, not just politiciansOnly an internal cultural shift and rigorous self-policing can deliver mainstream approval. Read More AI chatbots won’t save the media. But what powers them mightPublisher-built agents grounded in trusted archives may turn years of reporting into real products instead of just another chat widget. Read More This massive new data center is powered by used EV batteriesA new project from battery recycling startup Redwood Materials and data center builder Crusoe shows that it’s possible to build data centers cheaper and faster while also slashing emissions.Read More Why Trump’s AI diplomacy is doomed to failThis week, chips were on the menu in the White House Read More Even (especially) in the age of AI, here’s why I hire for character over skillBecause that’s what reveals true talent. Read More
Category:
E-Commerce
Before Wicked opened on Broadway in October 2003, the musicals production team took the show to the Curran Theatre in San Francisco for whats called an out of town tryout. The five-week run allowed the producers, writers, and director to work out the kinks ahead of the shows Broadway debut. During the San Francisco run, University of Southern California film student Jon M. Chu happened to be home for the weekend visiting his parents, who owned a Chinese restaurant called Chef Chus in Los Altos, California, just outside Silicon Valley. Chu was the youngest of five children growing up in a family that spent their free time playing instruments or going to the ballet, the opera, musicals, and the movies. It was the time of Michael Jordan on TV, and Steven Spielberg movies, Chu recalls. Michael Jackson videos were like mini musicals. Chu was raised on what he calls this beautiful idea of story, and went to film school with an inner theater geek driving his desire to learn the craft. So it wasnt a surprise when his mom, Ruth, suggested they catch the show at the Curran. Also at the show was Marc Platt, Wickeds producer, who spent those five weeks in San Francisco putting the finishing touches on what would eventually become one of Broadways biggest hits. Little did either of these men know that nearly 20 years later theyd pair up to bring Wicked to the big screen. I waited 20 some years to make the movie after I produced the stage show, Platt says. There were many reasons I waited; destiny was calling me to wait until Jon was available. Director Jon M. Chu on the set of Wicked: For Good with Ariana Grande as Glinda [Photo: Universal Studios] Their partnership is now one of Hollywoods great success stories. Together, Chu and Platt delivered a giant blockbuster that grossed $114 million in box-office sales on its opening weekend in November 2024 (it went on to gross nearly $750 million). Now theyre poised to do it again with the release of part two, Wicked: For Good. The Wicked films represent a big gamble for Universal Picturestwo lavish movie musicals released back-to-back with an estimated combined budget of roughly $300 million. All this in an era when studios are cutting costs and audiences are distracted by content that requires them to do nothing more than scroll their phones or stream TV from the living room couch. Still, Chu says he’s more optimistic about filmmaking than ever. Now is the moment, he says, to fight for big stories that can break through in a business thats increasingly driven by algorithms and a focus on the bottom line. I think if you want to fight that fight, youve got to play by the rules of this game, he says. You have to entertain the hell out of people. Jeff Goldblum (the Wizard of Oz) and Cynthia Erivo (Elphaba) in Wicked: For Good [Photo: Universal Studios] The Silicon Valley Showman meets the Hollywood Machine Chu grew up inside two great American experimentsthe immigrant dream and Silicon Valley innovation. His parents have run Chef Chus for decades. In fact, thats where he was gifted his first camera and editing software from some thoughtful customers who worked in the movie industry. In high school Chu convinced his teachers to allow him to turn in short edited videos rather than written papers, which provided him the opportunity to develop a fluency with the software hed inherited at the restaurant. He also landed gigs shooting weddings and bar mitzvahs in his hometown. He was an early adopter from adolescence, learning how to master After Effects and Pro Tools. In 2002, while at USC, Chu gained some notoriety for a short film called When the Kids Are Away, which he made with a grant from the Princess Grace Foundation. It was a full-blown movie musical with singing and dancing, all about how stay-at-home-mothers spend their time when children are in school and spouses are at work. To work on it with him, he pulled in film school classmate and budding cinematographer Alice Brooks. They would later reunite on the Wicked films. Chu had an early run of films that cemented his approach to musical storytelling. In his first studio film, 2008s Step Up 2 the Streets, Chu learned how to turn ideas into an actual production by marrying choreography and