Summer is the best time to start hosting parties, and if you’re planning on doing just that this year, you’re going to want to make sure you have everything you need to throw some ragers. What’s the most important thing you need for a great party in addition to the “vibe”? Oh, nothing. Just some amazing music and a way to share it. Case in point? A party speaker that’s up to the task that won’t force you to spend all of your savings. Sound like a plan Head to Amazon to get the JBL PartyBox for $249, down from its usual price of $349. That’s $100 off and a discount of 29%. See at Amazon Party all night for a great discount The JBL PartyBox is everything you need to get people moving and the party jumping. And it’s powerful enough to fill your space with 100 watts of signature JBL sound. It’s also compact enough to carry over your shoulder. It even comes with a padded strap so you can sling it across your back like a messenger bag. That makes it convenient for beach trips, park hangs, or just moving from room to room too. It also includes a built-in bottle opener, which sounds like a gimmick until you actually need one mid-party and realize how smart that feature is. Because someone’s going to have to bring a drink with a lid that doesn’t twist off, and that’s gonna harsh your vibe. This is more than just a speaker. It’s also a light show in a box, which you’ll soon find out when you start using it in earnest. Whether it’s a quiet night with friends or a full-on celebration, it’s a small touch that makes a big visual impact. And if you’re the karaoke type or know someone who is, you’ll love the wireless microphone that comes included. Tweak the bass, treble, and even echo effects, which makes this speaker double as your own personal DJ booth or karaoke machine. It’s also made to handle real life. With IPX4 splash resistance, it can stand up to light rain, poolside splashes, or accidental drink spills. The 6-hour battery life keeps things going long enough for most parties, and if you’ve got another compatible JBL speaker, you can pair them together for an even bigger sound experience. All this for under $300? You’ll want to get yours as soon as possible before the heat wave and the summer parties begin. See at Amazon

Last year, China saw a boom in foundation models, the do-everything large language models that underpin the AI revolution. This year, the focus has shifted to AI agents—systems that are less about responding to users’ queries and more about autonomously accomplishing things for them.  There are now a host of Chinese startups building these general-purpose digital tools, which can answer emails, browse the internet to plan vacations, and even design an interactive website. Many of these have emerged in just the last two months, following in the footsteps of Manus—a general AI agent that sparked weeks of social media frenzy for invite codes after its limited-release launch in early March.  These emerging AI agents aren’t large language models themselves. Instead, they’re built on top of them, using a workflow-based structure designed to get things done. A lot of these systems also introduce a different way of interacting with AI. Rather than just chatting back and forth with users, they are optimized for managing and executing multistep tasks—booking flights, managing schedules, conducting research—by using external tools and remembering instructions.  China could take the lead on building these kinds of agents. The country’s tightly integrated app ecosystems, rapid product cycles, and digitally fluent user base could provide a favorable environment for embedding AI into daily life.  For now, its leading AI agent startups are focusing their attention on the global market, because the best Western models don’t operate inside China’s firewalls. But that could change soon: Tech giants like ByteDance and Tencent are preparing their own AI agents that could bake automation directly into their native super-apps, pulling data from their vast ecosystem of programs that dominate many aspects of daily life in the country.  As the race to define what a useful AI agent looks like unfolds, a mix of ambitious startups and entrenched tech giants are now testing how these tools might actually work in practice—and for whom. Set the standard It’s been a whirlwind few months for Manus, which was developed by the Wuhan-based startup Butterfly Effect. The company raised $75 million in a funding round led by the US venture capital firm Benchmark, took the product on an ambitious global roadshow, and hired dozens of new employees.  Even before registration opened to the public in May, Manus had become a reference point for what a broad, consumer‑oriented AI agent should accomplish. Rather than handling narrow chores for businesses, this “general” agent is designed to be able to help with everyday tasks like trip planning, stock comparison, or your kid’s school project.  Unlike previous AI agents, Manus uses a browser-based sandbox that lets users supervise the agent like an intern, watching in real time as it scrolls through web pages, reads articles, or codes actions. It also proactively asks clarifying questions, supports long-term memory that would serve as context for future tasks. "Manus represents a promising product experience for AI agents," says Ang Li, cofounder and CEO of Simular, a startup based in Palo Alto, California, that’s building computer use agents, AI agents that control a virtual computer. “I believe Chinese startups have a huge advantage when it comes to designing consumer products, thanks to cutthroat domestic competition that leads to fast execution and greater attention to product details.” In the case of Manus, the competition is moving fast. Two of the most buzzy follow‑ups, Genspark and Flowith, for example, are already boasting benchmark scores that match or edge past Manus’s.  Genspark, led by former Baidu executives Eric Jing and Kay Zhu, links many small “super agents” through what it calls multi‑component prompting. The agent can switch among several large language models, accepts both images and text, and carries out tasks from making slide decks to placing phone calls. Whereas Manus relies heavily on Browser Use, a popular open-source product that lets agents operate a web browser in a virtual window like a human, Genspark directly integrates with a wide array of tools and APIs. Launched in April, the company says that it already has over 5 million users and over $36 million in yearly revenue. Flowith, the work of a young team that first grabbed public attention in April 2025 at a developer event hosted by the popular social media app Xiaohongshu, takes a different tack. Marketed as an “infinite agent,” it opens on a blank canvas where each question becomes a node on a branching map. Users can backtrack, take new branches, and store results in personal or sharable “knowledge gardens”—a design that feels more like project management software (think Notion) than a typical chat interface. Every inquiry or task builds its own mind-map-like graph, encouraging a more nonlinear and creative interaction with AI. Flowith’s core agent, NEO, runs in the cloud and can perform scheduled tasks like sending emails and compiling files. The founders want the app to be a “knowledge marketbase”, and aims to tap into the social aspect of AI with the aspiration of becoming “the OnlyFans of AI knowledge creators”. What they also share with Manus is the global ambition. Both Genspark and Flowith have stated that their primary focus is the international market. A global address Startups like Manus, Genspark, and Flowith—though founded by Chinese entrepreneurs—could blend seamlessly into the global tech scene and compete effectively abroad. Founders, investors, and analysts that MIT Technology Review has spoken to believe Chinese companies are moving fast, executing well, and quickly coming up with new products.  Money reinforces the pull to launch overseas. Customers there pay more, and there are plenty to go around. “You can price in USD, and with the exchange rate that’s a sevenfold multiplier,” Manus cofounder Xiao Hong quipped on a podcast. “Even if we’re only operating at 10% power because of cultural differences overseas, we’ll still make more than in China.” But creating the same functionality in China is a challenge. Major US AI companies including OpenAI and Anthropic have opted out of mainland China because of geopolitical risks and challenges with regulatory compliance. Their absence initially created a black market as users resorted to VPNs and third-party mirrors to access tools like ChatGPT and Claude. That vacuum has since been filled by a new wave of Chinese chatbots—DeepSeek, Doubao, Kimi—but the appetite for foreign models hasn’t gone away.  Manus, for example, uses Anthropic’s Claude Sonnet—widely considered the top model for agentic tasks. Manus cofounder Zhang Tao has repeatedly praised Claude’s ability to juggle tools, remember contexts, and hold multi‑round conversations—all crucial for turning chatty software into an effective executive assistant. But the company’s use of Sonnet has made its agent functionally unusable inside China without a VPN. If you open Manus from a mainland IP address, you’ll see a notice explaining that the team is “working on integrating Qwen’s model,” a special local version that is built on top of Alibaba’s open-source model.  