Where others have become obsessed with speed and trends, Sergio Membrillas has built a career on slowing down and staying true to his craft. Over 13 years as a professional illustrator, he's evolved a distinct style – bold, graphic, and quietly playful – that feels both timeless and entirely his own. But this evolution, as he's quick to point out, wasn't something he forced. "I believe evolution is essential for every artist," Sergio reflects. "Change is not something I resist; it's something I enjoy and welcome as part of the creative journey." His work, initially influenced by a love of Mid-Century graphic design, has gradually absorbed references as diverse as Etruscan art, Mesopotamian motifs, and early 20th-century traditional tattoos. It's a portfolio built on curiosity, not conformity. Despite his stylistic clarity, Sergio's process remains refreshingly analogue at its core. He still begins each project with pencil and paper, leaning into the tactility of drawing by hand. "I've always valued having a physical relationship with art," he says. "There's something essential about feeling the wood of the pencil, manually erasing mistakes, rather than simply double-tapping a screen." His philosophy of working slowly, attentively, and with purpose shows in the balanced compositions and confident use of negative space that has become his signature. Sergio credits his background in graphic design for this instinctive sense of structure. "There has always been a strong graphic sensibility within me," he says, noting that the discipline of design continues to inform his illustration practice, particularly in editorial and poster work where clarity and storytelling must co-exist. Yet, like many creative careers, his path into illustration wasn't plotted from the start. Initially, Sergio imagined illustration would complement his design projects, not become the main event. "At first, I thought I would incorporate my illustrations into my graphic design projects – but in the end, it turned out to be the other way around." A commission for EasyJet's in-flight magazine marked a pivotal moment that reframed illustration from a hobby to a profession. "It made me realize that being an illustrator is a real profession, just like any other." Fast-forward to today, and Sergio's client list reads like a who's who of publishing and design, from The New Yorker to Pentagram. While the calibre of collaborators has changed, what matters most remains the same: trust. "When a client trusts your work and approaches projects with an open mind, collaboration flourishes," he says. It's in these open, trusting relationships that Sergio finds the space to experiment and further develop his ideas. Unsurprisingly, editorial work holds a special place for him. "Editorial projects offer wider creative margins, allowing for greater flexibility and innovation," he explains. Compared to the tighter confines of advertising, editorial illustration offers the opportunity to tell nuanced stories, and Sergio is clear that he views every assignment, regardless of format, as a collaborative effort. If one thing is clear, though, it's that he's not that he's afraid to say no when needed. Maintaining a strong personal voice is non-negotiable. "I always strive to stay true to myself and ensure that every project I undertake reflects a clear personal signature," he says. For Sergio, authenticity isn't just a matter of artistic pride; it's what differentiates an illustrator in a saturated, increasingly automated industry. Valencia is home for Sergio, a city he credits with much of his creative energy. "Oh, Valencia! It's a beautiful city that inspires me and brings me joy," he says, describing it as a European cousin to Los Angeles with its sun-soaked streets and vibrant cultural scene. The blend of tradition and modernity fuels his practice, offering constant inspiration without the sensory overload that often accompanies larger creative hubs. Given the pressure many creatives feel to keep pace with shifting trends, Sergio's approach feels almost radical. "I'm not particularly interested in trends," he says. "What truly matters to me is the feeling of having done something meaningful and fulfilling by the end of the day." Instead of chasing what's fashionable, he draws inspiration from a surprisingly eclectic bookshelf, mixing everything from Wim Crouwel's graphic design classics to studies on Scandinavian tattoos and Alec Soth's photography. It's a reminder that fresh ideas rarely come from looking where everyone else is looking. In the era of AI and content overload, the role of the illustrator is changing, but Sergio remains optimistic. While machine learning might churn out images at record speed, it can't replicate the nuance and emotional intelligence that underpin great illustration. "Illustrators can tell stories, evoke emotions, and create meaningful connections that machines simply cannot replicate," he says. "Our role is shifting, but the value of authentic, thoughtful illustration remains indispensable." That insistence on authenticity carries through to the advice he offers younger illustrators navigating a commercial landscape. "Create work that makes you happy when you go to sleep at night," Sergio advises. "It's important to find projects that align with your values and passions so your artistic voice remains authentic." It's an ethos that has carried him through more than a decade of creative highs and industry shifts. Perhaps it's the real secret behind the clarity of his practice, which he has built not on chasing trends or algorithms but on careful craft, deliberate evolution, and the simple, enduring joy of a well-made pencil and a blank piece of paper.

Email authentication protocols like SPF, DKIM, and DMARC play a critical role in protecting domains from spoofing and phishing. However, when SEGs are introduced into the email path, the interaction with these protocols becomes more complex. Security gateways(SEGs) are a core part of many organizations’ email infrastructure. They act as intermediaries between the public internet and internal mail systems, inspecting, filtering, and routing messages. This blog examines how security gateways handle SPF, DKIM, and DMARC, with real-world examples from popular gateways such as Proofpoint, Mimecast, and Avanan. We’ll also cover best practices for maintaining authentication integrity and avoiding misconfigurations that can compromise email authentication or lead to false DMARC failures. Security gateways often sit at the boundary between your organization and the internet, managing both inbound and outbound email traffic. Their role affects how email authentication protocols behave. An inbound SEG examines emails coming into your organization. It checks SPF, DKIM, and DMARC to determine if the message is authentic and safe before passing it to your internal mail servers. An outbound SEG handles emails sent from your domain. It may modify headers, rewrite envelope addresses, or even apply DKIM signing. All of these can impact SPF,  DKIM, or DMARC validation on the recipient’s side. Understanding how SEGs influence these flows is crucial to maintaining proper authentication and avoiding unexpected DMARC failures. Inbound Handling of SPF, DKIM, and DMARC by Common Security Gateways When an email comes into your organization, your security gateway is the first to inspect it. It checks whether the message is real, trustworthy, and properly authenticated. Let’s look at how different SEGs handle these checks. Avanan (by Check Point) SPF: Avanan verifies whether the sending server is authorized to send emails for the domain by checking the SPF record. DKIM: It verifies if the message was signed by the sending domain and if that signature is valid. DMARC: It uses the results of the SPF and DKIM check to evaluate DMARC. However, final enforcement usually depends on how DMARC is handled by Microsoft 365 or Gmail, as Avanan integrates directly with them. Avanan offers two methods of integration:1. API integration: Avanan connects via APIs, no change in MX, usually Monitor or Detect modes.2. Inline integration: Avanan is placed inline in the mail flow (MX records changed), actively blocking or remediating threats. Proofpoint Email Protection SPF: Proofpoint checks SPF to confirm the sender’s IP is authorized to send on behalf of the domain. You can set custom rules (e.g. treat “softfail” as “fail”). DKIM: It verifies DKIM signatures and shows clear pass/fail results in logs. DMARC: It fully evaluates DMARC by combining SPF and DKIM results with alignment