Energy vs. hearts.Last week, Duolingo announced a major shift: they’re replacing their long-standing hearts system with a new “energy” mechanic. The company frames this as a way to make learning more motivating by focusing on rewarding correct answers rather than punishing mistakes.But after spending 48 hours thinking about this change and analyzing reactions across the internet, I’ve come to see it as something far more nuanced: a masterclass in behavioral psychology that serves both users and Duolingo’s bottom line in fascinating ways.What’s actually changingUnder the old hearts system, users started with five hearts and lost one for each mistake. When all hearts were depleted, you couldn’t continue learning until you either waited (hearts regenerated over time), watched ads, practiced old content to earn hearts back, or paid for premium.The new energy system gives users 25 energy units. Each lesson costs one unit to start, and mistakes also cost energy. The key difference? Users can earn bonus energy at random intervals when answering correctly.As Moses Wayne, a senior staff engineer at Duolingo, explains to The Verge: “We feel like this is a way that we can motivate you to focus on things you’re getting right rather than penalizing for the things that you’re making mistakes on.”It sounds positive on the surface. But there’s much more happening under the hood.The psychological switch: from loss to expenditureThe hearts system was built on loss aversion (our tendency to feel losses more intensely than equivalent gains). Users would desperately try to avoid mistakes to keep their hearts. In psychological terms, this creates anxiety that can inhibit learning, especially for a complex skill like language acquisition.The energy system cleverly shifts the frame from “losing hearts through mistakes” to “spending energy to learn.” This slight repositioning transforms the psychological experience. Instead of punishment, it’s now an investment.But here’s where it gets particularly interesting, and perhaps concerning.At time of writing, users will still be seeing this when they run out of hearts.Variable rewards: the slot machine mechanismThe new system incorporates one of the most powerful psychological techniques in existence: variable reward schedules, or what behavioral scientists call “intermittent reinforcement.”When completing lessons, users now receive random bonus energy for consecutive correct answers. This unpredictability is the same mechanism that makes slot machines so addictive. As behavioral psychologist B.F. Skinner discovered decades ago, unpredictable rewards create stronger habit loops than predictable ones.The uncertain nature of “when will I get bonus energy?” creates a dopamine response that keeps users engaged far longer than the more predictable hearts system. One psychology research article noted that “variable ratio schedules have two benefits: They result in the most instances of the behavior than any of the other schedules… and they result in behaviors that are ‘hard to extinguish,” meaning the behavior persists even when rewards decrease.Duolingo isn’t the first app to use this technique. It’s everywhere in modern digital products, from social media (random likes triggering dopamine) to games (loot boxes with variable rewards). What’s notable is how Duolingo is implementing it in an educational context.The inverse energy model: Finch vs. DuolingoTo understand what makes Duolingo’s approach unique, it’s worth comparing it to another popular app that uses an energy system: Finch.In Finch, a self-care app with a virtual pet, users earn energy by completing wellness tasks. As one reviewer describes it: “My effort gives [my pet] energy, but my lack of effort takes none away.” The energy you earn through self-care allows your pet to go on adventures and grow.Energy in Finch and Duolingo are both symbolized by a ⚡Duolingo takes the inverse approach, i.e. you spend energy to engage with the product. This fundamental difference reveals contrasting philosophies:Finch’s model: Do real-world activities → earn in-app energy → enjoy rewardsDuolingo’s model: Spend in-app energy → engage with content → occasionally get energy backWhile both use similar terminology, the direction of energy flow completely changes the behavioral dynamics. One rewards external action; the other creates a closed loop that keeps users in the app.The business of artificial scarcityLet’s not overlook the monetization angle. Super Duolingo subscribers get unlimited energy, just as they previously got unlimited hearts. The new system maintains the same fundamental constraint: free users eventually hit a wall where they need to either wait, work harder, or pay up.What the energy system might do more effectively is increase the perceived value of the premium subscription. By creating a more engaging core loop with the variable rewards, users may be more likely to convert when they hit energy limits.The strategy is brilliant, if perhaps ethically complex. By transforming a negative experience (punishment for mistakes) into a more positive ne (spending a resource with occasional bonuses), Duolingo has maintained its monetization constraints while making them feel less punitive.Critics of the heart system have long pointed out that unlimited hearts are one of the main selling points of Duolingo’s premium tier, suggesting the constraint exists primarily to drive conversions rather than improve learning. The energy system continues this pattern, just with more sophisticated psychological underpinnings.The learning science questionThe most important question is whether this change actually improves learning outcomes.Duolingo claims the energy system helps users “get through more lessons,” as the data shows users can do more with the new system. But completing more lessons doesn’t necessarily mean better language acquisition.Language learning requires making mistakes. It’s a fundamental part of the process. The old heart system discouraged risk-taking by heavily penalizing errors, potentially leading users to stick with easier content where they wouldn’t lose hearts.The energy system might ease this anxiety somewhat — making a mistake costs the same as starting a new lesson. But the fundamental constraint remains: you still have a limited resource that depletes when you make errors.A truly learning-optimized approach might involve no penalties for mistakes at all, focusing instead on spaced repetition, comprehensible input, and other evidence-based language acquisition methods. But such an approach might not drive the same level of engagement or monetization.Duloingo falls somewhere in the center of the bottom right quadrantUsers vs. learnersThis tension highlights a key challenge for educational technology: the gap between what creates engaged users and what creates successful learners.The techniques that make an app sticky (intermittent reinforcement, streaks, gamification) don’t always align with optimal learning science. In fact, they can sometimes work against it by encouraging shallow engagement over deep processing.Duolingo has always walked this line, using game mechanics to keep people engaged with language learning when they might otherwise quit. The energy system represents an evolution of this approach, using more sophisticated behavioral psychology to maintain the same fundamental constraints while making them feel better.The verdict: clever design for multiple objectivesAfter analyzing this change from multiple angles, I’m left impressed by its cleverness while somewhat ambivalent about its implications.The energy system:Creates a more positive psychological framing than heartsImplements powerful variable rewards that increase engagementMaintains the same fundamental monetization pressurePotentially allows for more mistakes, which is better for learningUses randomness to create stronger habit loopsIt’s a change that serves both user experience and business goals, likely increasing both engagement and conversion to premium. Whether it truly serves learning outcomes remains to be seen, though it’s probably at least marginally better than the hearts system in this regard.What fascinates me most is how this change reveals the sophistication of behavioral design in today’s digital products. What looks like a simple mechanic change is actually a carefully engineered system to shape behavior through psychological principles.The next time you find yourself compulsively swiping for another lesson, hoping for that bonus energy to drop, remember that this feeling isn’t accidental… it’s by design.And while I appreciate the brilliance of that design, I find myself wishing more of this psychological ingenuity was directed toward optimizing learning outcomes rather than engagement metrics. Perhaps that’s the next frontier for language learning apps: finding ways to make evidence-based learning as addictive as a slot machine.Sam Liberty is a gamification expert, applied game designer, and consultant. His clients include The World Bank, Click Therapeutics, and DARPA. He teaches game design at Northeastern University. He is the former Lead Game Designer at Sidekick Health.Duolingo’s small UI switch that changes everything was originally published in UX Collective on Medium, where people are continuing the conversation by highlighting and responding to this story.