character. Then, as director of Justin Bieber: Never Say Never, Chu gained an understanding of how powerful internet culture can be as a storytelling mechanism. It wasnt until he directed Crazy Rich Asians, though, that Chu truly turned a corner in his career. The film allowed him to blend cultural nuance, authentic storytelling, and mass appeal. All of this leveled up into Wicked, which is Chus most ambitious film by far. The movie required him to stretch beyond his creative instincts. Donna Langley, chairperson of Universal Pictures, says Chu is as much a whip-smart executive as he is a creative: He thrives at the intersection of commerce and art. He is deeply empathetic, pragmatic, and innovative. He sees challenges as opportunities, is extremely methodical in his planning, and yet remains flexible in his execution. [Photo: Universal Studios] Storytelling as empathy Chu sees storytelling as a kind of transcendent currency in the current age of filmmaking. It is one of the most powerful empathy engines we have other than travel, he says. It helps, of course, that with Wicked Chu happens to be telling a uniquely magical story, in a land that feels so fantastical and vibrant you cannot help but be transported. But Chus singular power in bringing Wicked to life is his ability to think like a storyteller and work like an engineer. Chus creative system is deliberate and disciplined. He recalls how in his early years of editing, he devised a system for nonlinear storytelling logic that helped him organize his ideas. He would assemble thousands of screenshots, textures, and notes and organize them into folders. These folders became a creative pantry of sorts that he began drawing from when developing a new film. [Photo: Universal Studios] When Chu and cinematographer Brooks were classmates at USC they bonded over their love of musicals. I think both of us are really emotional storytellers, says Brooks, who worked with Chu on both In the Heights and Wicked (among others). It’s about breaking a story from the inside out. She describes Chus process as incredibly intentional. [Photo: Universal Studios] She and Chu move at their own pace during the early stages of production and planning, trusting one anothers instincts as they go. When we first get a script, we break down each scene with one word, emotional intention, and every single camera choice and lighting choice comes from those intentions, Brooks says. It’s a long marinating process of letting ideas grow first, and then the technical ideas come very much as a secondary. On the set of Wicked: For Good, Chu implemented a process that required the actors to speak their lines while rolling before official filming began. While rehearsing the song For Good, actors Ariana Grande and Cynthia Erivo began to ad lib, singing to one another through a closet door. It was totally unplanned, but also part of Chus usual process ahead of a scene, so he let it go and the captured moment ended up one of the most memorable in the film. Process is how genius happens, he says. [Photo: Universal Studios] During the development of a movie, Chu doesnt think about box-office sales. Instead, hes locked in on making sure the film is as entertaining as possible. He views movies as a portal into another world, where the audience can experience things from someone elses perspective. To me, that is what we need to protect most in our culture, Chu says. I feel great responsibility that I have a microphone to be in that space. Chu began shooting both Wicked films in 2022 before AI really became a thing. He and the movies team of more than 700 visual effects experts were deeply focused on building a tactile but imperfect world. He wanted tables that wobbled and doors with cracks in them. For Wicked to work, nothing could feel overly manufactured or make believe. [Photo: Universal Studios] He insisted the world be touchable, that we could feel the scratches and dirt, says Platt. That quality is what allows the audience to feel the high stakes in Glinda and Elphabas relationship, Chu says, noting that everything on set within 40 feet was physically built. The imperfections enabled a kind of intimacy with the audience that AI could never replicate. This couldnt feel like a dream, because we were talking about real things and were digging at real truth, he says. Platt says this unwavering commitment to emotional resonance is what makes Chu unique at this moment in time. Even when it was challenging to make changes, or we had disagreements with the writers, there was always joy in the process, Platt says. When your director feels that way, it permeates all those working on a production. But also it made me confident in our collaboration, and in the outcome. In many ways, Chu represents a blueprint for what the movie business will need as technology continues to infuse cinma. Chu believes that as AI becomes a bigger part of how creatives make their work, it will only put more value on human curiosity. In the end, he says, its about building something people can feeleven in a world made of pixels.