checks. Administrators can configure how to handle messages that fail DMARC, such as rejecting, quarantining, or delivering them. Additionally, Proofpoint allows whitelisting specific senders you trust, even if their emails fail authentication checks. Integration Methods Inline Mode: In this traditional deployment, Proofpoint is positioned directly in the email flow by modifying MX records. Emails are routed through Proofpoint’s infrastructure, allowing it to inspect and filter messages before they reach the recipient’s inbox. This mode provides pre-delivery protection and is commonly used in on-premises or hybrid environments. API-Based (Integrated Cloud Email Security – ICES) Mode: Proofpoint offers API-based integration, particularly with cloud email platforms like Microsoft 365 and Google Workspace. In this mode, Proofpoint connects to the email platform via APIs, enabling it to monitor and remediate threats post-delivery without altering the email flow. This approach allows for rapid deployment and seamless integration with existing cloud email services. Mimecast SPF: Mimecast performs SPF checks to verify whether the sending server is authorized by the domain’s SPF record. Administrators can configure actions for SPF failures, including block, quarantine, permit, or tag with a warning. This gives flexibility in balancing security with business needs. DKIM: It validates DKIM signatures by checking that the message was correctly signed by the sending domain and that the content hasn’t been tampered with. If the signature fails, Mimecast can take actions based on your configured policies. DMARC: It fully evaluates DMARC by combining the results of SPF and DKIM with domain alignment checks. You can choose to honor the sending domain’s DMARC policy (none, quarantine, reject) or apply custom rules, for example, quarantining or tagging messages that fail DMARC regardless of the published policy. This allows more granular control for businesses that want to override external domain policies based on specific contexts. Integration Methods Inline Deployment: Mimecast is typically deployed as a cloud-based secure email gateway. Organizations update their domain’s MX records to point to Mimecast, so all inbound (and optionally outbound) emails pass through it first. This allows Mimecast to inspect, filter, and process emails before delivery, providing robust protection. API Integrations: Mimecast also offers API-based services through its Mimecast API platform, primarily for management, archival, continuity, and threat intelligence purposes. However, API-only email protection is not Mimecast’s core model. Instead, the APIs are used to enhance the inline deployment, not replace it. Barracuda Email Security Gateway SPF: Barracuda checks the sender’s IP against the domain’s published SPF record. If the check fails, you can configure the system to block, quarantine, tag, or allow the message, depending on your policy preferences. DKIM: It validates whether the incoming message includes a valid DKIM signature. The outcome is logged and used to inform further policy decisions or DMARC evaluations. DMARC: It combines SPF and DKIM results, checks for domain alignment, and applies the DMARC policy defined by the sender. Administrators can also choose to override the DMARC policy, allowing messages to pass or be treated differently based on organizational needs (e.g., trusted senders or internal exceptions). Integration Methods Inline mode (more common and straightforward): Barracuda Email Security Gateway is commonly deployed inline by updating your domain’s MX records to point to Barracuda’s cloud or on-premises gateway. This ensures that all inbound emails pass through Barracuda first for filtering and SPF, DKIM, and DMARC validation before being delivered to your mail servers. Deployment Behind the Corporate Firewall: Alternatively, Barracuda can be deployed in transparent or bridge mode without modifying MX records. In this setup, the gateway is placed inline at the network level, such as behind a firewall, and intercepts mail traffic transparently. This method is typically used in complex on-premises environments where changing DNS records is not feasible. Cisco Secure Email (formerly IronPort) Cisco Secure Email acts as an inline gateway for inbound email, usually requiring your domain’s MX records to point to the Cisco Email Security Appliance or cloud service. SPF: Cisco Secure Email verifies whether the sending server is authorized in the sender domain’s SPF record. Administrators can set detailed policies on how to handle SPF failures. DKIM: It validates the DKIM signature on incoming emails and logs whether the signature is valid or has failed. DMARC: It evaluates DMARC by combining SPF and DKIM results along with domain alignment checks. Admins can configure specific actions, such as quarantine, reject, or tag, based on different failure scenarios or trusted sender exceptions. Integration methods On-premises Email Security Appliance (ESA): You deploy Cisco’s hardware or virtual appliance inline, updating MX records to route mail through it for filtering. Cisco Cloud Email Security: Cisco offers a cloud-based email security service where MX records are pointed to Cisco’s cloud infrastructure, which filters and processes inbound mail. Cisco Secure Email also offers advanced, rule-based filtering capabilities and integrates with Cisco’s broader threat protection ecosystem, enabling comprehensive inbound email security. Outbound Handling of SPF, DKIM, and DMARC by Common Security Gateways When your organization sends emails, security gateways can play an active role in processing and authenticating those messages. Depending on the configuration, a gateway might rewrite headers, re-sign messages, or route them through different IPs – all actions that can help or hurt the authentication process. Let’s look at how major SEGs handle outbound email flow. Avanan – Outbound Handling and Integration Methods Outbound Logic Avanan analyzes outbound emails primarily to detect data loss, malware, and policy violations. In API-based integration, emails are sent directly by the original mail server (e.g., Microsoft 365 or Google Workspace), so SPF and DKIM signatures remain intact. Avanan does not alter the message or reroute traffic, which helps maintain full DMARC alignment and domain reputation. Integration Methods 1. API Integration: Connects to Microsoft 365 or Google Workspace via API. No MX changes are needed. Emails are scanned after they are sent, with no modification to SPF, DKIM, or the delivery path.  How it works: Microsoft Graph API or Google Workspace APIs are used to monitor and intervene in outbound emails. Protection level: Despite no MX changes, it can offer inline-like protection, meaning it can block, quarantine, or encrypt emails before they are delivered externally. SPF/DKIM/DMARC impact: Preserves original headers and signatures since mail is sent directly from Microsoft/Google servers. 2. Inline Integration: Requires changing MX records to route email through Avanan. In this mode, Avanan can intercept and inspect outbound emails before delivery. Depending on the configuration, this may affect SPF or DKIM if not properly handled. How it works: Requires adding Avanan’s Protection level: Traditional inline security with full visibility and control, including encryption, DLP, policy enforcement, and advanced threat protection. SPF/DKIM/DMARC impact: SPF configuration is needed by adding Avanan’s include mechanism to the sending domain’s SPF record. The DKIM record of the original sending source is preserved. For configurations, you can refer to the steps in this blog. Proofpoint – Outbound Handling and Integration Methods Outbound Logic Proofpoint analyzes outbound emails to detect and prevent data loss (DLP), to identify advanced threats (malware, phishing, BEC) originating from compromised internal accounts, and to ensure compliance. Their API integration provides crucial visibility and powerful remediation capabilities, while their traditional gateway (MX record) deployment delivers true inline, pre-delivery blocking for outbound traffic. Integration methods 1. API Integration: No MX record changes are required for this deployment method. Integration is done with Microsoft 365 or Google Workspace. How it works: Through its API integration, Proofpoint gains deep visibility into outbound emails and provides layered security and response features, including: Detect and alert: Identifies sensitive content (Data Loss Prevention violations), malicious attachments, or suspicious links in outbound emails. Post-delivery remediation (TRAP): A key capability of the API model is Threat Response Auto-Pull (TRAP), which enables Proofpoint to automatically recall, quarantine, or delete emails after delivery. This is particularly useful for internally sent messages or those forwarded to other users. Enhanced visibility: Aggregates message metadata and logs into Proofpoint’s threat intelligence platform, giving security teams a centralized view of outbound risks and user behavior. Protection level: API-based integration provides strong post-delivery detection and response, as well as visibility into DLP incidents and suspicious behavior.  