The studio admitted that artwork was used without permission in various assets found in Marathon's Closed Alpha.

Machine Learning Engineering Manager – MarketplaceRobloxSan Mateo, CA, United States2 hours agoApplyEvery day, tens of millions of people come to Roblox to explore, create, play, learn, and connect with friends in 3D immersive digital experiences– all created by our global community of developers and creators.At Roblox, we’re building the tools and platform that empower our community to bring any experience that they can imagine to life. Our vision is to reimagine the way people come together, from anywhere in the world, and on any device. We’re on a mission to connect a billion people with optimism and civility, and looking for amazing talent to help us get there.A career at Roblox means you’ll be working to shape the future of human interaction, solving unique technical challenges at scale, and helping to create safer, more civil shared experiences for everyone.Machine Learning Engineering Manager – MarketplaceAt Roblox, our Marketplace Search & Recommendation team powers the discovery of avatar content across the platform—from bodies, clothing, and accessories to fully styled outfits—spanning the Roblox app and immersive in-experience surfaces. We’re reimagining how discovery and personalization can unlock creativity and economic opportunity for millions of users and creators.We’re looking for a Machine Learning Engineering Manager to lead a team of exceptional ML engineers tackling core challenges across ranking, retrieval, and multimodal content understanding. This includes everything from advancing deep learning models for search and recommendation, to developing systems that understand the textual and visual semantics of avatar content and drive long-term value for the Roblox economy.You Will::Define and drive the technical roadmap for ML across Search, Recommendation, and Content Understanding.Lead the development of scalable, production-ready ML systems—from raw data to model deployment.Collaborate cross-functionally with product, design, and data science to build intelligent discovery experiences.Recruit, mentor, and grow a high-performing team of ML engineers.Establish engineering best practices to ensure system reliability, extensibility, and performance.You Have:Leadership Experience: 3+ years managing ML engineers or scientists; 5+ years building ML systems at scale in search, recommender systems, or ads.Strong Technical Foundation: Degree in Computer Science, Engineering, Math, Physics, or a related field—or equivalent hands-on experience.Depth in ML: Proven track record shipping deep learning models across ranking, retrieval, NLP, computer vision, or multimodal pipelines. Experience using transformer architectures and LLMs in production systems is a strong plus.End-to-End Execution: You’ve built ML systems from data ingestion to serving in production.Strategic Thinking: You’re comfortable balancing near-term impact with long-term innovation.Team-first Mindset: A collaborative leader who thrives on mentoring and enabling others to succeed.For roles that are based at our headquarters in San Mateo, CA: The starting base pay for this position is as shown below. The actual base pay is dependent upon a variety of job-related factors such as professional background, training, work experience, location, business needs and market demand. Therefore, in some circumstances, the actual salary could fall outside of this expected range. This pay range is subject to change and may be modified in the future. All full-time employees are also eligible for equity compensation and for benefits.Annual Salary Range$289,460 — $338,270 USDRoles that are based in our San Mateo, CA Headquarters are in-office Tuesday, Wednesday, and Thursday, with optional in-office on Monday and Friday (unless otherwise noted).Roblox provides equal employment opportunities to all employees and applicants for employment and prohibits discrimination and harassment of any type without regard to race, color, religion, age, sex, national origin, disability status, genetics, protected veteran status, sexual orientation, gender identity or expression, or any other characteristic protected by federal, state or local laws. Roblox also provides reasonable accommodations for all candidates during the interview process. Create Your Profile — Game companies can contact you with their relevant job openings. Apply

activisionByZack ZwiezenPublished2 hours agoWe may earn a commission from links on this page.Screenshot: Epic / Lucasfilm / KotakuSAG-AFTRA, the massive actors and media union with over 160,000 members, has filed an unfair labor practice charge with the National Labor Relations Board against Epic Games over its inclusion of an AI-powered Darth Vader in a recent Fortnite update. Suggested ReadingNintendo Switch 2 Price Is Set at $450 for Now, But Could Go Higher Share SubtitlesOffEnglishSuggested ReadingNintendo Switch 2 Price Is Set at $450 for Now, But Could Go Higher Share SubtitlesOffEnglishNintendo Switch 2 Price Is Set at $450 for Now, But Could Go HigherYou might have heard about the AI-powered Darth Vader that was added to Epic’s popular battle royale game last week as part of an ongoing Star Wars-themed season. Players could talk to this AI-powered Vader and ask him questions. The AI character’s voice was designed to mimic that of the late James Earl Jones and was done with his and his estate’s permission. Pretty quickly after he was added to the game, players got him to say slurs, swears, and other questionable stuff. Epic patched him up and seems to have gotten the Sith Lord under control. But now the developer is in hot water with one of the largest media unions in the country because it included this AI-powered Vader without first communicating or bargaining with SAG-AFTRA.On May 19, as reported by Eurogamer, SAG-AFTRA announced that it had filed an unfair labor practice charge against Epic Games and its subsidiary, Llama Productions. “We celebrate the right of our members and their estates to control the use of their digital replicas and welcome the use of new technologies to allow new generations to share in the enjoyment of those legacies and renowned roles,” said SAG-AFTRA in a statement. “However, we must protect our right to bargain terms and conditions around uses of voice that replace the work of our members, including those who previously did the work of matching Darth Vader’s iconic rhythm and tone in video games.”SAG-AFTRA says that Epic and Llama Productions “chose to replace the work of human performers with A.I.” and did so without “providing any notice of their intent to do this” and without allowing the union to bargain over fair terms.SAG-AFTRA voice actors are currently on strike against large video game companies including EA and Activision over various issues. The strike started in 2024. One of the big ones is more protection against AI copies, like Darth Vader in Fortnite.   .