On an unremarkable Monday last February, Andrej Karpathy fired off a tweet that gave the internet its new favorite buzz-phrase: vibe coding.Within hours, people watched tools like Bolt, v0, and Lovable conjure apps from mock-ups never designed or developed.The internet cheered—speed looks spectacular in demo reels—but more senior groups were quietly wincing as AI began to add technical debt to large codebases at previously impossible rates.Why demo-first AI fails mature teamsHere’s how senior designers, developers, and marketers feel the pain.Designers: Prefab design systems ≠ your design systemToday’s one-click generators choose their own colors, border-radii, and fonts. The instant they collide with a house style they don’t recognize, they hard-code new hex values that overwrite your brand tokens, throw off the grid, and leave designers sending screenshots of mismatched styles to the bug tracker.When you draw a picture, you spend your time carefully crafting the shapes and lines. Great design comes from consistency of intention.Today’s AI designers give you a page full of chicken scratches, an eraser, and a “good luck.”Developers: One-shot codegen ≠ production code.A “fully-working” React page arrives as a single 1,200-line component—no tests, no accessibility tags, no separation of concerns. The diff looks like a CVS receipt.The dev who merges it now basically owns a tarball that will resist every future refactor. Code feels like inheriting a junior’s side project, but the junior in this case is an LLM that never sleeps and can’t be taught.Clean, maintainable code is measured by how many lines you can still cut. Without context of your existing codebase or a lot of fine-tuning, AI’s tendency is to be too verbose.Marketers: Fake data ≠ live experiments.Marketing can now crank out a landing page in fifteen minutes, but every testimonial, price, and CTA is still lorem ipsum.Wiring the page to the CMS—and to analytics—means rewriting half the markup by hand. Every 10x sprint lands on an engineering backlog.You can generate pages, sure, but you’re gonna have to wait on the revenue.What’s left for us?Call it all the 80/20 hangover: AI generates the first 80% of the job in 20% of the time… then teams grind through the next 80% of the time patching the final 20% of the work.It’s the perfect inversion of healthy collaboration. Humans drown in drudgery and plead with the machine to handle the craft. But here’s the thing: The lesson isn’t that AI is snake oil; it’s that big team problems aren’t solved by fast AI alone.Sure, “Make me a meal-planning app” is a fun toy, but as we’ve seen, speed without fidelity just moves the bottleneck downstream.What we need is to respect the craft already living in your components, tokens, data, and tests. We need AI for grown-ups.The two-sided contract of AI for grown-upsHere’s the proposal: what if we started building AI with competent teams in mind?For that kind of tooling, two mindset shifts need to happen, both from the builders of AI software and the professionals that use it.Builders need to…Respect the stack. Every pixel, prop, token, schema, and test that already lives in a repo becomes the rails, not a suggestion. Professionals should be able to easily predict what AI outputs will look like. Stay in their lane. Instead of inventing Yet-Another-Canvas, the tooling embeds itself in the software the best teams already use: Figma, IDEs, headless CMSes, etc. Professionals work hard to choose their tools, and AI alone is not a compelling enough reason to switch. Expose control. Codegen happens in daylight—via CLI hooks, readable diffs, and human review gates—so senior engineers can keep the quality bar where it belongs. Professionals are smart enough to handle the machine, so long as it shows them what it’s doing.It’s not up to the professional world to keep perfect pace with the runaway train of AI. It’s the tool builders’ responsibility to stop breaking so much stuff.Professionals need to…Feed the context. Put in the work to document processes, map prototypes to real components, hand over design tokens, and write the first round of tests. Give the model a fighting chance at fidelity. Builders will use this to make AI generations far more deterministic.Stay accountable. A merged PR is still a human name on the line for the most intricate of crafts: UX polish, performance budgets, sparkling copy, etc. Builders don’t have to design all-star AI; they can focus on making it consistent at the grunt-work.Recognize AI limits. Designers, developers, and marketers use AI primarily as the translation layer between deeply skillful fields. Builders don’t need to make AI that replaces teams, but rather AI that fosters communication in the handoffs.It’s not up to the Builders to anticipate every team’s exact use case. Instead, they can trust that professionals will rise to the challenge of adapting AI to their needs. How AI for grown-ups works on a given MondayWell, that’s all great: People focus on people things; AI focuses on AI things. But what does it look like in practice?Let’s take the perspective of our three personas again—designers, developers, and marketers—and see what an ideal world looks like for them.Designers, still in FigmaA designer adjusts the corner radius on the hero card and watches the live hand-off panel flag the exact React prop it will touch.A token audit bot pings him only if the value breaks the design system—no more screenshot spreadsheets.Freed from red-lining, he spends the afternoon nudging micro-interactions and pairing with motion design, really making everything sing.Way more time on polish. No more token-drift bugs filed.Developers, still in the IDEA developer pulls the latest “design-to-code” PR: it’s already split into sensible components, test stubs green-lit, and the diff is small enough to review over coffee.She runs the performance benchmark—numbers hold, thanks to preset budgets the generator never crosses.With boilerplate gone, she dives into accessible keyboard flows and edge-case logic that actually moves the product forward.Review cycles on layout bugs drop by half with AI taking over scaffolding.Marketers, still in the CMSA marketer duplicates last week’s landing page variant, swaps headline copy, and clicks “Stage Experiment.”The AI wires analytics, snapshots the control, and opens a PR tagged for growth-team review—no Jira ticket, no dev backlog delay.They schedule the A/B test in the same dashboard and spend their saved hour crafting social hooks for the launch.Campaign velocity doubles, while engineering time spent on “tiny copy changes” shrinks to near-zero.Same tools, less tedium, deeper ownership. Each role feeds context to the system, stays accountable, and lets the AI do the grunt work—that’s the contract in action.Keep your Cursor, keep your ChatGPT—the glue-layer AI just has to play nice with them while sitting squarely in the team’s shared stack. How we’re building toward AI for grown-upsSo, how close are we to that perfect Monday?At Builder, we find ourselves in a unique position to help, and we’ve been pivoting to meet the demand, especially for large teams.A tiny history lessonBuilder launched in 2019 as a headless CMS plus visual editor whose goal was simple: let developers surface their own JavaScript framework components, and then let non-developers arrange those components without touching code.To pull that off, we built our editor on Mitosis, an open-source layer that describes components once and compiles them to whatever JS framework a team already runs.Because of that early direction, three “grown-up parts” of our product were firmly in place before the generative-AI wave arrived:Fully-featured visual editing: Our Webflow-like editor creates space for designers and marketers to tweak components and pages that compile to any JS framework. Deterministic component mapping: Every Figma frame and its design tokens can round-trip with a real, version-controlled component—not a look-alike snippet. A data source with real content: The CMS holds marketing content, sure, but also all the data engineers need for accurate UI states.So when large language models became practical, we didn’t bolt AI onto a blank slate; we layered it onto an editor, a mapper, and a CMS that were already respecting the stack.That head-start is why our current AI features