SPF/DKIM/DMARC impact: Proofpoint does not alter SPF, DKIM, or DMARC because emails are sent directly through Microsoft or Google servers. Since Proofpoint’s servers are not involved in the actual sending process, the original authentication headers remain intact. 2. Gateway Integration (MX Record/Smart Host): This method requires updating MX records or routing outbound mail through Proofpoint via a smart host. How it works: Proofpoint acts as an inline gateway, inspecting emails before delivery. Inbound mail is filtered via MX changes; outbound mail is relayed through Proofpoint’s servers. Threat and DLP filtering: Scans outbound messages for sensitive content, malware, and policy violations. Real-time enforcement: Blocks, encrypts, or quarantines emails before they’re delivered. Policy controls: Applies rules based on content, recipient, or behavior. Protection level: Provides strong, real-time protection for outbound traffic with pre-delivery enforcement, DLP, and encryption. SPF/DKIM/DMARC impact: Proofpoint becomes the sending server: SPF: You need to configure ProofPoint’s SPF. DKIM: Can sign messages; requires DKIM setup. DMARC: DMARC passes if SPF and DKIM are set up properly. Please refer to this article to configure SPF and DKIM for ProofPoint. Mimecast – Outbound Handling and Integration Methods Outbound Logic Mimecast inspects outbound emails to prevent data loss (DLP), detect internal threats such as malware and impersonation, and ensure regulatory compliance. It primarily functions as a Secure Email Gateway (SEG), meaning it sits directly in the outbound email flow. While Mimecast offers APIs, its core outbound protection is built around this inline gateway model. Integration Methods 1. Gateway Integration (MX Record change required) This is Mimecast’s primary method for outbound email protection. Organizations route their outbound traffic through Mimecast by configuring their email server (e.g., Microsoft 365, Google Workspace, etc.) to use Mimecast as a smart host. This enables Mimecast to inspect and enforce policies on all outgoing emails in real time. How it works: Updating outbound routing in your email system (smart host settings), or Using Mimecast SMTP relay to direct messages through their infrastructure. Mimecast then scans, filters, and applies policies before the email reaches the final recipient. Protection level: Advanced DLP: Identifies and prevents sensitive data leaks. Impersonation and Threat Protection: Blocks malware, phishing, and abuse from compromised internal accounts. Email Encryption and Secure Messaging: Applies encryption policies or routes messages via secure portals. Regulatory Compliance: Enforces outbound compliance rules based on content, recipient, or metadata. SPF/DKIM/DMARC impact: SPF: Your SPF record must include Mimecast’s SPF mechanism based on your region to avoid SPF failures. DKIM: A new DKIM record should be configured to make sure your emails are DKIM signed when routing through Mimecast. DMARC: With correct SPF and DKIM setup, Mimecast ensures DMARC alignment, maintaining your domain’s sending reputation. Please refer to the steps in this detailed article to set up SPF and DKIM for Mimecast. 2. API Integration (Complementary to Gateway) Mimecast’s APIs complement the main gateway by providing automation, reporting, and management tools rather than handling live outbound mail flow. They allow you to manage policies, export logs, search archived emails, and sync users. APIs enhance visibility and operational tasks but do not provide real-time filtering or blocking of outbound messages. Since APIs don’t process live mail, they have no direct effect on SPF, DKIM, or DMARC; those depend on your gateway (smart host) setup. Barracuda – Outbound Handling and Integration Methods Outbound Logic Barracuda analyzes outbound emails to prevent data loss (DLP), block malware, stop phishing/impersonation attempts from compromised internal accounts, and ensure compliance. Barracuda offers flexible deployment options, including both traditional gateway (MX record) and API-based integrations. While both contribute to outbound security, their roles are distinct. Integration Methods 1. Gateway Integration (MX Record / Smart Host) — Primary Inline Security How it works: All outbound emails pass through Barracuda’s security stack for real-time inspection, threat blocking, and policy enforcement before delivery. Protection level: Comprehensive DLP (blocking, encrypting, or quarantining sensitive content)  Outbound spam and virus filtering  Enforcement of compliance and content policies This approach offers a high level of control and immediate threat mitigation on outbound mail flow. SPF/DKIM/DMARC impact: SPF: Update SPF records to include Barracuda’s sending IPs or SPF include mechanism. DKIM: Currently, no explicit setup is needed; DKIM of the main sending source is preserved. Refer to this article for more comprehensive guidance on Barracuda SEG configuration. 2. API Integration (Complementary & Advanced Threat Focus) How it works: The API accesses cloud email environments to analyze historical and real-time data, learning normal communication patterns to detect anomalies in outbound emails. It also supports post-delivery remediation, enabling the removal of malicious emails from internal mailboxes after sending. Protection level: Advanced AI-driven detection and near real-time blocking of outbound threats, plus strong post-delivery cleanup capabilities. SPF/DKIM/DMARC impact: Since mail is sent directly by the original mail server (e.g., Microsoft 365), SPF and DKIM signatures remain intact, preserving DMARC alignment and domain reputation. Cisco Secure Email (formerly IronPort) – Outbound Handling and Integration Methods Outbound Logic Cisco Secure Email protects outbound email by preventing data loss (DLP), blocking spam and malware from internal accounts, stopping business email compromise (BEC) and impersonation attacks, and ensuring compliance. Cisco provides both traditional gateway appliances/cloud gateways and modern API-based solutions for layered outbound security. Integration Methods 1. Gateway Integration (MX Record / Smart Host) – Cisco Secure Email Gateway (ESA) How it works: Organizations update MX records to route mail through the Cisco Secure Email Gateway or configure their mail server (e.g., Microsoft 365, Exchange) to smart host outbound email via the gateway. All outbound mail is inspected and policies enforced before delivery. Protection level: Granular DLP (blocking, encrypting, quarantining sensitive content) Outbound spam and malware filtering to protect IP reputation Email encryption for sensitive outbound messages Comprehensive content and attachment policy enforcement SPF: Check this article for comprehensive guidance on Cisco SPF settings. DKIM: Refer to this article for detailed guidance on Cisco DKIM settings. 2. API Integration – Cisco Secure Email Threat Defense How it works: Integrates directly via API with Microsoft 365 (and potentially Google Workspace), continuously monitoring email metadata, content, and user behavior across inbound, outbound, and internal messages. Leverages Cisco’s threat intelligence and AI to detect anomalous outbound activity linked to BEC, account takeover, and phishing. Post-Delivery Remediation: Automates the removal or quarantine of malicious or policy-violating emails from mailboxes even after sending. Protection level: Advanced, AI-driven detection of sophisticated outbound threats with real-time monitoring and automated remediation. Complements gateway filtering by adding cloud-native visibility and swift post-send action. SPF/DKIM/DMARC impact: Since emails are sent directly by the original mail server, SPF and DKIM signatures remain intact, preserving DMARC alignment and domain reputation. If you have any questions or need assistance, feel free to reach out to EasyDMARC technical support.