Category:
E-Commerce
AI has made us faster and more productive at work. It drafts our emails, summarizes our meetings, and even reminds us to take breaks. But heres the problem: in our rush to embrace AI, its quietly eroding our relationships and how we build human connections at work and in our everyday lives. People are increasingly using tools like ChatGPT to help them write, coach, and communicate. And many are also turning to it for therapy and relationship advice. The problem is, AI doesnt truly understand people as unique individuals. It can mimic empathy, but it cant understand it. It can predict tone, but it cant sense intent. The way we communicate with one person shouldnt be the same as the way we communicate with the next, yet thats exactly what happens when we hand over the nuances of being human to a machine. And its showing up at work: 82% of employees now report burnout, and 85% have experienced conflict at work. The majority trace it back to miscommunication, misunderstandings, or feeling unseen. AI is teaching us to write better, but not necessarily to understand better. Written communication has never been more polished. Yet the more we optimize our words, the more disconnected we seem to feel from one another. The importance of respecting human nuance AI can help us communicate, but it shouldnt act as a crutch. The real opportunity lies in using it as a mirror, which helps us better understand ourselves and the people we work with. Rather than replacing emotional intelligence (EQ), many teams are turning to personality science, such as the Five-Factor Model, to help leaders recognize how different teammates prefer feedback, how they handle stress, or why two colleagues interpret the same message in completely different ways. For teams, and for counselors and coaches, the goal is similar: not to have AI communicate for us, but to help us communicate better with each individual that we engage with. Because no two people hear, feel, communicate, or respond to information in the same way. And while the best communicators already know this instinctively, in todays era of chatbots and synthetic personas, we often abandon that awareness. We need to go back to giving each message, each meeting, and each moment the same level of consideration. Whos on the other side? What do they value? How do they process information or emotion? Leaders who take the time to personalize their communication build trust faster and resolve conflict sooner. When we adapt our style to meet people where they are, we only get better outcomes and make sure that people feel seen. Why we need to leverage EQ to optimize communications and outcomes Emotional intelligence isnt disappearing because people lack empathy. Its slipping because were letting machines do more of the communicating unilaterally. A new study by the Wharton School and GBK Collective found that 43% of leaders warn of skill atrophy as automation takes over routine work. This includes how we communicate. Leadership happens in the spaces algorithms cannot see: a pause in a meeting, the tension after a missed deadline, or the silence that signals someone doesnt feel safe speaking up. When we lose sensitivity to those cues, collaboration breaks down. Teams still communicate, but they stop connecting, and thats when misunderstandings quietly multiply into conflict and burnout. Heres how to keep the balance of efficiency and connection at work: Pause before you send. Before you hit “approve” on an AI-generated message, ask yourself: Does this sound like me? Does this reflect what the other person needs to hear? Sometimes, a call or short message will land better than a polished paragraph. Use AI for preparation, not delivery. Let technology help you structure the what, but you bring in the who with the persons history, style, and emotional context in mind. Listen and follow up. After sending feedback or direction, prioritize follow-up and check-ins to make sure you keep building the relationship, while listening and applying feedback. Prioritize taking a relationship-first approach. Remember that every person interprets messages differently. Landing the right tone and approach, depending on the relationship, shows respect and builds your connection. The leaders who thrive wont be those who use AI to talk more. Theyll be the ones who use it to listen more intentionally, understand people, and communicate with individuals uniquely. Because in the end, our progress, happiness, and success depend on the quality of relationships that we have with one another.
Category:
E-Commerce
All news |
||||||||||||||||||
|
||||||||||||||||||