can focus on removing drudgery instead of reinventing fundamentals.What we’ve made so farOur product works as a bunch of incrementally adoptable layers. No big bang rewrites needed.Visual Editor → Code. Think Bolt or v0, but with Webflow-level tweakability. Prompt the AI to draft a page—or paste any URL for an instant, editable clone—then fine-tune any class, token, or breakpoint by hand. Drop in production-ready components from 21st.dev (or your own repo) and ship designs that respect your CSS, not ours.Figma → Visual Editor (and Code). Install the free Figma Visual Copilot plug-in, draw your frame, and click Export. Copilot converts the frame’s exact Auto-Layout geometry into clean, responsive framework code that can drop straight into the repo or open in Builder’s visual editor for tweaks. Designers still hand over a Figma link; developers run one CLI command to scaffold production-ready code—no guessing, no style drift.Repo → Visual Editor → Repo PR [coming soon]. We’ll be announcing a product shortly that allows anyone to import any GitHub repo to the Visual Editor and make changes that get auto-sent as a PR. Marketers and designers can file engineering tickets with the code already in nice shape.Component Mapping. Tell Builder which Figma button matches which code button once, click “Publish,” and you’re set. From then on, every design export that uses that component drops the real one from your repo into the generated diff—props, tokens, everything—so code reviews focus on ideas, not mismatched styles.Builder Publish: It’s not just a patch for lorem ipsum; it’s your entire headless CMS. Wire any page to real-time content and analytics so marketing can run A/B tests without tapping devs—and get the full schema, versioning, and single-source-of-truth perks that come with a modern CMS, all under the same roof.Where we’re still workingWe’re close, not finished. Next on our list:Less manual context. Auto-detect more tokens and components so mapping feels like autocomplete, not data entry. Deeper control. Let power users crack open every AI draft—props, tests, perf budgets—without leaving the editor. Broader design-system support. Shadcn today, your bespoke kit tomorrow. Mapping should be one click, not a weekend.That “perfect Monday” vision is where we’re headed, and our roadmap will get us there one feature at a time.We’re definitely interested in your feedback: What gaps do you still feel between design, code, and marketing, and where could an “AI for grown-ups” save you the most pain?Drop your thoughts in the comments or tag us on our socials (see footer for links).Turning the 80/20 right-side-upIf Software 2.0 taught us that networks can write code, and the Software 3.0 hype reminds us they can write whole interfaces, then the middle road we’ve walked today—call it Software 2.5—insists that learned programs still deserve linting, budgeting, and daylight.Boring constraints create extraordinary freedom.Speed is intoxicating, but alone it flips the 80/20 rule on its head, leaving humans elbow-deep in cleanup. AI for grown-ups reverses that trade.Tool builders promise rails, not rabbit holes. Professionals promise context and accountability. Every mapped component, every token lock, every human review makes the next generation more predictable—and predictability is what compounds into real velocity.We don’t need to be dragged kicking and screaming into the new world. We’re perfectly capable of walking there at our own pace, spending our hours on craft no model can fake. Introducing Visual Copilot: convert Figma designs to high quality code in a single click.Try Visual CopilotGet a demo

By Passant Rabie Published May 15, 2025 | Comments (0) | The Athena lander ended up lying on its side on the surface of the Moon. Intuitive Machines Earlier this year, an ill-fated lander skidded across the lunar surface and ended up lying sideways in a cold, dark crater. Intuitive Machines’ Athena lander was declared dead on arrival at the Moon. Now, the company has released a report detailing the chain of mishaps that led to its demise. Athena touched down on the Moon on Thursday, March 6, following a week-long journey through space. After a less-than-ideal touchdown, the lunar lander ended up on its side in a shallow crater in the Moon’s Mons Mouton region, 820 feet (250 meters) from its targeted landing site. During an earnings call on Wednesday, Intuitive Machines listed three factors that affected Athena’s ability to land on the Moon, SpaceNews reported. The first issue had to do with the lander’s laser altimeters, instruments designed to help it estimate its distance from the Moon during descent. “In the final phase of descent, we saw signal noise and distortion that did not allow for accurate altitude readings,” Intuitive Machines CEO Steve Altemus said during the call. The Houston-based company also blamed conditions at the Moon’s south polar region. According to Altemus, the sunlight’s low angles and the topography “created long shadows and dim lighting conditions that challenged the precision capability of our landing system.” The third factor that resulted in Athena’s lunar tumble was the lander’s inability to recognize craters on the Moon from lower altitudes. Intuitive Machines relied on photos captured by NASA’s Lunar Reconnoissance Orbiter to guide the lander’s optical navigation system during its descent. Those images, however, are captured at an altitude of 62 miles (100 kilometers) and therefore “could not accurately account for how craters appear at lower altitudes with south pole lighting conditions as you approach the landing site,” Altemus said. The company is treating Athena’s failed mission as a set of lessons learned as it prepares to launch its third lander sometime next year. “We’ve added dissimilar and redundant altimeters to the sensor suite, and they’re going through more rigorous and extreme flight-like testing than we’ve done before,” Altemus said in Tuesday’s call, according to Space.com. “We’ve incorporated an additional lighting-independent sensor for surface velocity measurements. We’ve expanded onboard terrain crater database for enhanced navigation across the surface of the moon.” Intuitive Machines launched its first lunar lander, Odysseus, in February 2024. It reached the lunar surface, but the landing wasn’t exactly smooth. One of the lander’s legs may have gotten caught during its descent, causing it to tip onto its side and end up lying sideways on a rock. The mission did operate for seven days on the lunar surface, making Intuitive Machines the first company to land a private lander on the Moon. “Moving forward, we will succeed,” Altemus said. “Land softly, land upright, land ready to operate.” Daily Newsletter You May Also Like By Passant Rabie Published May 10, 2025 By Passant Rabie Published May 9, 2025 By Passant Rabie Published May 8, 2025 By Isaac Schultz Published May 2, 2025 By Passant Rabie Published May 2, 2025 By Passant Rabie Published May 1, 2025

Discover the story behind the portrait of Belgium’s maverick salvager-designer, featured in the Circularity issue In the May 2025 issue of The Architectural Review, themed Circularity, the AR commissioned a Reputations profile of the little-known Belgian architect, salvager and entrepreneur Marcel Raymaekers. This maverick designer ‘never qualified as an architect’, write James Westcott, Lionel Devlieger, Arne Vande Capelle, Stijn Colon and Aude-Line Dulière in their text, but ‘that did not stop him from procuring the materials for, designing and building around 150 villas across Belgium.’ These eccentric buildings, which are often dismissed as so many ugly Belgian houses, provide an unlikely source of lessons for contemporary material reuse, the authors argue. For every Reputations story, the AR’s editors commission an illustrator to create a portrait of the subject, incorporating elements from their work, or evoking its spirit. Raymaekers’ portrait was created by Laslo Antal, a Hungarian artist and musician based in Berlin who contributed to the magazine in 2020 with a cover design for its Criticism issue, and in 2021 with a Reputations portrait of Kurt Schwitters in its Waste issue. ‘When creating the portrait of Marcel Raymaekers, my goal was to develop an idea that would integrate his architectural and design approach,’ Antal tells the AR. ‘I wanted to create an image using these concepts rather than simply collaging or drawing a building or a part of his existing architectural work.’ Beyond using the murky palette that Raymaekers frequently used in his exterior and interior designs, Antal was particularly inspired by Raymaekers’ practice of giving objects new and different meanings by repurposing them, ‘using the old in a new way,’ he explains. ‘I also found a strong personal connection to his approach of working in 3D (when planning a project) rather than drawing architectural plans on paper.’ The illustration invokes Raymaekers’ features by using protruding plastic cups for his eyes, and cardboard for his nose, eyebrows and mouth. Additionally, part of the neck was made using the same plastic cups cut in half – ‘a homage,’ Antal says, ‘to the fighter jet skylights that famously appear in his House Klechtermans of 1970.’ Read Marcel Raymaekers’ full Reputations profile, and explore the rest of the Circularity issue here. Laslo Antal’s portfolio can be found here. 2025-05-15 AR Editors Share AR May 2025CircularityBuy Now