Upcoming (serious) Web performance boostBy: Adam Scott5 June 2025Progress ReportSometimes, just adding a compiler flag can yield significant performance boosts. And that just happened.For about two years now, all major browsers have supported WASM (WebAssembly) SIMD. SIMD stands for “Single instruction, multiple data” and is a technology that permits CPUs to do some parallel computation, often speeding up the whole program. And that’s exactly why we tried it out recently.We got positive results.The need for performance on the WebThe Web platform is often overlooked as a viable target, because of its less-than-ideal environment and its perceived poor performance. And the perception is somewhat right: the Web environment has a lot of security-related quirks to take into account—the user needs to interact with a game frame before the browser allows it to play any sound1. Also, due to bandwidth and compatibility reasons, you rarely see high-fidelity games being played on a browser. Performance is better achieved when running software natively on the operating system.But don’t underestimate the potential of the Web platform. As I explained in broad terms at the talk I gave at the last GodotCon Boston 2025, the Web has caught up a lot since the days of Flash games. Not only are there more people playing Web games every year, but standards and browsers improve every year in functionality and in performance.And that’s why we are interested in using WASM SIMD.WASM SIMD BenchmarksOur resident benchmark expert Hugo Locurcio (better known as Calinou) ran the numbers for us on a stress test I made. We wanted to compare standard builds to builds with WASM SIMD enabled.Note: You may try to replicate his results, but be aware that he has a beast of a machine. Here are his PC’s specifications:CPU: Intel Core i9-13900KGPU: NVIDIA GeForce RTX 4090RAM: 64 GB (2×32 GB DDR5-5800 CL30)SSD: Solidigm P44 Pro 2 TBOS: Linux (Fedora 42)I built a Jolt physics stress test from a scene initially made by passivestar. By spawning more and more barrels into the contraption, we can easily test the performance difference between the WASM SIMD build and the other.Without WASM SIMDWith WASM SIMDImprovement (approx.)Test linksLinkLink-Firefox 138(“+100 barrels” 3 times)2×Firefox 138(“+100 barrels” 6 times)10.17×*Chromium 134(“+100 barrels” 3 times)1.37×Chromium 134(“+100 barrels” 6 times)14.17×**Please note that once the physics engine enters a “spiral of death”, it is common for the framerate to drop to single digits, SIMD or not. These tests don’t prove 10× to 15× CPU computing speed improvements, but rather that games will be more resilient to framerate drops on the same machine in the same circumstances. The 1.5× to 2× numbers are more representative here of the performance gains by WASM SIMD.What it means for your gamesStarting with 4.5 dev 5, you can expect your Web games to run a little bit more smoothly, without having to do anything. Especially when things get chaotic (for your CPU). It isn’t a silver bullet for poorly optimized games, but it will help nonetheless. Also, note that it cannot do anything for GPU rendering bottlenecks.Be aware that the stress tests are meant by nature to only test the worst case scenarios, so you may not see such large improvements in normal circumstances. But it’s nice to see such stark improvements when the worst happens.AvailabilityFrom here on out, the 4.5 release official templates will only support WebAssembly SIMD-compatible browsers in order to keep the template sizes small. We generally aim to maintain compatibility with the oldest devices we can. But in this case, the performance gains are too large to ignore and the chances of users having browsers that are that far out of date is too small relative to the potential benefits.If you need to use non-SIMD templates, don’t fret. You can always build the Godot Editor and the engine templates without WebAssembly SIMD support by using the wasm_simd=no build option.What’s next?As I wrote in my last blog post, we’re currently working very hard to make C#/.NET exports a reality. We do have a promising prototype, we just need to make sure that it’s production-ready.I also mentioned in that article that I wanted to concentrate on improving our asset loading game. Preloading an entire game before even starting it hinders the ability to use Godot for commercial Web games. Once something is implemented to improve that issue, count on me to share the news with you.It’s either that, or we return to the old days of spam-webpages using the “Congratulations, you won!” sound effect when you least expect it. ↩

Nailing cross-channel marketing for your app is crucial in today’s ecosystem. This isn’t just a speculation, it’s a hard fact: the average mobile marketer uses 5 or fewer media sources, but CPIs tend to be much higher for marketers using 6 or more media sources, according to a report by Singular and ironSource Aura.It’s time to start looking elsewhere for users. Something to consider is returning to old-school, tried-and-true strategies that have taken a backseat as the industry has advanced, such as web acquisition campaigns. There are two clicks in these campaigns - the first click directs users from an ad to a web landing page and the second click directs users from the web page to the app store. While the funnel to get to your app is longer, it’s actually an opportunity. Let’s dive into why and how you should be running web-to-app acquisition campaigns.3 benefits of web-to-app campaignsWeb acquisition campaigns became tried-and-true for a reason and many apps are beginning to take advantage of the strategy to reduce costs, get better visibility, and expand their reach.Reduce costsWhile the cost of app acquisition campaigns is increasing as the space becomes more saturated, web campaigns have been underutilized in recent years. This means many advertisers are finding that the cost of web campaigns are less than app store acquisitions. Much of these higher costs comes from high app store fees - Apple takes a 30% cut from large developers and a 15% cut from smaller developers. The web remains a relatively low-cost, accessible channel.Get better visibilityWith the ability to direct users to your owned mobile site before they click to download your app, you get better visibility into new users and where they came from - the app store is a black box where it’s nearly impossible to match the user who clicked an ad to the user who downloaded your app. On top of that, longer funnels give you more control and insight into each step users take to get to your app - allowing you to better optimize your campaigns. For example, you can try out different web landing pages to find the most effective way to increase users’ understanding of your app before they download it. This increases your likelihood of, first, motivating users to download your app, and second, encouraging them to stick around.Expand your reachInternet users worldwide spent 415.5 minutes a day online in 2021 according to Oberlo. App users spend at most 300 mins a day in apps according to TechCrunch. The internet is still more prevalent than apps because, let’s face it, the web predates the app store. Even more so, users have become especially savvy on the web, visiting the search bar to research a brand, fact check a friend, or dive deeper into a news headline. This means that by running web campaigns, you’re reaching a massive audience - nearly the entire world population.Andre Kempe, Founder and CEO of Admiral Media says, “marketers consider web acquisition important because they’re reaching a different type of audience that may convert to becoming an active app user, even though they were surfing on a news website… When you start advertising for the web, you reach an audience you haven’t reached before. This is definitely a growth opportunity." Listen to the full podcast.Web campaigns are a tried and true strategy, which is what makes them so effective in today’s ecosystem. Realizing that you should be considering this acquisition channel, let’s discuss how web onboarding can help you maximize success.Maximize success with web onboardingTypically, users complete onboarding directly on your app. Web onboarding, however, means letting users register, subscribe, and sometimes pay on the web, before heading to your app. This way, users are fully prepared to start using your app before even clicking download, ensuring they can engage at full capacity as soon as download is complete.When users onboard through the app, they may not immediately decide to engage but at least the app is still on their phones reminding them everyday of its value - they’re likely to open it again at some point in time. When a user onboards on the web and decides your app isn’t valuable, they’ll never end up actually downloading it - it’s gone and forgotten before it even started. That said, with web onboarding, the users who end up downloading your app will do so with high-intent to actively engage. Plus, you don’t have to pay for users that download without ever engaging. Whatever category your app is - fitness, food delivery, dating, reading, learning, etc. - you can use the web touchpoint to drive quality users to your app.Ensure the transition is seamlessFirst, ensure the web experience mimics the app experience enough that users don’t feel jipped when they actually start using your app. The transition from the web to the app should be seamless. If you’re offering users a taste of your app, the web touchpoint should look exactly like your app or, at least, include all of the same features. If the buttons to adjust settings are on the right side of the page during onboarding, they should be on the right side in the app as well. To ensure users stick around, you have to give them what they signed up for.Give users a sneak peakYou should also determine your main selling point and give users a sneak peak. For example, if you’re a reading app that allows users to adjust the settings of the page, you could give users premium access to the first chapter of a book allowing them to adjust font size or brightness. Readers will not want to leave the experience behind, encouraging them to set up their subscription and input their credit card information on the web before heading to the app to continue reading the book. The best part - when they open your app for the first time they’ll be able to jump right into the app experience.Learn more about how to improve your app’s onboarding experienceWeb-to-app campaigns should not be ignored in this competitive landscape - lower costs, more visibility, better optimization, and strong reach. Going even further to onboard users through the web, you’re improving the quality of the users entering your app for the first time.