An engineer overseeing ByteDance’s work on developing an agent, who spoke to MIT Technology Review anonymously to avoid sanction, said that the absence of Claude Sonnet models “limits everything we do in China.” DeepSeek’s open models, he added, still hallucinate too often and lack training on real‑world workflows. Developers we spoke with rank Alibaba’s Qwen series as the best domestic alternative, yet most say that switching to Qwen knocks performance down a notch. Jiaxin Pei, a postdoctoral researcher at Stanford’s Institute for Human‑Centered AI, thinks that gap will close: “Building agentic capabilities in base LLMs has become a key focus for many LLM builders, and once people realize the value of this, it will only be a matter of time." For now, Manus is doubling down on audiences it can already serve. In a written response, the company said its “primary focus is overseas expansion,” noting that new offices in San Francisco, Singapore, and Tokyo have opened in the past month. A super‑app approach Although the concept of AI agents is still relatively new, the consumer-facing AI app market in China is already crowded with major tech players. DeepSeek remains the most widely used, while ByteDance’s Doubao and Moonshot’s Kimi have also become household names. However, most of these apps are still optimized for chat and entertainment rather than task execution. This gap in the local market has pushed China’s big tech firms to roll out their own user-facing agents, though early versions remain uneven in quality and rough around the edges.  ByteDance is testing Coze Space, an AI agent based on its own Doubao model family that lets users toggle between “plan” and “execute” modes, so they can either directly guide the agent’s actions or step back and watch it work autonomously. It connects up to 14 popular apps, including GitHub, Notion, and the company’s own Lark office suite. Early reviews say the tool can feel clunky and has a high failure rate, but it clearly aims to match what Manus offers. Meanwhile, Zhipu AI has released a free agent called AutoGLM Rumination, built on its proprietary ChatGLM models. Shanghai‑based Minimax has launched Minimax Agent. Both products look almost identical to Manus and demo basic tasks such as building a simple website, planning a trip, making a small Flash game, or running quick data analysis. Despite the limited usability of most general AI agents launched within China, big companies have plans to change that. During a May 15 earnings call, Tencent president Liu Zhiping teased an agent that would weave automation directly into China’s most ubiquitous app, WeChat.  Considered the original super-app, WeChat already handles messaging, mobile payments, news, and millions of mini‑programs that act like embedded apps. These programs give Tencent, its developer, access to data from millions of services that pervade everyday life in China, an advantage most competitors can only envy. Historically, China’s consumer internet has splintered into competing walled gardens—share a Taobao link in WeChat and it resolves as plaintext, not a preview card. Unlike the more interoperable Western internet, China’s tech giants have long resisted integration with one another, choosing to wage platform war at the expense of a seamless user experience. But the use of mini‑programs has given WeChat unprecedented reach across services that once resisted interoperability, from gym bookings to grocery orders. An agent able to roam that ecosystem could bypass the integration headaches dogging independent startups. Alibaba, the e-commerce giant behind the Qwen model series, has been a front-runner in China's AI race but has been slower to release consumer-facing products. Even though Qwen was the most downloaded open-source model on Hugging Face in 2024, it didn’t power a dedicated chatbot app until early 2025. In March, Alibaba rebranded its cloud storage and search app Quark into an all-in-one AI search tool. By June, Quark had introduced DeepResearch—a new mode that marks its most agent-like effort to date.  ByteDance and Alibaba did not reply to MIT Technology Review’s request for comments. “Historically, Chinese tech products tend to pursue the all-in-one, super-app approach, and the latest Chinese AI agents reflect just that,” says Li of Simular, who previously worked at Google DeepMind on AI-enabled work automation. “In contrast, AI agents in the US are more focused on serving specific verticals.” Pei, the researcher at Stanford, says that existing tech giants could have a huge advantage in bringing the vision of general AI agents to life—especially those with built-in integration across services. “The customer-facing AI agent market is still very early, with tons of problems like authentication and liability,” he says. “But companies that already operate across a wide range of services have a natural advantage in deploying agents at scale.”

AGI will be embedded into humanoid robots, which makes white-collar and blue-collar jobs a target ... More for walking/talking automation.getty In today’s column, I examine the highly worrisome qualms expressed that the advent of artificial general intelligence (AGI) is likely to usurp white-collar jobs. The stated concern is that since AGI will be on par with human intellect, any job that relies principally on intellectual pursuits such as typical white-collar work will be taken over via the use of AGI. Employers will realize that rather than dealing with human white-collar workers, they can more readily get the job done via AGI. This, in turn, has led to a rising call that people should aim toward blue-collar jobs, doing so because (presumably) those forms of employment will not be undercut via AGI. Sorry to say, that misses the bigger picture, namely that AGI when combined with humanoid robots is coming not only for white-collar jobs but also blue-collar jobs too. It is a proverbial double-whammy when it comes to the attainment of AGI. Let’s talk about it. This analysis of an innovative AI breakthrough is part of my ongoing Forbes column coverage on the latest in AI, including identifying and explaining various impactful AI complexities (see the link here). Heading Toward AGI And ASI First, some fundamentals are required to set the stage for this weighty discussion. There is a great deal of research going on to further advance AI. The general goal is to either reach artificial general intelligence (AGI) or maybe even the outstretched possibility of achieving artificial superintelligence (ASI). AGI is AI that is considered on par with human intellect and can seemingly match our intelligence. ASI is AI that has gone beyond human intellect and would be superior in many if not all feasible ways. The idea is that ASI would be able to run circles around humans by outthinking us at every turn. For more details on the nature of conventional AI versus AGI and ASI, see my analysis at the link here. We have not yet attained AGI. In fact, it is unknown as to whether we will reach AGI, or that maybe AGI will be achievable in decades or perhaps centuries from now. The AGI attainment dates that are floating around are wildly varying and wildly unsubstantiated by any credible evidence or ironclad logic. ASI is even more beyond the pale when it comes to where we are currently with conventional AI. AGI Problem Only Half Seen Before launching into the primary matter at hand in this discussion, let’s contemplate a famous quote attributed to Charles Kettering, a legendary inventor, who said, “A problem well-stated is a problem half-solved.” I bring this up because those loud clamors right now about the assumption that AGI will replace white-collar workers are only seeing half of the problem. The problem as they see it is that since AGI is intellectually on par with humans, and since white-collar workers mainly use intellect in their work endeavors, AGI is going to be used in place of humans for white-collar work. I will in a moment explain why that’s only half of the problem and there is a demonstrative need to more carefully and fully articulate the nature of the problem. Will AGI Axiomatically Take White-Collar Jobs On a related facet, the belief that AGI will axiomatically replace white-collar labor makes a number of other related key assumptions. I shall briefly explore those and then come back to why the problem itself is only half-baked. The cost of using AGI for doing white-collar work will need to be presumably a better ROI choice over human workers. If not, then an employer would be wiser to stick with humans rather than employing AGI. There seems to often be an unstated belief that AGI is necessarily going to be a less costly route than employing humans. We don’t know yet what the cost of using AGI will be. It could be highly expensive. Indeed, some are worried that the world will divide into the AGI haves and AGI have-nots, partially due to the exorbitant cost that AGI might involve. If AGI is free to use, well, that would seem to be the nail in the coffin related to using human workers for the same capacity. Another angle is that AGI is relatively inexpensive in comparison to human labor. In that case, the use of AGI is likely to win over human labor usage. But if the cost of AGI is nearer to the cost of human labor (all in), or more so, then employers would rationally need to weigh the use of one versus the other. Note that when referring to the cost of human labor, there is more to that calculation than simply the dollar-hour labor rate per se. There are lots of other less apparent costs, such as the cost to manage human labor, the cost of dealing with HR-related issues, and many other factors that come into the weighty matter. Thus, an AGI versus human labor ROI will be more complex than it might seem at an initial glance. In addition, keep in mind that AGI