John Edwards, Technology Journalist & AuthorMay 15, 20255 Min ReadTuta via Alamy Stock PhotoGetting a new IT project off the ground requires a concept, a goal, meticulous planning, and, perhaps most importantly, a willingness to assume a certain amount of risk. It's this last point that often keeps IT leaders awake at night. Addressing risk prior to launching an IT project is crucial, since it helps to set clear expectations, identify potential obstacles, and create a roadmap for handling unforeseen challenges, says Anbang Xu, founder of JoggAI, an AI-powered video platform, and a former senior product manager at Apple and senior software engineer at Google. "Particularly when it comes to AI-driven projects, understanding the risks -- whether they're related to technology, data security, scalability, or talent -- can make the difference between a smooth launch and a failed initiative," he observes in an email interview. "By addressing these concerns early, you also ensure that resources are allocated effectively, and your team is better equipped to handle surprises." Building a strong mechanism to address risk during project planning is crucial to ensure that all decisions are made with transparency into cost, benefit, and risk, says Ola Chowning, a partner with technology research and advisory firm ISG. "Project decisions that don't estimate risk and its tangible or intangible impacts at the start, or even during a project, likely don't have the data needed to make informed decisions and balance cost, risk, and benefit," she explains in an online interview. "More importantly, failing to address risk can waste time if an event occurs and the team is left scrambling to determine how to respond." Related:There are also secondary risks, including budget, resources, and technology challenges, says Munir Hafez, CTO of Independence Pet Holdings via email (Hafez was senior vice president and CIO at credit reporting firm TransUnion at the time of the interview.) "Each of these risks should be addressed before launching an IT project to ensure that the project is starting off on a solid foundation and will gain alignment with all stakeholders." Accurate Estimations By quantifying risks, project managers can objectively assess their potential likelihood and impact, allowing efficient resource allocation and the development of appropriate mitigation strategies, Hafez says. "This structured approach ensures that no significant risks are overlooked and that the project remains on track." The best way to estimate risk is to divide the project into small components and assess the uncertainty around each unit, Xu says. This process includes examining the project's technical feasibility, the team's ability to handle complexity, and any potential dependencies on third-party tools or services. Xu states he relies on a combination of historical data from similar past projects and predictive models -- particularly when it comes to AI and machine learning applications, where it's often unknown how algorithms will perform at scale. "This allows us to develop a more nuanced risk profile, instead of relying on assumptions," he explains. Related:Many organizations use a risk scoring tool, built on previous project experiences and common risk categories, to guide project teams on risk types for their particular project, Chowning says. "Some [organizations] even have statistics on how often risks have occurred in the past and the average impact on previous projects." Risk Tips If a risk estimate appears overly optimistic, the leader should reassess the situation and adjust the project plan accordingly, Hafez says. "This may involve consulting industry experts or revisiting lessons learned from previous internal projects to accurately reflect the risk." The biggest mistake is underestimating the impact of technical debt and the complexity of scaling solutions, Xu warns. "Leaders often focus too much on the initial development phase and don’t account for the ongoing maintenance and optimization required," he explains. "Risk should be measured not only by the initial launch, but also by the long-term implications of maintaining and improving the post-launch system." Related:Final Thoughts As a project progresses, existing risks can evolve and new risks may emerge, Hafez says. "Regularly reviewing and updating the risk management plan ensures that the project will remain resilient and adaptable to these changes." Fostering a culture of open communication and encouraging team members to report potential risks early can significantly enhance a project's ability to respond effectively, Hafez says. This proactive approach helps avoid the "watermelon effect," where performance metrics appear positive (green) on the surface, but reveal underlying problems (red) when examined more closely. Embrace flexibility, Xu advises. In tech, especially in fields such as AI, things are constantly evolving. "A leader’s ability to adapt and pivot when necessary is as critical as the ability to plan," he notes. That's why it's important to continue refining risk assessments throughout the project's life. It’s important for project team and business decision-makers to understand that risks aren't a bad thing, Chowning says. "The underlying objective of risk management is to understand and mitigate risk, and to be prepared if risk events occur." Making high project risk scores a "bad thing" will simply drive team leaders and members to underestimate, ignore, or even attempt to hide potential risk. 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

Part antiques dealer, part architect, this Belgian maverick is an unlikely source of lessons for contemporary material reuse Marcel Raymaekers (1933–) never qualified as an architect. This did not stop him from procuring the materials for, designing and building around 150 villas across Belgium. The exact number is unknown: Belgium’s architectural establishment mostly ignored and rejected Raymaekers just as he ignored and rejected them, and as a result his work has never been collated or taken seriously. Raymaekers himself never kept records of his work; he regarded the drawing board as a ‘torture device’, preferring sketches and improvisation on site, determined by the material batches he had to hand. The extent of Raymaekers’ oeuvre is also hard to fathom because he wanted to keep it under the radar of the tax authorities. They caught up with him eventually; he was bankrupted in 2014. Today, Raymaekers is a 92‑year‑old tenant living inside what was once his personal empire, Queen of the South, in the Limburg countryside in north‑western Belgium. This compound was – and to a limited extent, still is – a vast salvage yard packed with components scavenged from aggressive postwar demolition. This puzzling place, completed in 1972 but constantly expanded, is stuffed with billowing pitched roofs, impossible archways, numerous appendages and antechambers. In its heyday – the 1970s and ’80s – Queen of the South also had a nightclub, a restaurant, a jazz venue, Raymaekers’ own ultra‑luxe apartment, and an estaminet – a plush, wood‑panelled, banquet‑seated, mirror‑pillared café‑bar. Marcel Raymaekers was born in 1933 near Leuven in Belgium. In 1950, he enrolled to study architecture at the Sint-Lukas School in Brussels but left a year later without a degree. Though not officially an architect, Raymaekers proceeded to design around 150 projects, from suburban homes to hotels and music venues – in 1986, he completed the Orlando discotheque, which burned down in 2014 Credit: Roger Dyckmans It was here that Raymaekers would sit – when he was not criss‑crossing Belgium scouting for materials – scoping out clients as they stumbled giddily into his world, overwhelmed by the aspiration and status anxiety that Queen of the South was designed to induce. The project was also a cultural hub attracting misfits and eccentrics, wheelers and dealers, experts and charlatans. The novelist Hugo Claus had his 50th birthday at Queen of the South, honouring Raymaekers in his speech. Raymaekers and his wife Hilde did interviews for newspapers, magazines and TV, often while reclining on their bed. The media was not interested in him as an accomplished architect, but as a purveyor of kitsch, an absurdity.  