One of the key use cases for generative AI involves answering questions over private datasets, with retrieval-augmented generation (RAG) as the go-to framework. As new RAG techniques emerge, there’s a growing need to benchmark their performance across diverse datasets and metrics.  To meet this need, we’re introducing BenchmarkQED, a new suite of tools that automates RAG benchmarking at scale, available on GitHub (opens in new tab). It includes components for query generation, evaluation, and dataset preparation, each designed to support rigorous, reproducible testing.   BenchmarkQED complements the RAG methods in our open-source GraphRAG library, enabling users to run a GraphRAG-style evaluation across models, metrics, and datasets. GraphRAG uses a large language model (LLM) to generate and summarize entity-based knowledge graphs, producing more comprehensive and diverse answers than standard RAG for large-scale tasks.  In this post, we walk through the core components of BenchmarkQED that contribute to the overall benchmarking process. We also share some of the latest benchmark results comparing our LazyGraphRAG system to competing methods, including a vector-based RAG with a 1M-token context window, where the leading LazyGraphRAG configuration showed significant win rates across all combinations of quality metrics and query classes. In the paper, we distinguish between local queries, where answers are found in a small number of text regions, and sometimes even a single region, and global queries, which require reasoning over large portions of or even the entire dataset.  Conventional vector-based RAG excels at local queries because the regions containing the answer to the query resemble the query itself and can be retrieved as the nearest neighbor in the vector space of text embeddings. However, it struggles with global questions, such as, “What are the main themes of the dataset?” which require understanding dataset qualities not explicitly stated in the text.   AutoQ: Automated query synthesis This limitation motivated the development of GraphRAG a system designed to answer global queries. GraphRAG’s evaluation requirements subsequently led to the creation of AutoQ, a method for synthesizing these global queries for any dataset. AutoQ extends this approach by generating synthetic queries across the spectrum of queries, from local to global. It defines four distinct classes based on the source and scope of the query (Figure 1, top) forming a logical progression along the spectrum (Figure 1, bottom). Figure 1. Construction of a 2×2 design space for synthetic query generation with AutoQ, showing how the four resulting query classes map onto the local-global query spectrum.  AutoQ can be configured to generate any number and distribution of synthetic queries along these classes, enabling consistent benchmarking across datasets without requiring user customization. Figure 2 shows the synthesis process and sample queries from each class, using an AP News dataset. Figure 2. Synthesis process and example query for each of the four AutoQ query classes.  About Microsoft Research Advancing science and technology to benefit humanity View our story Opens in a new tab AutoE: Automated evaluation framework  Our evaluation of GraphRAG focused on analyzing key qualities of answers to global questions. The following qualities were used for the current evaluation: Comprehensiveness: Does the answer address all relevant aspects of the question?  Diversity: Does it present varied perspectives or insights?  Empowerment: Does it help the reader understand and make informed judgments?  Relevance: Does it address what the question is specifically asking?   The AutoE component scales evaluation of these qualities using the LLM-as-a-Judge method. It presents pairs of answers to an LLM, along with the query and target metric, in counterbalanced order. The model determines whether the first answer wins, loses, or ties with the second. Over a set of queries, whether from AutoQ or elsewhere, this produces win rates between competing methods. When ground truth is available, AutoE can also score answers on correctness, completeness, and related metrics. An illustrative evaluation is shown in Figure 3. Using a dataset of 1,397 AP News articles on health and healthcare, AutoQ generated 50 queries per class (200 total). AutoE then compared LazyGraphRAG to a competing RAG method, running six trials per query across four metrics, using GPT-4.1 as a judge. These trial-level results were aggregated using metric-based win rates, where each trial is scored 1 for a win, 0.5 for a tie, and 0 for a loss, and then averaged to calculate the overall win rate for each RAG method. Figure 3. Win rates of four LazyGraphRAG (LGR) configurations across methods, broken down by the AutoQ query class and averaged across AutoE’s four metrics: comprehensiveness, diversity, empowerment, and relevance. LazyGraphRAG outperforms comparison conditions where the bar is above 50%. The four LazyGraphRAG conditions (LGR_b200_c200, LGR_b50_c200, LGR_b50_c600, LGR_b200_c200_mini) differ by query budget (b50, b200) and chunk size (c200, c600). All used GPT-4o mini for relevance tests and GPT-4o for query expansion (to five subqueries) and answer generation, except for LGR_b200_c200_mini, which used GPT-4o mini throughout. Comparison systems were GraphRAG (Local, Global, and Drift Search), Vector RAG with 8k- and 120k-token windows, and three published methods: LightRAG (opens in new tab), RAPTOR (opens in new tab), and TREX (opens in new tab). All methods were limited to the same 8k tokens for answer generation. GraphRAG Global Search used level 2 of the community hierarchy. LazyGraphRAG outperformed every comparison condition using the same generative model (GPT-4o), winning all 96 comparisons, with all but one reaching statistical significance. The best overall performance came from the larger budget, smaller chunk size configuration (LGR_b200_c200). For DataLocal queries, the smaller budget (LGR_b50_c200) performed slightly better, likely because fewer chunks were relevant. For ActivityLocal queries, the larger chunk size (LGR_b50_c600) had a slight edge, likely because longer chunks provide a more coherent context. Competing methods performed relatively better on the query classes for which they were designed: GraphRAG Global for global queries, Vector RAG for local queries, and GraphRAG Drift Search, which combines both strategies, posed the strongest challenge overall. Increasing Vector RAG’s context window from 8k to 120k tokens did not improve its performance compared to LazyGraphRAG. This raised the question of how LazyGraphRAG would perform against Vector RAG with 1-million token context window containing most of the dataset. Figure 4 shows the follow-up experiment comparing LazyGraphRAG to Vector RAG using GPT-4.1 that enabled this comparison. Even against the 1M-token window, LazyGraphRAG achieved higher win rates across all comparisons, failing to reach significance only for the relevance of answers to DataLocal queries. These queries tend to benefit most from Vector RAG’s ranking of directly relevant chunks, making it hard for LazyGraphRAG to generate answers that have greater relevance to the query, even though these answers may be dramatically more comprehensive, diverse, and empowering overall. Figure 4. Win rates of LazyGraphRAG (LGR) over Vector RAG across different context window sizes, broken down by the four AutoQ query classes and four AutoE metrics: comprehensiveness, diversity, empowerment, and relevance. Bars above 50% indicate that LazyGraphRAG outperformed the comparison condition.  AutoD: Automated data sampling and summarization Text datasets have an underlying topical structure, but the depth, breadth, and connectivity of that structure can vary widely. This variability makes it difficult to evaluate RAG systems consistently, as results may reflect the idiosyncrasies of the dataset rather than the system’s general capabilities. The AutoD component addresses this by sampling datasets to meet a target specification, defined by the number of topic clusters (breadth) and the number of samples per cluster (depth). This creates consistency across datasets, enabling more meaningful comparisons, as structurally aligned datasets lead to comparable AutoQ queries, which in turn support consistent AutoE evaluations. AutoD also includes tools for summarizing input or output datasets in a way that reflects their topical coverage. These summaries play an important role in the AutoQ query synthesis process, but they can also be used more broadly, such as in prompts where context space is limited. Since the release of the GraphRAG paper, we’ve received many requests to share the dataset of the Behind the Tech (opens in new tab) podcast transcripts we used in our evaluation. An updated version of this dataset is now available in the BenchmarkQED repository (opens in new tab), alongside the AP News dataset containing 1,397 health-related articles, licensed for open release.   We hope these datasets, together with the BenchmarkQED tools (opens in new tab), help accelerate benchmark-driven development of RAG systems and AI question-answering. We invite the community to try them on GitHub (opens in new tab).  Opens in a new tab