would seemingly be readily switched on and off, and have other capacities that human labor would not equally tend to allow. The Other Half Is Coming Too Assume that by and large the advent of AGI will decimate the need for white-collar human labor. The refrain right now is that people should begin tilting toward blue-collar jobs as an alternative to white-collar jobs. This is a logical form of thinking in the sense that AGI as an intellectual mechanism would be unable to compete in jobs that involve hands-on work. A plumber needs to come to your house and do hands-on work to fix your plumbing. This is a physicality that entails arriving at your physical home, physically bringing and using tools, and physically repairing your faulty home plumbing. A truck driver likewise needs to sit in the cab of a truck and drive the vehicle. These are physically based tasks. There is no getting around the fact that these are hands-on activities. Aha, yes, those are physical tasks, but that doesn’t necessarily mean that only human hands can perform them. The gradual emergence of humanoid robots will provide an alternative to human hands. A humanoid robot is a type of robot that is built to resemble a human in form and function. You’ve undoubtedly seen those types of robots in the many online video recordings showing them walking, jumping, grasping at objects, and so on. A tremendous amount of active research and development is taking place to devise humanoid robots. They look comical right now. You watch those videos and laugh when the robot trips over a mere stick lying on the ground, something that a human would seldom trip over. You scoff when a robot tries to grasp a coffee cup and inadvertently spills most of the coffee. It all seems humorous and a silly pursuit. Keep in mind that we are all observing the development process while it is still taking place. At some point, those guffaws of the humanoid robots will lessen. Humanoid robots will be as smooth and graceful as humans. This will continue to be honed. Eventually, humanoid robots will be less prone to physical errors that humans make. In a sense, the physicality of a humanoid robot will be on par with humans, if not better, due to its mechanical properties. Do not discount the coming era of quite physically capable humanoid robots. AGI And Humanoid Robots Pair Up You might remember that in The Wonderful Wizard of Oz, the fictional character known as The Strawman lacked a brain. Without seeming to anthropomorphize humanoid robots, the current situation is that those robots typically use a form of AI that is below the sophistication level of modern generative AI. That’s fine for now due to the need to first ensure that the physical movements of the robots get refined. I have discussed that a said-to-be realm of Physical AI is going to be a huge breakthrough with incredible ramifications, see my analysis at the link here. The idea underlying Physical AI is that the AI of today is being uplifted by doing data training on the physical world. This also tends to include the use of World Models, consisting of broad constructions about how the physical world works, such as that we are bound to operate under conditions of gravity, and other physical laws of nature, see the link here. The bottom line here is that there will be a close pairing of robust AI with humanoid robots. Imagine what a humanoid robot can accomplish if it is paired with AGI. I’ll break the suspense and point out that AGI paired with humanoid robots means that those robots readily enter the blue-collar worker realm. Suppose your plumbing needs fixing. No worries, a humanoid robot that encompasses AGI will be sent to your home. The AGI is astute enough to carry on conversations with you, and the AGI also fully operates the robot to undertake the plumbing tasks. How did the AGI-paired humanoid robot get to your home? Easy-peasy, it drove a car or truck to get there. I’ve previously predicted that all the work on devising autonomous vehicles and self-driving cars will get shaken up once we have suitable humanoid robots devised. There won’t be a need for a vehicle to contain self-driving capabilities. A humanoid robot will simply sit in the driver’s seat and drive the vehicle. This is a much more open-ended solution than having to craft components that go into and onto a vehicle to enable self-driving. See my coverage at the link here. Timing Is Notable One of the reasons that many do not give much thought to the pairing of AGI with humanoid robots is that today’s humanoid robots seem extraordinarily rudimentary and incapable of performing physical dexterity tasks on par with human capabilities. Meanwhile, there is brazen talk that AGI is just around the corner. AGI is said to be within our grasp. Let’s give the timing considerations a bit of scrutiny. There are three primary timing angles: Option 1: AGI first, then humanoid robots. AGI is attained before humanoid robots are sufficiently devised. Option 2: Humanoid robots first, then AGI. Humanoid robots are physically fluently adept before AGI is attained. Option 3: AGI and humanoid robots arrive about at the same time. AGI is attained and at the same time, it turns out that humanoid robots are fluently adept too, mainly by coincidence and not due to any cross-mixing. A skeptic would insist that there is a fourth possibility, consisting of the possibility that we never achieve AGI and/or we fail to achieve sufficiently physically capable humanoid robots. I am going to reject that possibility. Perhaps I am overly optimistic, but it seems to me that we will eventually attain AGI, and we will eventually attain physically capable humanoid robots. I shall next respectively consider each of the three genuinely reasonable possibilities. Option 1: AGI First, Then Humanoid Robots What if we manage to attain AGI before we manage to achieve physically fluent humanoid robots? That’s just fine. We would indubitably put AGI to work as a partner with humans in figuring out how we can push along the budding humanoid robot development process. It seems nearly obvious that with AGI’s capable assistance, we would overcome any bottlenecks and soon enough arrive at top-notch physically adept humanoid robots. At that juncture, we would then toss AGI into the humanoid robots and have ourselves quite an amazing combination. Option 2: Humanoid Robots First, Then AGI Suppose that we devise very physically adept humanoid robots but have not yet arrived at AGI. Are we in a pickle? Nope. We could use conventional advanced AI inside those humanoid robots. The combination would certainly be good enough for a wide variety of tasks. The odds are that we would need to be cautious about where such robots are utilized. Nonetheless, we would have essentially walking, talking, and productive humanoid robots. If AGI never happens, oh well, we end up with pretty good humanoid robots. On the other hand, once we arrive at AGI, those humanoid robots will be stellar. It’s just a matter of time. Option 3: AGI And Humanoid Robots At The Same Time Let’s consider the potential of AGI and humanoid robots perchance being attained around the same time. Assume that this timing isn’t due to an outright cross-mixing with each other. They just so happen to advance on a similar timeline. I tend to believe that’s the most likely of the three scenarios. Here’s why. First, despite all the hubris about AGI being within earshot, perhaps in the next year or two, which is a popular pronouncement by many AI luminaries, I tend to side with recent surveys of AI developers that put the date around the year 2040 (see my coverage at the link here). Some AI luminaires sneakily play with the definition of AGI in hopes of making their predictions come true sooner, akin to moving the goalposts to easily score points. For my coverage on Sam Altman’s efforts of moving the cheese regarding AGI attainment, see the link here. Second, if you are willing to entertain the year 2040 as a potential date for achieving AGI, that’s about 15 years from now. In my estimation, the advancements being made in humanoid robots will readily progress such that by 2040 they will be very physically adept. Probably be sooner, but let’s go with the year 2040 for ease of contemplation. In my view, we will likely have humanoid robots doing well enough that they will be put into use prior to arriving at AGI. The pinnacle of robust humanoid robots and the attainment of AGI will roughly coincide with each other. Two peas in a pod.Impact Of Enormous Consequences In an upcoming column posting, I will examine the enormous consequences of having AGI paired with fully physically capable humanoid robots. As noted above, this will have a humongous impact on white-collar work and blue-collar work. There will be gargantuan economic impacts, societal impacts, cultural impacts, and so on. Some final thoughts for now. A single whammy is already being hotly debated. The debates currently tend to be preoccupied with the loss of white-collar jobs due to the attainment of AGI. A saving grace seems to be that at least blue-collar jobs are going to be around and thriving, even once AGI is attained. The world doesn’t seem overly gloomy if you can cling to the upbeat posture that blue-collar tasks remain intact. The double whammy is a lot more to take in. But the double whammy is the truth. The truth needs to be faced. If you are having doubts as a human about the future, just remember the famous words of Vince Lombardi: “Winners never quit, and quitters never win.” Humankind can handle the double whammy. Stay tuned for my upcoming coverage of what this entails.