What Raymaekers had to sell his clients was more than simply antiques or even houses composed of them (buy enough materials and he would design you a house for no charge: that was always the deal). What Raymaekers was really selling were dreams of nobility: the life of a lord and lady of the manor, set apart (and above) from an increasingly modernised, homogenised world. Sensuality and hedonism, expressed through haptic, resplendent materials, was a big part of the allure too – the promise of a new way of life. One of Raymaekers’ bigger projects was a love hotel, Rubensexclusief, near Diest, completed in 1979. Each chamber was bestowed with an excess of padded velvet upholstery, often creeping up the walls and curving onto the ceilings. In the lobby, a salvaged confession booth was intended as a hiding place for guilty parties. Raymaekers brought several clients here to celebrate a successful transaction or to introduce them to his material language before starting design negotiations. His clients were middle‑class professionals – teachers, dentists, entrepreneurs, doctors, engineers, florists, pharmacists. Wealthy, but not extraordinarily rich. Many of his houses allowed them to live and work – interfacing with the public – on the same property. House Kelchtermans from 1970, for example, designed with architect Jos Witters, is composed of three pyramids – a home, a doctor’s office and a garage – structured with reclaimed oak beams. The roof tiles were salvaged by the client himself, Dr Kelchtermans, who received tips on where to find them from his patients, many of whom were farmers whose buildings were undergoing modernisation or demolition. Inside, it is all split‑levels, enormous fireplaces and complex, spiralling spaces. But the house’s biggest flourish is the skylights: 23 cupolas from decommissioned Lockheed T‑33 fighter jet cockpits. Raymaekers had found them languishing in a salvage yard on the other side of Flanders, saw their architectural potential and persuaded the client to buy them. Client participation was everything. The Boncher family’s house, completed in 1984, is a mash‑up of components from a derelict slaughterhouse in the city of Tienen and an army barracks in Verviers. The Bonchers themselves spent weeks carefully dismantling the bricks, Gobertange white stone and fleur‑de‑lis roof tiles from the slaughterhouse. Raymaekers had the grand entrance and guardhouses from the barracks available in his stock at Queen of the South, probably at a time‑sensitive discount. Many of the construction and final design decisions for House Boncher were left to happenstance; as the contractors built the walls up row‑by‑row, whenever a hole was needed for a window, Raymaekers would decide the dimensions only at that moment, based on whatever blue limestone window frame he happened to have in stock. Occasionally, the masons had to deconstruct part of a wall if a larger‑than‑anticipated window became available.  ‘Raymaekers’ houses are intelligent, instinctive assemblages of unpredictable material streams’ The interior, too, required improvisation. A white stone staircase, salvaged from a church pulpit, was meant to spiral up to the main bedroom of House Boncher but turned out to be about 200mm too short. The solution? Raymaekers and the contractors decided to build a hefty bump into the floor, consisting of cobbles and decorative tiling. It gave the staircase the necessary boost to reach up to the bedroom. It was typical of how designing exclusively with old materials required Raymaekers to empower and trust his contractors. Their design contributions, extemporisation and management of materials and how they might fit together were intrinsic to the realisation of every project. Working with large batches helped simplify matters – or allowed further complexity. When the Antwerp townhouse of art nouveau architect Joseph Bascourt was demolished to make way for the extension of a car park, Raymaekers snapped up the facade. Without knowing or caring how the pieces were originally composed, Raymaekers shuffled them into a new configuration for the facade of a flower shop a client had commissioned in the Brussels hinterland, completed in 1987. Raymaekers moved Bascourt’s decorative entablature from the top of the facade down to eye level, the better to be enjoyed; that this meant the windows on the top floor would now poke strangely above the roofline was fine with both Raymaekers and the client. The shreds of Raymaekers’ reputation that survive today, and what can be retrieved and reconstructed of his private and professional habits, are complex at best. His charisma and determination enabled him to convince artisans, contractors, labourers and clients to stretch themselves beyond what they thought possible, adding their own talent to his difficult buildings. But Raymaekers could also be intimidating and domineering. Collaborators often stepped away from his all‑consuming process; clients who considered him too pushy and found themselves running out of money could cease working with him. But this was not the case for his spouses, sons, grandchildren and daughters‑in‑law. They were bound to him by more than his business practice and art, and some of them suffered enduring and even unbearable distress. Their suffering is the shadow behind Raymaekers’ work. When researchers from Belgian design practice Rotor and the University of Ghent – also the authors of this article – started showing up at Queen of the South in 2011, it was not because of Raymaekers’ reputation; it was merely to include Queen of the South on Opalis’s database, a roster of salvage dealers covering much of north‑western Europe. We only realised the extent and importance of his work after several years and at least three visits. Raymaekers had refused all interviews since his bankruptcy, but after some persuasion, a meeting was scheduled in the courtyard on a grey winter day. Triggered by newspaper cuttings, his own limited archive of photographs and strolls through the stock, a picture emerged of the rich reuse ecosystem in which he was active in his glory days. It is this ecosystem, along with the remarkable and necessary fluidity of Raymaekers’ practice with old building materials, that was critical to his success. Belgium was blessed in the postwar period with a remarkably robust network of demolition contractors who were also salvage dealers, antiques dealers who were also designers, industrial scrapyards willing to let architects pick through their mountains of waste. The landscape was ripe for reuse. So was the legislative field. Raymaekers’ oeuvre could be mistaken, at a cursory glance, for just more Ugly Belgian Houses (several of them have indeed appeared on Hannes Coudenys’ infamous blog). His houses appear to be kitsch, chaotic, brazen, overstuffed with mixed metaphors. But they are much more than that. They are intelligent, instinctive assemblages of unpredictable material streams. They are bracing, never boring. And though they are wildly out‑of‑step with today’s tastes, the houses – and moreover, the kind of innovative practice that built them – have a lot to teach spatial practitioners about reusing architectural materials and circularity in the construction industry. And as with any of those so‑called ‘ugly’ Belgian houses, the label is a trivialisation (and an elitist one) of something systemic and generative. What makes possible such expressionistic and characterful suburban houses in Belgium – whatever one thinks of their quality – is government policy since the postwar era. While other European countries went all in on standardised social housing, in Belgium, the 1948 De Taeye Law offered construction grants and a state guarantee on mortgages, triggering families to initiate, help design and sometimes even execute the construction of their own homes, tapping into a rural tradition of self‑reliance. Raymaekers worked at a time when modernisation – and the demolition it demanded – was churning out a constant flow of antique materials. It was also a time when the merger and exchange of roles in the architectural process – from material procurement, to design and collaboration live on site – was still possible. Material reuse today is much harder. It must work precisely against the linear flows of extraction, capital, efficiency and predictability – all supercharged by the digital and its requirement of an almost omniscient predictability, and by increasingly demanding rules around compliance and liability. But to explore the wild potential of material reuse – not just to reduce embodied carbon, but to unleash new design potential and a richer culture around material (heritage) and practice – it must be possible to imagine ways in which material procurement, construction sites and the architects’ role can change fundamentally again.  Illustration: Laslo Antal for The Architectural Review. Read more about the process of making this portrait here 2025-05-15 Justinien Tribillon Share AR May 2025CircularityBuy Now

John Edwards, Technology Journalist & AuthorMay 15, 20255 Min ReadJorge Pérez via Alamy Stock PhotoGetting a new IT project off the ground requires a concept, a goal, meticulous planning, and, perhaps most importantly, a willingness to assume a certain amount of risk. It's this last point that often keeps IT leaders awake at night. Addressing risk prior to launching an IT project is crucial, since it helps to set clear expectations, identify potential obstacles, and create a roadmap for handling unforeseen challenges, says Anbang Xu, founder of JoggAI, an AI-powered video platform, and a former senior product manager at Apple and senior software engineer at Google. "Particularly when it comes to AI-driven projects, understanding the risks -- whether they're related to technology, data security, scalability, or talent -- can make the difference between a smooth launch and a failed initiative," he observes in an email interview. "By addressing these concerns early, you also ensure that resources are allocated effectively, and your team is better equipped to handle surprises." Building a strong mechanism to address risk during project planning is crucial to ensure that all decisions are made with transparency into cost, benefit, and risk, says Ola Chowning, a partner with technology research and advisory firm ISG. "Project decisions that don't estimate risk and its tangible or intangible impacts at the start, or even during a project, likely don't have the data needed to make informed decisions and balance cost, risk, and benefit," she explains in an online interview. "More importantly, failing to address risk can waste time if an event occurs and the team is left scrambling to determine how to respond." Related:There are also secondary risks, including budget, resources, and technology challenges, says Munir Hafez, CTO of Independence Pet Holdings via email (Hafez was senior vice president and CIO at credit reporting firm TransUnion at the time of the interview.) "Each of these risks should be addressed before launching an IT project to ensure that the project is starting off on a solid foundation and will gain alignment with all stakeholders." Accurate Estimations By quantifying risks, project managers can objectively assess their potential likelihood and impact, allowing efficient resource allocation and the development of appropriate mitigation strategies, Hafez says. "This structured approach ensures that no significant risks are overlooked and that the project remains on track." The best way to estimate risk is to divide the project into small components and assess the uncertainty around each unit, Xu says. This process includes examining the project's technical feasibility, the team's ability to handle complexity, and any potential dependencies on third-party tools or services. Xu states he relies on a combination of historical data from similar past projects and predictive models -- particularly when it comes to AI and machine learning applications, where it's often unknown how algorithms will perform at scale. "This allows us to develop a more nuanced risk profile, instead of relying on assumptions," he explains. Related:Many organizations use a risk scoring tool, built on previous project experiences and common risk categories, to guide project teams on risk types for their particular project, Chowning says. "Some [organizations] even have statistics on how often risks have occurred in the past and the average impact on previous projects." Risk Tips If a risk estimate appears overly optimistic, the leader should reassess the situation and adjust the project plan accordingly, Hafez says. "This may involve consulting industry experts or revisiting lessons learned from previous internal projects to accurately reflect the risk." The biggest mistake is underestimating the impact of technical debt and the complexity of scaling solutions, Xu warns. "Leaders often focus too much on the initial development phase and don’t account for the ongoing maintenance and optimization required," he explains. "Risk should be measured not only by the initial launch, but also by the long-term implications of maintaining and improving the post-launch system." Related:Final Thoughts As a project progresses, existing risks can evolve and new risks may emerge, Hafez says. "Regularly reviewing and updating the risk management plan ensures that the project will remain resilient and adaptable to these changes." Fostering a culture of open communication and encouraging team members to report potential risks early can significantly enhance a project's ability to respond effectively, Hafez says. This proactive approach helps avoid the "watermelon effect," where performance metrics appear positive (green) on the surface, but reveal underlying problems (red) when examined more closely. Embrace flexibility, Xu advises. In tech, especially in fields such as AI, things are constantly evolving. "A leader’s ability to adapt and pivot when necessary is as critical as the ability to plan," he notes. That's why it's important to continue refining risk assessments throughout the project's life. It’s important for project team and business decision-makers to understand that risks aren't a bad thing, Chowning says. "The underlying objective of risk management is to understand and mitigate risk, and to be prepared if risk events occur." Making high project risk scores a "bad thing" will simply drive team leaders and members to underestimate, ignore, or even attempt to hide potential risk. 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