Max Towers / Max DudlerSave this picture!© Stefan MüllerResidential Architecture, Residential•Aesch, Switzerland Architects: Max Dudler Area Area of this architecture project Area:  26500 m² Year Completion year of this architecture project Year:  2024 Photographs Photographs:Stefan MüllerMore SpecsLess Specs Save this picture! Text description provided by the architects. The two Max Towers are the centerpiece of the residential and commercial quarter Aere, taking shape in the former industrial district of Stöcklin, which spreads between Reinach and Aesch south of Basel. The tower volumes, roughly 40 meters high, are striking for the continuous balconies that surround each level over a recessed ground floor. The girdle of balconies merges the urban flair of high-rise living with the qualities of generous garden space associated with a family house.Save this picture!Save this picture!Save this picture!The tectonic arrangement is completed on the outside by a mantle of solid columns with cornices integrated into monolithic ceilings. The inner façade behind these constitutes a second skin, benefiting from greater liberty as it steps forward and back to generate an efficient footprint. This meandering balcony zone enlivens the relationship between indoors and outdoors while serving as a filter and as a layer of flexible space.Save this picture!Save this picture!The entrance hallways in Tower 2 create an arrival zone and link the public space outside with the central garden that forms an open yet shielded hub for the new quarter.Save this picture! Project gallerySee allShow less Project locationAddress:Aesch, SwitzerlandLocation to be used only as a reference. It could indicate city/country but not exact address.About this officeMax DudlerOffice••• MaterialsMaterials and TagsPublished on June 06, 2025Cite: "Max Towers / Max Dudler" 06 Jun 2025. ArchDaily. Accessed . <https://www.archdaily.com/1030845/max-towers-max-dudler&gt ISSN 0719-8884Save世界上最受欢迎的建筑网站现已推出你的母语版本!想浏览ArchDaily中国吗?是否 You've started following your first account!Did you know?You'll now receive updates based on what you follow! Personalize your stream and start following your favorite authors, offices and users.Go to my stream

Mario Kart World – is it a backwards step? (Nintendo) The Friday letters page is surprised the Sony State of Play didn’t go down better with more people, as readers share images of the Switch 2 midnight launches. To join in with the discussions yourself email gamecentral@metro.co.uk Initial experience I’ve spent a few hours playing Mario Kart World in single and multiplayer, and my first impression is that it’s a bit… boring.The free roam aspect of the game seemed like it had heaps of potential but, for the most part, it feels aimless. Sure, there are things to collect and discover, but in-between there’s just driving. Lots and lots of driving. It’s not exactly a world brimming with things to do in the same way as the sandbox areas in Super Mario Odyssey, for example. As for the races, in principle it makes sense that the tracks need to be wider to accommodate 24 racers. However, the races themselves don’t generally feel any busier because you’re still, generally, just competing against those closest to your level of ability. Consequently, I felt the game lacks the chaotic energy of jostling your rivals to get ahead because there’s so much space on the otherwise well-designed courses. I know it’s unfair to compare Mario Kart World to Mario Kart 8 at this point, given the latter spanned two consoles, with a plethora of content reflecting that. But it’s hard not to feel that while the free roam section is a solid demonstration of technical advancement (even if the inability to integrate it with Grand Prix mode seems like a missed opportunity), the core gameplay is consistent, at best. In some respects, I’d say it’s more of a step backwards, if anything. That being said, I don’t think this is the same version of Mario Kart World we’ll be playing five years from now. Free roam mode feels like a playground in which Nintendo can experiment, and I fully expect them to do so. There’ll undoubtedly be a slew of DLC and perhaps better integration of the various modes into one seamless experience. Expert, exclusive gaming analysis Sign up to the GameCentral newsletter for a unique take on the week in gaming, alongside the latest reviews and more. Delivered to your inbox every Saturday morning. In the unlikely event that this is the final iteration of Mario Kart World, it’s currently a six or (generous) seven out of 10 and probably the least enthralled I’ve ever been with a new Mario Kart title. But I’ll give Nintendo the benefit of the doubt for now because this doesn’t feel like a game that’s even close to being finished.Needlemouse Just add SonyFirstly, I’m surprised by the relatively lukewarm response to the State of Play from some people; I thought it was one of the best in recent memory and there were only a few games that I personally had little interest in. The main reason for writing in is the recent addition of Destiny 2: The Final Shape to PS Plus subscription. In short, I played a lot of Destiny 2 back in the day but have subsequently not played any of the most recent DLCs/additions. Can The Final Shape be played as a standalone? (both logistically and in terms of the story)?John GC: We think people were upset that Sony themselves had nothing to show from their first parties, even though we agree it was overall a good show. As far as we understand, you don’t need any other expansions to play The Final Shape. Only one Just a quick heads up for anyone buying digital content on either a Switch or Switch 2 console, if you are Switch online subscriber you can purchase a voucher for £84 that lets you purchase any two titles from a list of games.This reduces each game to a reasonable £42 and the vouchers are valid for a year. I didn’t get an alert about this until after I’d purchased Super Smash Bros. Ultimate, which I paid £59.99 for. I did at least then go on to buy a voucher for £84 and redeemed each token for Super Mario Bros. Wonder and Super Mario 3D Land + Bowser’s Fury.Charlie H. GC: It’s important to note these can’t be used for Switch 2 exclusive games. Email your comments to: gamecentral@metro.co.uk Midnight gathering I’ve seen a lot of pictures from the Switch 2 launch, so I thought I’d send you a picture of the queue at Smyth’s in Dublin city centre, a couple of minutes after midnight.I think there were a couple of hundred people there when I arrived. if I’d known then that I wouldn’t get out of the store until 2:30am, I might have gone home and tried again in the morning. But in the end, I’m glad I waited. I’ve just been tinkering with it so far, hooked it up to the TV, and played a little bit of Mario Kart World and Cyberpunk 2077, so I can’t really give any conclusions. But when my (dedicated non-gamer) wife saw it, her reaction was, ‘Oh, it looks much better! Can I play it?’ So I think Nintendo are on to another winner.Mickah Night-time rendezvous (Mickah) American retail I was on my way down to Smyths Leeds to collect my Switch 2 Mario Kart bundle, that I had pre-ordered a month ago.I called in at Costco on the way and was surprised to find they had the Mario Kart bundle for £419.99, so I decided to get one from there. It was late afternoon Thursday and they looked to have plenty left. It is limited to one per customer. I will let my pre-order lapse so it will cancel in two days.Martin GC: We had no idea Costco existed in the UK. Maybe we’re not the only ones and that’s why they have so many left. C’mon DoreenSomebody (not me, don’t know how) should start a petition to get Squirrel Girl in that Marvel game for you guys. How good did it all look? How good was that showcase? What have Sony been up too? That’s what. Mortal Kombat: Legacy Kollection is already on my wishlist. That surfing sword game wasn’t bad either. With Nintendo too, great times ahead.Indiegaz (PSN ID) GC: At least she’s in Marvel Rivals. Autumnal purchase So I won’t be buying the Nintendo Switch 2 at launch. I’ll be instead visiting Tenerife in August. But I do hope the system reviews well and those who have purchased it are zooming around on Mario Kart World. I’ll be hopefully purchasing the device before autumn of this year. But you never know what could happen.So my sister paid for my ticket and it of course must be paid back. A debt is a debt. Which is quite fitting, since I just finished my playthrough of Red Dead Redemption and just as John Marston paid his debt to a life of crime and Edgar Ross paid his debt to a vengeful son. I’ll be hopefully paid up in less than three months, then I’ll be purchasing the Switch 2 and with more information on the games, reviews, and what’s to come for the future. For now however, I look towards a week spent in the beaches of Spain and my focus on the remainder of my maths course. Also, a playthrough of Resident Evil 3 remake. Only six hours long. Can’t complain.Shahzaib Sadiq Free for all Borderlands 2 and Hellslave are currently free on Steam on PC. Hellslave is available for free until Sunday, 15th June. Also, Deathloop is currently free on Epic Games Store.I hope everyone who gets a Switch 2 enjoys it. I will have to probably get one next year now, as I need to buy a new gaming computer because my current gaming PC will not upgrade to Windows 11, unfortunately.Andrew J. Old reliable Just writing this after a couple of hours with the Switch 2. I always end up with whatever Nintendo’s latest console is at some point in its lifetime (going back to the Game Boy Advance) but for the first time I decided to jump onboard day one.The price rises for the PlayStation 5 and Xbox Series X have shown that the assumption consoles will get cheaper in the medium term isn’t a guarantee, and the feature list for the Switch 2 ticked pretty much every box I wanted from an upgrade to the original. Bump in specs, sturdier Joy-Con, and a new interface with the mouse controls. What I’ve not seen many people comment on yet is the set-up experience. It isn’t always a given with Nintendo that this is going to be smooth, but I found that the transfer from my original to Switch 2 was seamless. I had one error message in setting up GameChat, but restarting the process fixed that, and now I can settle into at least another few years of Nintendo gaming joy. As I approach the business end of 40 years old, it’s a comfort to know that the likes of Mario, Yoshi, and the other inhabitants of the Mushroom Kingdom are still karting away 30-odd years after the SNES original.Electric Crocosaurus GC: That’s a cool name. Inbox also-rans Yes! My Switch 2 has turned up and have the rest of the week off, and it’s Summer Game Fest on Friday night. Now that is what I call eating well for games fans!LemptonAs promised, here is the pic from the queue at Smyths toy store at midnight. Sorry I couldn’t get one from inside but they were only letting two at a time in. Not a bad turn out. I had about 20 people behind me too.woz_007 (NN ID) The way launches used to be (woz_007) Email your comments to: gamecentral@metro.co.uk More Trending The small printNew Inbox updates appear every weekday morning, with special Hot Topic Inboxes at the weekend. Readers’ letters are used on merit and may be edited for length and content. You can also submit your own 500 to 600-word Reader’s Feature at any time via email or our Submit Stuff page, which if used will be shown in the next available weekend slot. You can also leave your comments below and don’t forget to follow us on Twitter. GameCentral Sign up for exclusive analysis, latest releases, and bonus community content. This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply. Your information will be used in line with our Privacy Policy

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.”

John Edwards, Technology Journalist & AuthorJune 5, 20254 Min ReadSailGP Australia, USA, and Great Britain racing on San Francisco Bay, CaliforniaDannaphotos via Alamy StockWarren Jones is CTO at SailGP, the organizer of what he describes as the world's most exciting race on water. The event features high-tech F50 boats that speed across the waves at 100 kilometers-per-hour (62 miles-per-hour).  Working in cooperation with Oracle, Jones focuses on innovative solutions for remote broadcast production, data management and distribution, and a newly introduced fan engagement platform. He also leads the team that has won an IBC Innovation Awards for its ambitious and ground-breaking remote production strategy. Among the races Jones organizes is the Rolex SailGP Championship, a global competition featuring national teams battling each other in identical high-tech, high-speed 50-foot foiling catamarans at celebrated venues around the world. The event attracts the sport's top athletes, with national pride, personal glory, and bonus prize money of $12.8 million at stake. Jones also supports event and office infrastructures in London and New York, and at each of the global grand prix events over the course of the season. Prior to joining SailGP, he was IT leader at the America's Cup Event Authority and Oracle Racing. In an online interview, Jones discusses the challenges he faces in bringing reliable data services to event vessels, as well as onshore officials and fans. Related:Warren JonesWhat's the biggest challenge you've faced during your tenure? One of the biggest challenges I faced was ensuring real-time data transmission from our high-performance F50 foiling catamarans to teams, broadcasters, and fans worldwide. SailGP relies heavily on technology to deliver high-speed racing insights, but ensuring seamless connectivity across different venues with variable conditions was a significant hurdle. What caused the problem? The challenge arose due to a combination of factors. The high speeds and dynamic nature of the boats made data capture and transmission difficult. Varying network infrastructure at different race locations created connectivity issues. The need to process and visualize massive amounts of data in real time placed immense pressure on our systems. How did you resolve the problem? We tackled the issue by working with T-Mobile and Ericsson in a robust and adaptive telemetry system capable of transmitting data with minimal latency over 5G. Deploying custom-built race management software that could process and distribute data efficiently [was also important]. Working closely with our global partner Oracle, we optimized Cloud Compute with the Oracle Cloud.  Related:What would have happened if the problem wasn't quickly resolved? Spectator experience would have suffered. Teams rely on real-time analytics for performance optimization, and broadcasters need accurate telemetry for storytelling. A failure here could have resulted in delays, miscommunication, and a diminished fan experience. How long did it take to resolve the problem? It was an ongoing challenge that required continuous innovation. The initial solution took several months to implement, but we’ve refined and improved it over multiple seasons as technology advances and new challenges emerge. Who supported you during this challenge? This was a team effort -- with our partners Oracle, T-Mobile, and Ericsson with our in-house engineers, data scientists, and IT specialists all working closely. The support from SailGP's leadership was also crucial in securing the necessary resources. Did anyone let you down? Rather than seeing it as being let down, I'd say there were unexpected challenges with some technology providers who underestimated the complexity of what we needed. However, we adapted by seeking alternative solutions and working collaboratively to overcome the hurdles. What advice do you have for other leaders who may face a similar challenge? Related:Embrace adaptability. No matter how well you plan, unforeseen challenges will arise, so build flexible solutions. Leverage partnerships. Collaborate with the best in the industry to ensure you have the expertise needed. Stay ahead of technology trends. The landscape is constantly evolving; being proactive rather than reactive is key. Prioritize resilience. Build redundancy into critical systems to ensure continuity even in the face of disruptions. Is there anything else you would like to add? SailGP is as much a technology company as it is a sports league. The intersection of innovation and competition drives us forward and solving challenges like these is what makes this role both demanding and incredibly rewarding. About the AuthorJohn EdwardsTechnology Journalist & AuthorJohn Edwards is a veteran business technology journalist. His work has appeared in The New York Times, The Washington Post, and numerous business and technology publications, including Computerworld, CFO Magazine, IBM Data Management Magazine, RFID Journal, and Electronic Design. He has also written columns for The Economist's Business Intelligence Unit and PricewaterhouseCoopers' Communications Direct. John has authored several books on business technology topics. His work began appearing online as early as 1983. Throughout the 1980s and 90s, he wrote daily news and feature articles for both the CompuServe and Prodigy online services. His "Behind the Screens" commentaries made him the world's first known professional blogger.See more from John EdwardsWebinarsMore WebinarsReportsMore ReportsNever Miss a Beat: Get a snapshot of the issues affecting the IT industry straight to your inbox.SIGN-UPYou May Also Like