Events Letterheads Per L'Horta: An Intimate International Meet The masses amass in Almàssera for an inspiring four days painting in the Valencian sun. Better Letters Jun 5, 2025 • 8 min read Letterheads Per L'Horta in Almàssera, Valencia, 1–4 May 2025. This time last month, over 45 guests from 11 countries were feeling the post-Letterheads blues after four days in the small town of Almàssera, just outside Valencia, Spain. Letterheads Per L'Horta was organised by Nico Barrios, and it was a wonderfully intimate experience, with a host of activities to enjoy and learn from.Something that made the event feel extra special was the involvement of people from the local community, who were just as much a part of it as those that had travelled from as far afield as Australia and Mexico to attend. This included bidding in the auction for a souvenir of the long weekend in May spent with friends, new and old.Almàssera and L'HortaAlmàssera is a small town set within a vast expanse of small-scale agricultural production. While each plot of land is known as a huerto (allotment), they are collectively referred to as horta, which doesn't really have a direct translation. The Horta Nord (North Horta) that surrounds Almàssera is the largest and best surviving example of this type of terrain.We were based in the town's Museu de l'Horta (Horta Museum), which consists of an old and a modern building with a yard between them that housed the panel jam area.A traditional alquería (farmhouse) in l'horta, a view down on the meet, and the tents protecting the panel jam area.Inside the modern building there was a selection of pieces from Juan Nava's 2022 Gráfica Urbana de Valencia (Urban Graphics of Valencia) exhibition. There was also a trip down memory lane for Valencian locals in the form of another exhibition, L'ombra de les lletres (the shadow of the letters), with photos of signs spanning the period 1880–2000.L'ombra de les lletres was originally curated by Tomàs Gorria in 2024. Pedal PowerAlmàssera, and the city of Valencia, are easily navigated by bicycle, which Nico used to facilitate a cycling tour of the old signs of l'horta. In addition to the stories of the individual companies advertised, he was also able to identify the painters responsible for some of the signs.The tour took guests into the heart of l'horta, which, as a largely agricultural area, boasts a surprising number of old and hand-painted signs.Panel JammingAfter a windy first day or so, the event was bathed in beautiful Mediterranean sunshine. The protective tents were essential, although those in the middle had to carefully manage their colour schemes in light of the red hue they cast across the easels.Getting painty in l'horta: Nathan Collis, Xis Gomes, Maria Cano, Mike Meyer, and Loughlin Brady Smith.Panels set to dry in the early evening sun.WorkshopsAcross the first three days, Thursday to Saturday, there was a series of lettering and calligraphy workshops that were also open to those outside of the Letterheads event proper.Pictured are workshops being led by Ester Gradolí, Juanjo López, and Joan Quiros.TV TimeThe meet was profiled in the local newspaper on the day before it opened, and then a TV crew turned up to cover proceedings.Local press coverage and Letterheads Per L'Horta host Nico Barrios being interviewed for the TV report. 0:00 /1:34 Letterheads Per L'Horta makes the news! If you look closely at the top of the paper that Daniel Esteve Carbonell is working on it says "Collons de rètol" (it's only a fucking sign) which clearly escaped the attention of the censors. Talks & DemosIn addition to workshops, the museum building also hosted a busy programme of talks. These were delivered by the Asociación de Diseñadores de la Comunidad Valenciana (Valencian Graphic Design Association), the errorerror.studio creative typography studio, graphic designer Juan Nava, and type designer Juanjo López.Juan Nava talking about the evolution of his Letras Recuperadas (Recovered Letters) project, previously featured here at bl.ag online.One of the highlights was hearing from veteran local sign painters Ricardo Moreno and Paco Vivó, both of whom appear in the Tipos Que Importan film that was screened. They were interviewed by Nico and brought a host of goods with them, including their sign kits, photographic portfolios, work samples, books, and other reference materials.Ricardo (in glasses) and Paco (with beard) were mobbed after talking about their lives on the brush in Valencia.Following the session, everyone moved outside to watch Paco Vivó paint one of the motifs that he produced many times in his career: the Pepsi-Cola bottle top.Paco Vivó painted his demonstration piece on a canvas which was subsequently sold in the auction.Meanwhile, over in the town square, David Vanderh had set up his screenprinting station to apply Nico's event design in a single colour to any material that the public brought to him.The live screenprinting was in just blue, while the official event t-shirt combined this with a striking orange.Panels on Show and on SaleOn the Sunday, a small exhibition was mounted with the panels that folks could bid on in the auction. This was an open invitation, with those from the neighbourhood stopping by to inspect and snag some goods.Panels getting ready for new owners in the charity auction.Panels by Veronika Skilte (Vermut), Joe Coleman (Mental on the Rental), Rachel E Millar (Rotulos, Gracias), and Victor Calligraphy.This panel (right) by Joe Coleman was inspired by the truck lettering that was a lucky incidental on the earlier cycling tour.The auction raised over 2,000€ in support of those affected by the devastating DANA floods in 2024.The assembled crowd were ready with open wallets as the auction got underway.The auction was expertly hosted by Mike Meyer and Nico Barrios, with Nil Muge logging all the winning bids and accounting for the cash payments.Thank YouAs with any event, the photos never show the challenges that must be overcome behind the scenes. Some of these were substantial but Nico took each one in his stride, maintaining a smile throughout. Thank you, Nico, for facilitating these special days that will live long in the collective memory.Letterheads Per L'Horta host Nico Barrios.Letterheads Per L'Horta was hosted by Nico Barrios with the support of the following organisations: AVV Carraixet d'Almàssera; Ajuntament d'Almàssera; BLAG; A.S. Handover; 1 Shot; ADCV; gráffica. Also check out the event's dedicated Instagram account, @letterheadsperlhorta, for even more photos and videos. More LetterheadsFuture Meets

“Is that concrete all around, or is it in my head?” asked Ian Hunter in “All the Young Dudes,” the song David Bowie wrote for Mott the Hoople in 1972. Concrete is all around us, and we haven’t quite wrapped our heads around it. It’s one of the indispensable materials of modernity; as we try to decarbonize the built environment, it’s part of the problem, and innovations in its composition may become part of the solution. Understanding its history more clearly, the Skyscraper Museum’s new exhibition in Manhattan implies, just might help us employ it better. Concrete is “the second most used substance in the world, after water,” the museum’s founder/director/curator Carol Willis told AN during a recent visit. For plasticity, versatility, and compressive strength, reinforced concrete is hard to beat, though its performance is more problematic when assessed by the metric of embodied and operational carbon, a consideration the exhibition acknowledges up front. In tall construction, concrete has become nearly hegemonic, yet its central role, contend Willis and co-curator Thomas Leslie, formerly of Foster + Partners and now a professor at the University of Illinois, Urbana-Champaign, is underrecognized by the public and by mainstream architectural history. The current exhibition aims to change that perception. The Skyscraper Museum in Lower Manhattan features an exhibition, The Modern Concrete Skyscraper, which examines the history of material choices in building tall towers. (Courtesy the Skyscraper Museum) The Modern Concrete Skyscraper examines the history of tall towers’ structural material choices, describing a transition from the early dominance of steel frames to the contemporary condition, in which most large buildings rely on concrete. This change did not happen instantly or for any single reason but through a combination of technical and economic factors, including innovations by various specialists, well-recognized and otherwise; the availability of high-quality limestone deposits near Chicago; and the differential development of materials industries in nations whose architecture grew prominent in recent decades. As supertalls reach ever higher—in the global race for official height rankings by the Council on Tall Buildings and Urban Habitat (CTBUH) and national, corporate, or professional bragging rights—concrete’s dominance may not be permanent in that sector, given the challenge of pumping the material beyond a certain height. (The 2,717-foot Burj Khalifa, formerly Burj Dubai, uses concrete up to 1,987 and steel above that point; Willis quotes SOM’s William Baker describing it as “the tallest steel building with a concrete foundation of 156 stories.”) For the moment, however, concrete is ahead of its chief competitors, steel and (on a smaller scale) timber. Regardless of possible promotional inferences, Willis said, “we did not work with the industry in any way for this exhibition.” “The invention of steel and the grid of steel and the skeleton frame is only the first chapter of the history of the skyscraper,” Willis explained. “The second chapter, and the one that we’re in now, is concrete. Surprisingly, no one had ever told that story of the skyscraper today with a continuous narrative.” The exhibition traces the use of concrete back to the ancient Roman combination of aggregate and pozzolana—the chemical formula for which was “largely lost with the fall of the Roman Empire,” though some Byzantine and medieval structures approximated it. From there, the show explores comparable materials’ revival in 18th-century England, the patenting of Portland cement by Leeds builder Joseph Aspdin in 1824, the proof-of-concept concrete house by François Coignet in 1856, and the pivotal development of rebar in the mid-19th century, with overdue attention to Ernest Ransome’s 1903 Ingalls Building in Cincinnati, then the world’s tallest concrete building at 15 stories and arguably the first concrete skyscraper. The exhibition includes a timeline that depicts concrete’s origins in Rome to its contemporary use in skyscraper construction. (Courtesy the Skyscraper Museum) Baker’s lectures, Willis reported, sometimes pose a deceptively simple question: “‘What is a skyscraper?’ In 1974, when the World Trade Center and Sears Tower are just finished, you would say it’s a very tall building that is built of steel, an office building in North America. But if you ask that same question today, the answer is: It’s a building that is mixed-use, constructed of concrete, and [located] in Asia or the Middle East.” The exhibition organizes the history of concrete towers by eras of engineering innovation, devoting special attention to the 19th- and early-20th-century “patent era” of Claude Allen Porter Turner (pioneer in flat-slab flooring and mushroom columns) and Henry Chandlee Turner (founder of Turner Construction), Ransome (who patented twisted-iron rebar), and François Hennebique (known for the re-inforced concrete system exemplified by Liverpool’s Royal Liver Building, the world’s tallest concrete office building when completed in 1911). In the postwar era, “concrete comes out onto the surface [as] both a structural material and aesthetic.” Brutalism, perhaps to some observers’ surprise, “does not figure very large in high-rise design,” Willis said, except for Paul Rudolph’s Tracey Towers in the Bronx. The exhibition, however, devotes considerable attention to the work of Pier Luigi Nervi, Bertrand Goldberg (particularly Marina City), and SOM’s Fazlur Khan, pioneer of the structural tube system in the 1960s and 1970s—followed by the postmodernist 1980s, when concrete could express either engineering values or ornamentation. The exhibition highlights a number of concrete towers, including Paul Rudolph’s Tracey Towers in the Bronx. (Courtesy the Skyscraper Museum) “In the ’90s, there were material advances in engineering analysis and computerization that helped to predict performance, and so buildings can get taller and taller,” Willis said. The current era, if one looks to CTBUH rankings, is dominated by the supertalls seen in Dubai, Shanghai, and Kuala Lumpur, after the Petronas Towers (1998) “took the title of world’s tallest building from North America for the first time and traumatized everybody about that.” The previous record holder, Chicago’s Sears (now Willis) Tower, comprised steel structural tubes on concrete caissons; with Petronas, headquarters of Malaysia’s national petroleum company of that name, a strong concrete industry was represented but a strong national steel industry was lacking, and as Willis frequently says, form follows finances. In any event, by the ’90s concrete was already becoming the standard material for supertalls, particularly on soft-soiled sites like Shanghai, where its water resistance and compressive strength are well suited to foundation construction. Its plasticity is also well suited to complex forms like the triangular Burj, Kuala Lumpur’s Merdeka 118, and (if eventually completed) the even taller Jeddah Tower, designed to “confuse the wind,” shed vortices, and manage wind forces. Posing the same question Louis Kahn asked about the intentions of a brick, Willis said, with concrete “the answer is: anything you want.” The exhibition is front-loaded with scholarly material, presenting eight succinct yet informative wall texts on the timeline of concrete construction. The explanatory material is accompanied by ample photographs as well as structural models on loan from SOM, Pelli Clarke & Partners, and other firms. Some materials are repurposed from the museum’s previous shows, particularly Supertall! (2011–12) and Sky High and the Logic of Luxury (2013–14). The models allow close examination of the Burj Khalifa, Petronas Towers, Jin Mao Tower, Merdeka 118, and others, including two unbuilt Chicago projects that would have exceeded 2,000 feet: the Miglin-Beitler Skyneedle (Cesar Pelli/Thornton Tomasetti) and 7 South Dearborn (SOM). The Burj, Willis noted, was all structure and no facade for a time: When its curtain-wall manufacturer, Schmidlin, went bankrupt in 2006, it “ended up going to 100 stories without having a stitch of glass on it,” temporarily becoming a “1:1 scale model of the structural system up to 100 stories.” Its prominence justifies its appearance here in two models, including one from RWDI’s wind-tunnel studies. Eero Saarinen’s only skyscraper, built for CBS in 1965 and also known as “Black Rock,” under construction in New York City. (Courtesy Eero Saarinen Collection, Manuscripts, and Archives, Yale University Library) The exhibition opened in March, with plans to stay up at least through October (Willis prefers to keep the date flexible), with accompanying lectures and panels to be announced on the museum’s website (skyscraper.org). Though the exhibition’s full textual and graphic content is available online, the physical models alone are worth a trip to the Battery Park City headquarters. Intriguing questions arise from the exhibition without easy answers, setting the table for lively discussion and debate. One is whether the patenting of innovations like Ransome bar and the Système Hennebique incentivized technological progress or hindered useful technology transfer. Willis speculated, “Did the fact that there were inventions and patents mean that competition was discouraged, that the competition was only in the realm of business, rather than advancing the material?” A critical question is whether research into the chemistry of concrete, including MIT’s 2023 report on the self-healing properties of Roman pozzolana and proliferating claims about “green concrete” using alternatives to Portland cement, can lead to new types of the material with improved durability and lower emissions footprints. This exhibition provides a firm foundation in concrete’s fascinating history, opening space for informed speculation about its future. Bill Millard is a regular contributor to AN.

Last year, China saw a boom in foundation models, the do-everything large language models that underpin the AI revolution. This year, the focus has shifted to AI agents—systems that are less about responding to users’ queries and more about autonomously accomplishing things for them.  There are now a host of Chinese startups building these general-purpose digital tools, which can answer emails, browse the internet to plan vacations, and even design an interactive website. Many of these have emerged in just the last two months, following in the footsteps of Manus—a general AI agent that sparked weeks of social media frenzy for invite codes after its limited-release launch in early March.  These emerging AI agents aren’t large language models themselves. Instead, they’re built on top of them, using a workflow-based structure designed to get things done. A lot of these systems also introduce a different way of interacting with AI. Rather than just chatting back and forth with users, they are optimized for managing and executing multistep tasks—booking flights, managing schedules, conducting research—by using external tools and remembering instructions.  China could take the lead on building these kinds of agents. The country’s tightly integrated app ecosystems, rapid product cycles, and digitally fluent user base could provide a favorable environment for embedding AI into daily life.  For now, its leading AI agent startups are focusing their attention on the global market, because the best Western models don’t operate inside China’s firewalls. But that could change soon: Tech giants like ByteDance and Tencent are preparing their own AI agents that could bake automation directly into their native super-apps, pulling data from their vast ecosystem of programs that dominate many aspects of daily life in the country.  As the race to define what a useful AI agent looks like unfolds, a mix of ambitious startups and entrenched tech giants are now testing how these tools might actually work in practice—and for whom. Set the standard It’s been a whirlwind few months for Manus, which was developed by the Wuhan-based startup Butterfly Effect. The company raised $75 million in a funding round led by the US venture capital firm Benchmark, took the product on an ambitious global roadshow, and hired dozens of new employees.  Even before registration opened to the public in May, Manus had become a reference point for what a broad, consumer‑oriented AI agent should accomplish. Rather than handling narrow chores for businesses, this “general” agent is designed to be able to help with everyday tasks like trip planning, stock comparison, or your kid’s school project.  Unlike previous AI agents, Manus uses a browser-based sandbox that lets users supervise the agent like an intern, watching in real time as it scrolls through web pages, reads articles, or codes actions. It also proactively asks clarifying questions, supports long-term memory that would serve as context for future tasks. "Manus represents a promising product experience for AI agents," says Ang Li, cofounder and CEO of Simular, a startup based in Palo Alto, California, that’s building computer use agents, AI agents that control a virtual computer. “I believe Chinese startups have a huge advantage when it comes to designing consumer products, thanks to cutthroat domestic competition that leads to fast execution and greater attention to product details.” In the case of Manus, the competition is moving fast. Two of the most buzzy follow‑ups, Genspark and Flowith, for example, are already boasting benchmark scores that match or edge past Manus’s.  