Iron Man was a resounding success when it opened in 2008, earning more than half a billion dollars and declaring in no uncertain terms that Marvel Studios—and the Marvel Cinematic Universe, as it later came to be known—had arrived. Although Marvel was also developing movies about Thor, Captain America, and Ant-Man at the time, Iron Man 2 was pushed to the front of the line and set for release in May 2010, just two years after the first film. The idea was to cash in on the surprise success of Iron Man and give the other movies a little more time to breathe and develop, with the ultimate goal of teaming all the characters up in The Avengers. But perhaps due to its rushed production schedule, studio growing pains, and the need to lay down breadcrumbs for the stories to come, Iron Man 2 was a shaky affair when it eventually came out. Not that this hurt its box office potential. It grossed more than the first film. Still, it did not escape the notice of critics and hardcore fans, resulting in a film that still regularly lands near the bottom of MCU movie rankings. Strangely, however, Marvel Studios has repeated many of the same mistakes it made on Iron Man 2 in later films while also taking some of the wrong lessons from the troubled production of the MCU’s first sequel. A Rushed Development With Iron Man 2 announced days after the release of Iron Man (even before the relative failure of The Incredible Hulk that same year), Marvel had to scramble to come up with a story and lock down its principal players. Robert Downey Jr., of course, was signed to reprise the role of Tony Stark at a much higher salary. Gwyneth Paltrow was also confirmed to return, as was director Jon Favreau (who would also once again play Stark security chief Happy Hogan). Also returning were Samuel L. Jackson and Clark Gregg as Nick Fury and Agent Phil Coulson, respectively, expanding the presence of S.H.I.E.L.D. that was briefly established in the first film. Downey and Favreau tapped writer/actor Justin Theroux to pen the screenplay, which would introduce villains Justin Hammer and Whiplash (actually a sort of hybrid of that character and the Crimson Dynamo) into the story after comic book arch-nemesis The Mandarin was nixed as both too fantastical and potentially racist. Also slated to make her debut in the movie: Black Widow (Scarlett Johansson), a nod to the impending creation of the Avengers but one of the many elements of a movie that was potentially piling up too many plot threads. Although initial reports suggested that at least part of the film would adapt the famous “Demon in a Bottle” storyline, in which Tony Stark becomes a full-blown alcoholic, that aspect of the narrative was toned down due to concerns over toy sales and younger viewers’ perception of Iron Man. The movie reportedly began filming without a finished script, something that would hamper future Marvel movies and lead the studio to make reshoots a permanent part of every film’s production schedule. This necessitated Theroux having to do daily rewrites on the set with Favreau and Downey, with the latter’s ad-libbing and improvisatory style of acting further muddying what was already a confusing and tonally uneven plot. New Faces, New Problems As mentioned, Downey, Paltrow, Favreau, Gregg, and Jackson (the latter after some torturous negotiations) all returned to their roles in Iron Man 2, but Terence Howard, who co-starred in Iron Man as Tony Stark’s best friend, Col. James “Rhodey” Rhodes, did not. Depending on who you ask, Howard was either too difficult or too expensive—or both—to keep around for the sequel. He was replaced by Don Cheadle, who’s played Rhodey (and become a fan favorite, at least until Secret Invasion retconned him as a Skrull) ever since. Ironically Marvel should have learned from this and the recasting of Bruce Banner (from Edward Norton to Mark Ruffalo) that it was okay to recast a major role once in a while. But the biggest mistake Marvel made in this regard was arguably announcing that T’Challa/Black Panther would not be recast after the tragic death of Chadwick Boseman, which left the MCU without one of its most potentially important superheroes (Marvel has since seemingly eased up on this, with Harrison Ford replacing the late William Hurt as “Thunderbolt” Ross and rumors circulating that a new T’Challa from a different universe will soon appear). Still, the casting of Iron Man 2’s villains, Justin Hammer and Ivan Vanko/Whiplash, set some precedents that would later haunt the MCU in a different way. Hammer was played by Sam Rockwell (an early candidate for Iron Man himself) and acquitted himself well, but his role was in some ways superfluous to a story already packed with new characters and themes and his function in the movie was not always clear (he’s also yet to reappear in the MCU). But it was Mickey Rourke being courted to play Whiplash that perhaps caused the most angst. With Rourke coming in hot off his Oscar-nominated, Golden Globe-winning comeback performance in The Wrestler, Marvel wanted to seize a prestige casting coup—a move similar to its hiring of recent Oscar winner Brie Larson years later to play Captain Marvel. It could be argued that neither Rourke nor Larson was ultimately the right fit for their roles, but luring award winners to a comic book universe was too tempting for Marvel to pass up. In retrospect, they might have even dodged a bullet when Joaquin Phoenix passed on playing Doctor Strange given the actor’s reportedly mercurial methods. Rourke allegedly made bizarre demands to Marvel on Iron Man 2. This included that Vanko have a pet bird and wear his hair in a bun. He also reportedly clashed with Favreau during filming. Later he would say publicly he was unhappy with the final product, claiming that a lot of his performance was left on the cutting room floor. Whiplash in the finished film is certainly a strange character and the first of many undercooked Marvel villains to come. But that didn’t stop the studio from pursuing other actors and filmmakers basking in the glow of a high-profile awards run. Too Much of Everything Whereas the first Iron Man was a sleek, crisply told origin story, Iron Man 2 was, narratively speaking, a mess. The film dealt with Tony Stark’s inability to reconcile his personal identity with his new public persona as Iron Man, his lingering issues with his father, Howard Stark (John Slattery), his burgeoning drinking problem (less visible in the finished film, but still there), and his discovery that the core element in his arc reactor was slowly poisoning him. Join our mailing list Get the best of Den of Geek delivered right to your inbox! And that was just Tony; the rest of the film introduced Black Widow and the beginnings of the Avengers, put Rhodey in his War Machine armor for the first time, expanded the role of Nick Fury and S.H.I.E.L.D., set out easter eggs for Thor and Captain America, and tried to balance the roles of two villains whose motivations and tones were not at all well-defined or complementary. It was, in short, a hodgepodge of ideas, plot points, and hints at future movies that all but dragged itself over the finish line without a coherent structure—a problem that would crop up again in the MCU. The franchise’s obsession with world-building at the expense of focus and logic reared its head in Avengers: Age of Ultron, which brought four new characters into an already crowded ensemble (Scarlet Witch, Quicksilver, Vision, and the title villain) while also planting seeds about Asgard, Wakanda, and the Infinity Stones (some of which, to be fair, did pay off later). Even the underrated Black Panther: Wakanda Forever, released years later, acted partly as a commercial for the still-unaired Ironheart TV series (ironically The Marvels worked in reverse, requiring audiences to watch two TV shows to understand some of what was happening). Part of the charm of the Marvel Cinematic Universe, of course, is watching out for those many easter eggs, post-credit scenes, and surprise cameos that remind us that it is indeed a universe, one that’s constantly evolving and expanding. But some of those universe-building elements have worked better than others over the years, and with Iron Man 2, it felt like sensory overload—not that that stopped Marvel from doing it again and again in the years since. Iron Man 2: The Gift That Keeps on Giving? Iron Man 2 was an even bigger success at the box office ($624 million) that its predecessor ($585 million), but it also cost more ($200 million as opposed to $140 million) and was less enthusiastically received by critics. It was a troubled, patchwork project from the start, but even as Marvel Studios went on to hit heights of success no one could have possibly imagined for a filmed comic book universe, the studio continually made the same mistakes that dogged the production of Iron Man 2 and kept it from being the sequel it could have been. In some cases those mistakes may have temporarily dinged the brand (Age of Ultron), but in the long run, their impact has not done permanent damage to the MCU or the vast majority of its output. Still, with the last few years being particularly rocky ones for the MCU in terms of quality and coherence, it might not hurt Kevin Feige and company to fire up Iron Man 2 once in a while and remember it as a cautionary tale of what not to do.