Key Takeaways The European Commission will introduce the new International Digital Strategy that will focus on tech collaboration with the US and other countries. This is in contradiction to the growing pressure in Europe that calls for tech sovereignty and reducing tech dependence on the US. Low venture capital funding and a diverse regulatory landscape have been major challenges for tech innovation in Europe. Europe is seeing an increasing push to establish technological sovereignty and reduce the region’s reliance on US technology. However, there’s always been an undertone of acceptance that the EU is years behind the US when it comes to technological innovation and advancements. Now, the European Commission is planning to acknowledge this publicly. The EC will introduce a new International Digital Strategy, which will focus on collaboration with the United States and other tech-forward countries, such as Japan, South Korea, and India. The lawmakers believe that “decoupling” from the West is unrealistic and will instead push Europe further back in the technological race. Europe’s Call for Tech Sovereignty Several prominent political leaders and lawmakers have advocated for European sovereignty over technology and artificial intelligence.  For instance, Emmanuel Macron, the president of France, said in a speech in 2024 that Europe’s strategic autonomy is a conscious choice to end the region’s dependence on others. In an earlier speech, he also said that if Europe fails to build champions in areas such as digital and artificial intelligence, its choices will be dictated by others.  Similarly, Thierry Breton, the former EU Commissioner for Internal Markets, said that digital spending in the EU will breach the 20% target, underlining the importance of investing in European tech sovereignty. He focused on Europe’s declining market share in the semiconductor industry and called for the development of groundbreaking European tech. The Eurostack movement in the EU has also been gaining a lot of traction and support from various think tanks, academic researchers, and industry voices. Eurostack staff calls for the development of an indigenous infrastructure stack in the European region, including cloud, AI, semiconductors, digital services, and data centers. The main aim is to reduce Europe’s dependence on Chinese and US technology. This growing concern among EU well-wishers is quite understandable. Excessive reliance on the United States puts the US (meaning Donald Trump) in a controlling position, where he can arm-twist the European Union in matters of trade and even politics.  We have already seen an example of this during the tariff war, where Trump imposed a 25% tariff on automobiles (and their parts) and steel and aluminum products imported from the EU. Also, the fact that Google, Microsoft, and Amazon account for around 69% of the cloud market in the EU is quite concerning, too. Europe Wakes Up to Reality Despite positive speeches and statements by people in power in the EU, the fact remains that Europe is still lagging behind the United States. Bulgarian lawmaker Eva Maydell said that Europe should “sober up” and accept that the train has left the station. Dan Nechita, the current EU director for the Transatlantic Policy Network, said that it’s not the right time to be politically absolute and say that “we are going to do everything in Europe.” EU tech chief Virkkunen has been working hard to bring home the support of influential tech lobbies and emphasized the need to continue working with the United States. To put it in a nutshell, the European Commission is ready to accept the fact that the damage is done, and now Europe needs to play second fiddle and forge strategic partnerships with key technological leaders (primarily in the US) to stay alive in the race. Lack of Investments and the EU’s Regulation-First Approach One of the major reasons behind the European Union’s sluggish tech development is the lack of venture capital investments in the region. As per an IMF post, EU VC funds raised around $130B between 2013 and 2022. During the same time, the US raised $924B.  Similarly, annual VC investments in the EU are only 0.2% of the GDP when compared to 0.7% in the US. Currently, the United States accounts for a massive 52% of the global VC funds, whereas the EU holds just 5%.  Lack of investments has forced European startups to look elsewhere, especially the United States, for funding and support. One of the major reasons for such low venture capital interest in Europe is its diversity. Each of the 27 countries in the EU has its own regulatory and legal challenges, which make it difficult for multinational corporations to operate in the region.  Plus, the EU has adopted a regulation-first approach, which is very different from the United States. Of course, this approach has its own merits, but it has surely slowed down the speed of technological investments in the region.  For instance, the GDPR puts a truckload of regulatory and moral responsibility on companies to protect user data. Similarly, the AI Act focuses on the more ethical development of artificial intelligence that’s aligned with human values and refrains companies from exploiting public data. Sure, all of these are positive tech regulations important to protect the long-term sovereignty of the public at large. Even global tech advocates have praised the EU’s efforts for the development of safe and ethical technologies. However, we cannot ignore the fact that this has come at the cost of sluggish tech investments and overall growth. With its back against the wall, Europe needs to reassess its strengths and focus on areas such as open-source technologies, such as France’s La Suite numérique, and government-backed technological initiatives to have a say in the upcoming artificial intelligence and semiconductor race. Krishi is a seasoned tech journalist with over four years of experience writing about PC hardware, consumer technology, and artificial intelligence.  Clarity and accessibility are at the core of Krishi’s writing style. He believes technology writing should empower readers—not confuse them—and he’s committed to ensuring his content is always easy to understand without sacrificing accuracy or depth. Over the years, Krishi has contributed to some of the most reputable names in the industry, including Techopedia, TechRadar, and Tom’s Guide. A man of many talents, Krishi has also proven his mettle as a crypto writer, tackling complex topics with both ease and zeal. His work spans various formats—from in-depth explainers and news coverage to feature pieces and buying guides.  Behind the scenes, Krishi operates from a dual-monitor setup (including a 29-inch LG UltraWide) that’s always buzzing with news feeds, technical documentation, and research notes, as well as the occasional gaming sessions that keep him fresh.  Krishi thrives on staying current, always ready to dive into the latest announcements, industry shifts, and their far-reaching impacts.  When he's not deep into research on the latest PC hardware news, Krishi would love to chat with you about day trading and the financial markets—oh! And cricket, as well. View all articles by Krishi Chowdhary Our editorial process The Tech Report editorial policy is centered on providing helpful, accurate content that offers real value to our readers. We only work with experienced writers who have specific knowledge in the topics they cover, including latest developments in technology, online privacy, cryptocurrencies, software, and more. Our editorial policy ensures that each topic is researched and curated by our in-house editors. We maintain rigorous journalistic standards, and every article is 100% written by real authors.