Genspark, led by former Baidu executives Eric Jing and Kay Zhu, links many small “super agents” through what it calls multi‑component prompting. The agent can switch among several large language models, accepts both images and text, and carries out tasks from making slide decks to placing phone calls. Whereas Manus relies heavily on Browser Use, a popular open-source product that lets agents operate a web browser in a virtual window like a human, Genspark directly integrates with a wide array of tools and APIs. Launched in April, the company says that it already has over 5 million users and over $36 million in yearly revenue. Flowith, the work of a young team that first grabbed public attention in April 2025 at a developer event hosted by the popular social media app Xiaohongshu, takes a different tack. Marketed as an “infinite agent,” it opens on a blank canvas where each question becomes a node on a branching map. Users can backtrack, take new branches, and store results in personal or sharable “knowledge gardens”—a design that feels more like project management software (think Notion) than a typical chat interface. Every inquiry or task builds its own mind-map-like graph, encouraging a more nonlinear and creative interaction with AI. Flowith’s core agent, NEO, runs in the cloud and can perform scheduled tasks like sending emails and compiling files. The founders want the app to be a “knowledge marketbase”, and aims to tap into the social aspect of AI with the aspiration of becoming “the OnlyFans of AI knowledge creators”. What they also share with Manus is the global ambition. Both Genspark and Flowith have stated that their primary focus is the international market. A global address Startups like Manus, Genspark, and Flowith—though founded by Chinese entrepreneurs—could blend seamlessly into the global tech scene and compete effectively abroad. Founders, investors, and analysts that MIT Technology Review has spoken to believe Chinese companies are moving fast, executing well, and quickly coming up with new products.  Money reinforces the pull to launch overseas. Customers there pay more, and there are plenty to go around. “You can price in USD, and with the exchange rate that’s a sevenfold multiplier,” Manus cofounder Xiao Hong quipped on a podcast. “Even if we’re only operating at 10% power because of cultural differences overseas, we’ll still make more than in China.” But creating the same functionality in China is a challenge. Major US AI companies including OpenAI and Anthropic have opted out of mainland China because of geopolitical risks and challenges with regulatory compliance. Their absence initially created a black market as users resorted to VPNs and third-party mirrors to access tools like ChatGPT and Claude. That vacuum has since been filled by a new wave of Chinese chatbots—DeepSeek, Doubao, Kimi—but the appetite for foreign models hasn’t gone away.  Manus, for example, uses Anthropic’s Claude Sonnet—widely considered the top model for agentic tasks. Manus cofounder Zhang Tao has repeatedly praised Claude’s ability to juggle tools, remember contexts, and hold multi‑round conversations—all crucial for turning chatty software into an effective executive assistant. But the company’s use of Sonnet has made its agent functionally unusable inside China without a VPN. If you open Manus from a mainland IP address, you’ll see a notice explaining that the team is “working on integrating Qwen’s model,” a special local version that is built on top of Alibaba’s open-source model.  An engineer overseeing ByteDance’s work on developing an agent, who spoke to MIT Technology Review anonymously to avoid sanction, said that the absence of Claude Sonnet models “limits everything we do in China.” DeepSeek’s open models, he added, still hallucinate too often and lack training on real‑world workflows. Developers we spoke with rank Alibaba’s Qwen series as the best domestic alternative, yet most say that switching to Qwen knocks performance down a notch. Jiaxin Pei, a postdoctoral researcher at Stanford’s Institute for Human‑Centered AI, thinks that gap will close: “Building agentic capabilities in base LLMs has become a key focus for many LLM builders, and once people realize the value of this, it will only be a matter of time." For now, Manus is doubling down on audiences it can already serve. In a written response, the company said its “primary focus is overseas expansion,” noting that new offices in San Francisco, Singapore, and Tokyo have opened in the past month. A super‑app approach Although the concept of AI agents is still relatively new, the consumer-facing AI app market in China is already crowded with major tech players. DeepSeek remains the most widely used, while ByteDance’s Doubao and Moonshot’s Kimi have also become household names. However, most of these apps are still optimized for chat and entertainment rather than task execution. This gap in the local market has pushed China’s big tech firms to roll out their own user-facing agents, though early versions remain uneven in quality and rough around the edges.  ByteDance is testing Coze Space, an AI agent based on its own Doubao model family that lets users toggle between “plan” and “execute” modes, so they can either directly guide the agent’s actions or step back and watch it work autonomously. It connects up to 14 popular apps, including GitHub, Notion, and the company’s own Lark office suite. Early reviews say the tool can feel clunky and has a high failure rate, but it clearly aims to match what Manus offers. Meanwhile, Zhipu AI has released a free agent called AutoGLM Rumination, built on its proprietary ChatGLM models. Shanghai‑based Minimax has launched Minimax Agent. Both products look almost identical to Manus and demo basic tasks such as building a simple website, planning a trip, making a small Flash game, or running quick data analysis. Despite the limited usability of most general AI agents launched within China, big companies have plans to change that. During a May 15 earnings call, Tencent president Liu Zhiping teased an agent that would weave automation directly into China’s most ubiquitous app, WeChat.  Considered the original super-app, WeChat already handles messaging, mobile payments, news, and millions of mini‑programs that act like embedded apps. These programs give Tencent, its developer, access to data from millions of services that pervade everyday life in China, an advantage most competitors can only envy. Historically, China’s consumer internet has splintered into competing walled gardens—share a Taobao link in WeChat and it resolves as plaintext, not a preview card. Unlike the more interoperable Western internet, China’s tech giants have long resisted integration with one another, choosing to wage platform war at the expense of a seamless user experience. But the use of mini‑programs has given WeChat unprecedented reach across services that once resisted interoperability, from gym bookings to grocery orders. An agent able to roam that ecosystem could bypass the integration headaches dogging independent startups. Alibaba, the e-commerce giant behind the Qwen model series, has been a front-runner in China's AI race but has been slower to release consumer-facing products. Even though Qwen was the most downloaded open-source model on Hugging Face in 2024, it didn’t power a dedicated chatbot app until early 2025. In March, Alibaba rebranded its cloud storage and search app Quark into an all-in-one AI search tool. By June, Quark had introduced DeepResearch—a new mode that marks its most agent-like effort to date.  ByteDance and Alibaba did not reply to MIT Technology Review’s request for comments. “Historically, Chinese tech products tend to pursue the all-in-one, super-app approach, and the latest Chinese AI agents reflect just that,” says Li of Simular, who previously worked at Google DeepMind on AI-enabled work automation. “In contrast, AI agents in the US are more focused on serving specific verticals.” Pei, the researcher at Stanford, says that existing tech giants could have a huge advantage in bringing the vision of general AI agents to life—especially those with built-in integration across services. “The customer-facing AI agent market is still very early, with tons of problems like authentication and liability,” he says. “But companies that already operate across a wide range of services have a natural advantage in deploying agents at scale.”

Independent UK production company Chief has completed The Masterplan, a new documentary exploring the architectural and cultural transformation of Battersea Power Station in London. After decades of disuse, areas of the station were repurposed by WilkinsonEyre for housing and retail through a £9 billion ($12.2 billion USD) overhaul completed in 2022, with Foster + Partners designing a new Apple Store for the station in 2023. Directed by Chief’s Nate Camponi and commissioned by Stripe Communications and the Battersea Power Station Development Group, the new film presents an in-depth account of the power station’s history, from its origins as an industrial landmark to its revival as a mixed-use cultural destination. The film uses a combination of historical narrative, visual storytelling, and personal reflections to document the complex process of redeveloping the site and its evolving relationship with the city. Jason Flemyng as The Raconteur in The Masterplan. Image courtesy: ChiefA...