Are you building a cooperative multiplayer game with Unity? We’ve got the perfect on-demand resource for you: A four-part webinar series from the Multiplayer team – now available on YouTube. This series dives into the Boss Room sample game to explore how you can build a production-ready multiplayer game with Unity and Netcode for GameObjects.Creating multiplayer games isn’t just about using the right APIs. The first episode in the series covers best practices and provides advice when introducing a chaotic tech like the internet between your players when developing a multiplayer game.In the first episode, we cover:What is Netcode for GameObjects (NGO) and Unity Gaming Services (UGS)?NetworkVariables vs RPCs and reasons to choose which for each featureAuthority and why we chose server authority for most of our gameplayMovements and alternative models like Client Authority for dealing with lag, and optimization tricks for pathfinding-based movementsBasic game implementation and how we used State (netvar, network list) and RPCsEpisode two covers the implementation of character abilities in Boss Room and how to approach ability implementation in your game.We dive into:The general flow of implementing server-driven abilitiesReliability and what it means for RPCs to be unreliableMelee attacks and the anticipation of animation issuesTank shield functionality and interaction with the environmentArcher powershots and Area of Effect (AOE), which require tracking inputs over time on the client side onlyMage firebolt and bandwidth useThe general architecture of our ability systemWe also look at:Dynamically spawned objectsSpawning static in-scene placed objectsSpawning destroyable in-scene placed objectsLessons learned around zombie imps when late joiningCustom spawning logic like pools and how to integrate them with NGOIn the third session, we cover a few more game implementation details before diving into making your multiplayer game resilient to players.We explore:Physics interactions, objects and parenting, scenes and scene managementUsing Relay and Lobby to take your game online and why you can’t use NGO for joining playersCharacter selection screen syncing between players using NGO and custom serializationMaking your game resilient to players coming and going, and reconnection flows, late join support, disconnection, and timeoutsHandling race conditions and various tests to performIn the fourth and final session of the series, we explore game development best practices for multiplayer games.This episode covers:Bandwidth optimizations and tools to profile, optimizing NetworkTransforms, and other tipsDebugging tools like network condition simulators, stats monitoring, and custom debug flowsProject architecture (action system, player architecture, etc.)The future of Boss RoomWatch the series? Share your feedback so we can plan even better sessions in the future.If you’re ready to dive into your next multiplayer project with Unity and Netcode for GameObjects, here are some helpful resources to get you started:Learn more about Netcode for GameObjectsExplore our samples: 3D Boss Room, 2D Galactic Kittens, Bitesize SamplesRead the documentationGot questions or feedback about multiplayer game development in Unity? Feel free to join the discussion on Discord or in the forums.

The Pokémon Fossil Museum, a Japanese exhibition that pairs man-made Pokémon fossils with real-world ancient life forms, will make its international debut at Chicago’s Field Museum in 2026, the museum and The Pokémon Company International announced Wednesday. Younger fans who started their adventure in Paldea and OG Kanto fans alike can view real dinosaurs and ancient life forms alongside their Pokémon equivalents. Attendees will discover how Field Museum icons like SUE the T. Rex and the Chicago Archaeopteryx compare to Pokémon such as Tyrantrum and Archeops, while exploring how fossils are uncovered and studied in both the Pokémon universe and the real world. The exhibition features Pokémon “Professors,” the ever-adorable Excavator Pikachu complete with a cowboy hat, and Field scientists exploring fossils from both worlds. “Aligned with the Field Museum’s mission to connect us to the natural world, the Pokémon Fossil Museum will inspire visitors to make new discoveries about some of their favorite Pokémon while offering a pathway to science,” said Jaap Hoogstraten, head of exhibitions at the Field Museum, in a news release, “especially budding paleontologists.” You can see a Japanese-language tour of the Pokémon Fossil Museum exhibition in the video below for a sneak peek at what to expect. Created by Japan’s National Museum of Nature and Science and The Pokémon Company, Pokémon Fossil Museum is a traveling exhibition usually scheduled in museums throughout Japan. This year, Pokémon Fossil Museum is currently located at Fukui Prefectural Dinosaur Museum until May 25. It will then move to Nagasaki City’s Dinosaur Museum from June 7 until Sept. 21. Its last stop will be the Tokushima Prefectural Museum in October. Pokémon Fossil Museum will open at the Field Museum on May 22, 2026. We asked the Field Museum when tickets will go on sale, and a representative told Polygon, “We have not yet announced when tickets will be available. Anyone interested in being the first to buy tickets can sign up for email updates on our website.” If interest in the Pokémon Fossil Museum is anything like the fervor associated with the Pokémon exhibition at Amsterdam’s Van Gogh Museum, which prompted some very bad behavior when scalpers realized they could sell museum-exclusive merchandise to hardcore Pokémon fans, you’ll want to make a plan to visit sooner rather than later. Let’s hope the Field Museum and The Pokémon Company have learned a few lessons since then, because the Japanese versions of the exhibit have some extremely cool merch.