Last year, China saw a boom in foundation models, the do-everything large language models that underpin the AI revolution. This year, the focus has shifted to AI agents—systems that are less about responding to users’ queries and more about autonomously accomplishing things for them.  There are now a host of Chinese startups building these general-purpose digital tools, which can answer emails, browse the internet to plan vacations, and even design an interactive website. Many of these have emerged in just the last two months, following in the footsteps of Manus—a general AI agent that sparked weeks of social media frenzy for invite codes after its limited-release launch in early March.  These emerging AI agents aren’t large language models themselves. Instead, they’re built on top of them, using a workflow-based structure designed to get things done. A lot of these systems also introduce a different way of interacting with AI. Rather than just chatting back and forth with users, they are optimized for managing and executing multistep tasks—booking flights, managing schedules, conducting research—by using external tools and remembering instructions.  China could take the lead on building these kinds of agents. The country’s tightly integrated app ecosystems, rapid product cycles, and digitally fluent user base could provide a favorable environment for embedding AI into daily life.  For now, its leading AI agent startups are focusing their attention on the global market, because the best Western models don’t operate inside China’s firewalls. But that could change soon: Tech giants like ByteDance and Tencent are preparing their own AI agents that could bake automation directly into their native super-apps, pulling data from their vast ecosystem of programs that dominate many aspects of daily life in the country.  As the race to define what a useful AI agent looks like unfolds, a mix of ambitious startups and entrenched tech giants are now testing how these tools might actually work in practice—and for whom. Set the standard It’s been a whirlwind few months for Manus, which was developed by the Wuhan-based startup Butterfly Effect. The company raised $75 million in a funding round led by the US venture capital firm Benchmark, took the product on an ambitious global roadshow, and hired dozens of new employees.  Even before registration opened to the public in May, Manus had become a reference point for what a broad, consumer‑oriented AI agent should accomplish. Rather than handling narrow chores for businesses, this “general” agent is designed to be able to help with everyday tasks like trip planning, stock comparison, or your kid’s school project.  Unlike previous AI agents, Manus uses a browser-based sandbox that lets users supervise the agent like an intern, watching in real time as it scrolls through web pages, reads articles, or codes actions. It also proactively asks clarifying questions, supports long-term memory that would serve as context for future tasks. “Manus represents a promising product experience for AI agents,” says Ang Li, cofounder and CEO of Simular, a startup based in Palo Alto, California, that’s building computer use agents, AI agents that control a virtual computer. “I believe Chinese startups have a huge advantage when it comes to designing consumer products, thanks to cutthroat domestic competition that leads to fast execution and greater attention to product details.” In the case of Manus, the competition is moving fast. Two of the most buzzy follow‑ups, Genspark and Flowith, for example, are already boasting benchmark scores that match or edge past Manus’s.  Genspark, led by former Baidu executives Eric Jing and Kay Zhu, links many small “super agents” through what it calls multi‑component prompting. The agent can switch among several large language models, accepts both images and text, and carries out tasks from making slide decks to placing phone calls. Whereas Manus relies heavily on Browser Use, a popular open-source product that lets agents operate a web browser in a virtual window like a human, Genspark directly integrates with a wide array of tools and APIs. Launched in April, the company says that it already has over 5 million users and over $36 million in yearly revenue. Flowith, the work of a young team that first grabbed public attention in April 2025 at a developer event hosted by the popular social media app Xiaohongshu, takes a different tack. Marketed as an “infinite agent,” it opens on a blank canvas where each question becomes a node on a branching map. Users can backtrack, take new branches, and store results in personal or sharable “knowledge gardens”—a design that feels more like project management software (think Notion) than a typical chat interface. Every inquiry or task builds its own mind-map-like graph, encouraging a more nonlinear and creative interaction with AI. Flowith’s core agent, NEO, runs in the cloud and can perform scheduled tasks like sending emails and compiling files. The founders want the app to be a “knowledge marketbase”, and aims to tap into the social aspect of AI with the aspiration of becoming “the OnlyFans of AI knowledge creators”. What they also share with Manus is the global ambition. Both Genspark and Flowith have stated that their primary focus is the international market. A global address Startups like Manus, Genspark, and Flowith—though founded by Chinese entrepreneurs—could blend seamlessly into the global tech scene and compete effectively abroad. Founders, investors, and analysts that MIT Technology Review has spoken to believe Chinese companies are moving fast, executing well, and quickly coming up with new products.  Money reinforces the pull to launch overseas. Customers there pay more, and there are plenty to go around. “You can price in USD, and with the exchange rate that’s a sevenfold multiplier,” Manus cofounder Xiao Hong quipped on a podcast. “Even if we’re only operating at 10% power because of cultural differences overseas, we’ll still make more than in China.” But creating the same functionality in China is a challenge. Major US AI companies including OpenAI and Anthropic have opted out of mainland China because of geopolitical risks and challenges with regulatory compliance. Their absence initially created a black market as users resorted to VPNs and third-party mirrors to access tools like ChatGPT and Claude. That vacuum has since been filled by a new wave of Chinese chatbots—DeepSeek, Doubao, Kimi—but the appetite for foreign models hasn’t gone away.  Manus, for example, uses Anthropic’s Claude Sonnet—widely considered the top model for agentic tasks. Manus cofounder Zhang Tao has repeatedly praised Claude’s ability to juggle tools, remember contexts, and hold multi‑round conversations—all crucial for turning chatty software into an effective executive assistant. But the company’s use of Sonnet has made its agent functionally unusable inside China without a VPN. If you open Manus from a mainland IP address, you’ll see a notice explaining that the team is “working on integrating Qwen’s model,” a special local version that is built on top of Alibaba’s open-source model.  An engineer overseeing ByteDance’s work on developing an agent, who spoke to MIT Technology Review anonymously to avoid sanction, said that the absence of Claude Sonnet models “limits everything we do in China.” DeepSeek’s open models, he added, still hallucinate too often and lack training on real‑world workflows. Developers we spoke with rank Alibaba’s Qwen series as the best domestic alternative, yet most say that switching to Qwen knocks performance down a notch. Jiaxin Pei, a postdoctoral researcher at Stanford’s Institute for Human‑Centered AI, thinks that gap will close: “Building agentic capabilities in base LLMs has become a key focus for many LLM builders, and once people realize the value of this, it will only be a matter of time.” For now, Manus is doubling down on audiences it can already serve. In a written response, the company said its “primary focus is overseas expansion,” noting that new offices in San Francisco, Singapore, and Tokyo have opened in the past month. A super‑app approach Although the concept of AI agents is still relatively new, the consumer-facing AI app market in China is already crowded with major tech players. DeepSeek remains the most widely used, while ByteDance’s Doubao and Moonshot’s Kimi have also become household names. However, most of these apps are still optimized for chat and entertainment rather than task execution. This gap in the local market has pushed China’s big tech firms to roll out their own user-facing agents, though early versions remain uneven in quality and rough around the edges.  ByteDance is testing Coze Space, an AI agent based on its own Doubao model family that lets users toggle between “plan” and “execute” modes, so they can either directly guide the agent’s actions or step back and watch it work autonomously. It connects up to 14 popular apps, including GitHub, Notion, and the company’s own Lark office suite. Early reviews say the tool can feel clunky and has a high failure rate, but it clearly aims to match what Manus offers. Meanwhile, Zhipu AI has released a free agent called AutoGLM Rumination, built on its proprietary ChatGLM models. Shanghai‑based Minimax has launched Minimax Agent. Both products look almost identical to Manus and demo basic tasks such as building a simple website, planning a trip, making a small Flash game, or running quick data analysis. Despite the limited usability of most general AI agents launched within China, big companies have plans to change that. During a May 15 earnings call, Tencent president Liu Zhiping teased an agent that would weave automation directly into China’s most ubiquitous app, WeChat.  Considered the original super-app, WeChat already handles messaging, mobile payments, news, and millions of mini‑programs that act like embedded apps. These programs give Tencent, its developer, access to data from millions of services that pervade everyday life in China, an advantage most competitors can only envy. Historically, China’s consumer internet has splintered into competing walled gardens—share a Taobao link in WeChat and it resolves as plaintext, not a preview card. Unlike the more interoperable Western internet, China’s tech giants have long resisted integration with one another, choosing to wage platform war at the expense of a seamless user experience. But the use of mini‑programs has given WeChat unprecedented reach across services that once resisted interoperability, from gym bookings to grocery orders. An agent able to roam that ecosystem could bypass the integration headaches dogging independent startups. Alibaba, the e-commerce giant behind the Qwen model series, has been a front-runner in China’s AI race but has been slower to release consumer-facing products. Even though Qwen was the most downloaded open-source model on Hugging Face in 2024, it didn’t power a dedicated chatbot app until early 2025. In March, Alibaba rebranded its cloud storage and search app Quark into an all-in-one AI search tool. By June, Quark had introduced DeepResearch—a new mode that marks its most agent-like effort to date.  ByteDance and Alibaba did not reply to MIT Technology Review’s request for comments. “Historically, Chinese tech products tend to pursue the all-in-one, super-app approach, and the latest Chinese AI agents reflect just that,” says Li of Simular, who previously worked at Google DeepMind on AI-enabled work automation. “In contrast, AI agents in the US are more focused on serving specific verticals.” Pei, the researcher at Stanford, says that existing tech giants could have a huge advantage in bringing the vision of general AI agents to life—especially those with built-in integration across services. “The customer-facing AI agent market is still very early, with tons of problems like authentication and liability,” he says. “But companies that already operate across a wide range of services have a natural advantage in deploying agents at scale.”

The wait is almost over – Nintendo’s Switch 2 officially ships starting today, and the gaming community is electrified with anticipation. With its announcement sending shockwaves through the industry, the Switch 2 promises upgraded hardware, enhanced graphics, and a fresh wave of exclusive titles that have fans and newcomers alike counting down the hours. Social media and gaming forums are buzzing with unboxing plans, early impressions from pre-release units, and speculation about which features will define this next generation of hybrid gameplay. As any veteran Switch owner knows, the right accessories can dramatically enhance the gaming experience, and the Switch 2 is no exception. Whether it’s for extended play sessions at home, on-the-go protection, or boosting multiplayer fun, accessories are the unsung heroes that maximize the console’s potential. From ergonomic thumb grips and bumper cases to game-cartridge-holders, a well-equipped setup can make all the difference – and JSAUX seems to have everything sorted. Designer: JSAUX 1. JSAUX Carrying Case for Nintendo Switch 2 No one wants to risk their brand-new Switch 2 rattling around in a flimsy pouch. JSAUX’s Carrying Case ditches the generic for a custom-molded fit, wrapping your console in thick, dual-layer EVA hard-shell that shrugs off drops and shields against everyday scuffs. The plush microfiber lining is a small luxury, keeping your screen pristine even after you inevitably toss the case into a crowded backpack. There’s room for everything you actually need: 15 game card slots, a mesh pocket for cables and earbuds, and enough space for your Switch 2 with its own protective case attached. JSAUX nails the details with a smooth dual zipper, a Velcro strap to keep your console locked in place, and an adjustable stand that turns any table into a mobile gaming station. Whether you’re heading out for a weekend or just want to keep your gear tidy at home, the comfortable handle and detachable shoulder strap make carrying a breeze. It’s protection, organization, and a flex of practicality, without the bulk. Why We Recommend It This is the Switch 2 case for real-world gamers – the ones who actually take their console outside, swap cartridges in weird places, and need gear that isn’t just stylish, but built for chaos. The customizable storage and soft-but-tough protection mean your Switch 2 and accessories survive every bump, drop, or accidental coffee spill. It’s the kind of case that lets you focus on the game, not what’s happening to your hardware. Click Here to Buy Now: $14.39 $21.99 (34% off, use coupon code “JSSWITCH2”). This code can be stacked with existing product deals or discounts. Hurry, deal ends in 48-hours! 2. JSAUX 3-Pack Tempered Glass Screen Protector for Nintendo Switch 2 You know the moment you peel the plastic off a brand-new console? That pristine look deserves to last. JSAUX’s Tempered Glass Screen Protector for Switch 2 nails that first-day shine with edge-to-edge coverage, engineered specifically for Nintendo’s 2025 flagship. The guide frame makes installation foolproof – seriously, it’s almost impossible to misalign or trap bubbles, even if you’re all thumbs and shaky from launch-day excitement. Once on, the 99% transparency keeps every pixel crisp and colors vibrant, so Breath of the Wild 2 or Mario Kart pop just like on day one. Built from dual-tempered 9H glass, it shrugs off scratches, rogue keys, and the random chaos of backpack travel. The 0.3mm thickness is as subtle as it gets – no loss of touch accuracy, no weird drag. And thanks to a nano-oleophobic coating, smudges wipe away quickly, so you’re never gaming through a fingerprint haze. Whether you’re marathon grinding or passing the Switch to a pizza-fingered friend, this protector is ready for real-world gaming. Why We Recommend It Forget triple-checking your bag for sharp objects – this JSAUX screen protector means your Switch 2 just isn’t precious anymore, it’s practical. Installation is stress-free, the display stays flawless, and your frantic touchscreen mashing never feels laggy. For anyone tired of treating their console like fine china, this is the protection that lets you just play. Click Here to Buy Now: $5.59 $9.99 (44% off, use coupon code “JSSWITCH2”). This code can be stacked with existing product deals or discounts. Hurry, deal ends in 48 hours! 3. JSAUX Protective Case Compatible with Nintendo Switch 2 Skip the bulky bricks—this JSAUX Protective Carry Case is precision-engineered for the Switch 2 and the Switch 2 only. You get a snug, movement-free fit that feels purpose-built because it is, and it won’t work with any extra shells or cases. The hard PC exterior shrugs off scratches, dust, and splashes, holding up under café tables and cross-country flights alike. Inside, soft silicone strips keep your Switch 2 floating safely, while the clever raised sections mean your joysticks never take the brunt of an accidental bump. Traveling light? The detachable wrist strap is a small convenience that makes a big difference, and the slim profile means this case slides into any backpack or shoulder bag without hogging space. Built-in game card slots are a thoughtful touch – no more loose cartridges rattling around or getting lost in the depths of your bag. It’s the kind of accessory that makes daily Switch 2 use seamless, keeping your console clean, protected, and always ready to play. Why We Recommend It No-nonsense protection, zero wasted space. This case is for people who toss their Switch 2 into a bag and actually live life. The fit is tight, the shell is tough, and the joystick protection is smarter than most cases twice the price. Game card slots mean your library travels with you, not in a Ziploc. It’s the everyday armor you don’t have to think about, but you’ll notice the moment you need it. Click Here to Buy Now: $13.59 $25.99 (47% off, use coupon code “JSSWITCH2”). This code can be stacked with existing product deals or discounts. Hurry, deal ends in 48-hours! 4. JSAUX Thumb Grips for Nintendo Switch 2 If you’ve ever wished for a thumbstick upgrade on the Switch 2, these JSAUX Thumb Grips are as close as it gets to pro-level control, without needing a soldering iron or a warranty-voiding moment. Designed specifically for the Nintendo Switch 2, they snap on with a reassuring snugness and stay put through frantic boss fights or marathon Splatoon sessions. Three sizes (1, 2, and 3) are included in the kit, so you can fine-tune each stick for your hand size and play style – no more settling for “one size fits all” discomfort. Function isn’t sacrificed for comfort, either. The soft-touch finish and sculpted ergonomic shape actually take the edge off finger fatigue, so you can grind through RPGs or rack up wins in fighters without your thumb locking up. Each cap height serves a purpose: short for fast, twitchy moves in action games, tall for accuracy and steady aim in shooters, original size to simply boost grip. Mix and match to build your own hybrid layout. Why We Recommend It Who knew a 10-second upgrade could totally shift your game? These JSAUX Thumb Grips for Switch 2 let you dial in comfort and precision, whether you’re craving snap reflexes in Smash or pixel-perfect aim in Splatoon. No slipping, no awkward fit, no thumb fatigue after hours – just that satisfying soft grip and a custom feel that actually matches your play style. For anyone tired of stock sticks, this is the tweak you’ll wish you’d made sooner. Click Here to Buy Now: $6.39 $12.99 (50% off, use coupon code “JSSWITCH2”). This code can be stacked with existing product deals or discounts. Hurry, deal ends in 48-hours! 5. JSAUX Dockable Case for Nintendo Switch 2 The JSAUX Protective Case for Nintendo Switch 2 is engineered for gamers who want real protection without sacrificing style or convenience. Its split design means you can effortlessly detach the JoyCons without wrestling with stubborn plastic, and the ultra-thin 1mm PC back shell slides right into the official dock – no need to remove the case every time you charge or play on the big screen. The U-shaped cutout perfectly frames the Switch 2’s built-in kickstand, so you can go from handheld to tabletop mode in seconds, without ever fumbling or forcing a fit. Hybrid construction is the secret sauce here. The JoyCon covers fuse a transparent, scratch-resistant PC front with a soft-touch TPU back, letting the Switch 2’s design shine while adding grip and comfort for marathon sessions. Every port and button stays fully accessible thanks to precise cutouts, and popping the covers on or off is a breeze. No more risking cracked clips or flaky plastic – just smooth, reliable protection dialed in for everyday use. Why We Recommend It Forget clunky cases that ruin the Switch’s sleek vibe or force you to undock every time you want to play on your TV. JSAUX nails the essentials: dock-friendly fit, proper grip, zero interference with the kickstand, and a clear finish that doesn’t hide your console’s good looks. It’s protection that respects the Switch 2’s design – no bulk, no hassle, just smart, seamless defense you barely notice until you need it. Click Here to Buy Now: $12.99 $19.99 (35% off, use coupon code “JSSWITCH2”). This code can be stacked with existing product deals or discounts. Hurry, deal ends in 48-hours!The post Don’t Unbox Your Nintendo Switch 2 Without These Gaming Accessories first appeared on Yanko Design.

Looking to learn a new language or refresh your skills before traveling this summer? Here are the best language learning apps of 2025.