Because strategy games tend to be so intricately designed, we often see developers try to create equally intricate ad monetization strategies. However, overcomplicating ad placements can actually harm potential revenue. Meanwhile, keeping placements simple makes them feel like a natural part of the user experience - enabling you to engage players without disrupting your game economy.In this article, we’ll get into our recommendations for rewarded video ad placements that work best in strategy games and tips on how to integrate them. But first, let’s talk about some common mistakes we see strategy game developers make when they first start monetizing.Top 3 rewarded video mistakes for strategy gamesMistake #1: “I’ll just place a rewarded video anywhere I want”Your first instinct may be to place rewarded video traffic drivers in many - sometimes random - places throughout the game, like inside the mini games. But more isn’t always better, and can even hurt the game economy. Instead, you want to start simply and turn to your core game loop. That’s because by adjusting your rewarded video placements to meet the basic needs of your players, you’re encouraging more engagement with your ads - ultimately increasing revenue.For example, to progress through a game, strategy players constantly need basic resources like coins, wood, or mana - but this currency comes and goes quickly throughout levels. By offering users to watch a rewarded video in exchange for these basic resources, they’re more likely to engage because they’ll get what they need. When you offer them outside the core loop, they have no need for the reward - so there’s less of a reason for them to engage.Another example? If the core loop dictates that users need to buy a new weapon for soft currency, then you can make coins a reward users receive.Mistake #2: “I’ll offer rare currency to all my players”It’s true that rewarded video placements should be available to all players, but the rewards should differ depending on the players’ level - are they beginner or advanced? For beginners, currency should mostly be easy to get. For advanced players, offer rare currency at higher levels in the game. Why? If you offer rare currency to all your players in the first few levels of the game, beginners may not know what to do with it - so they won’t engage with the rewarded video. That’s why saving rare currency for advanced players, and offering basic currency early in the game helps you capture a wider audience and increase engagement.Mistake #3: “I want to make additional rewarded video currency to reward players because it’s easier to develop”The easier it is for strategy players to reach the resource they need, the more they’ll continue engaging with rewarded videos. Often we see strategy games create an additional currency just for rewarded video, which players then need to cash in later to receive the game’s actual core currency. But this puts an extra step in between the users and what they really want: immediate rewards. Not only is a third currency an additional step for players, they’ll often forget about spending it, delaying their reward.In fact, one of our developer partners originally created movie tickets as an additional currency for rewarded video but then ended up switching their strategy. They kept the movie ticket placement but also added a new placement without movie tickets, so that the players could have the option to receive a direct reward of gold coins.The developer saw a huge spike in all KPIs with no harm to the game economy and IAP. The placement worked well because it eliminated the in-between currency (movie tickets) while giving direct access to the currency (gold coins) players wanted.Top 5 rewarded video placements for strategy gamesAs a general rule of thumb, rewarded video works best when it’s shown as often as possible. Start monetizing users starting on D0, but also focus monetization efforts on D7 and D30. The players who stay beyond these days are the ones that rewarded video can be most useful for. That’s because the more players progress in a strategy game, the more valuable the rewards will be for them to stay and succeed.With that in mind, let’s dive deeper into how to boost ad engagement with rewarded video placements.1. Home screenStrategy players spend most of their time on the home screen, toggling between different actions to take - so rewarded video here is potentially your top placement. Because this placement will be your most exposed, users have more opportunities to engage with it.Turn to your core loop to decide what rewards users will be looking for the most, but always A/B test so you know which rewards drive the most revenue and engagement. The most common rewards to offer on the home screen are basic resources, soft currency, and gachas, since players constantly need these to progress through the game.To increase engagement rates, you can also add a time limit to the home screen placement. This will give your players a sense of urgency - encouraging them to engage immediately. They’re more likely to tap the button to watch an ad because they’re afraid of missing out on the opportunity.2. In-game storeStrategy players who need hard currency are most likely to head to the in-game store. Since the majority of players don’t spend real money to earn rewards, offering a rewarded video here in return for valuable resources is a great opportunity. Generally, rewards in the store aren't specific items but still give value - like treasure chests or other gachas.To get users to keep coming back to the store, set the pace for rewarded videos to be between 4 to 6 hours. Use this as a starting point and then keep A/B testing to find the ideal pace for your game. Not only does this promote more traffic to the store, but done right, users might also make an in-app purchase down the line - a double win for you.3. Pre and post level boostersLevel-based games like match-3 often show rewarded videos before and after levels to engage users at critical moments - giving users the resources they need to progress. Though strategy games aren’t centered around levels in the same way, the ones that do have level-based gameplay, can also leverage this placement.For example, your pre-level booster can offer users to upgrade their characters or get additional resources they need to complete that level - essentially giving them a leg up before they start. Your post-level booster can extend play sessions when players fail (think: earning an additional life).Be sure to carefully limit this placement, as it can be risky if players are given too much help. If players can too easily boost their performance and pass levels a lot faster, they’ll get bored and leave the game. It’s about finding the right balance.4. Time boostFor strategy games where wait time for tasks increase depending on progress, a time boost placement will help them progress faster.When using this placement, make sure that it’s aligned with a player’s progression and level. For example, if a player needs to wait 20 minutes, cutting their time by 10 minutes would be substantial. But if a player needs to wait 3 hours, you’d want to adapt the time you’re taking off for them and cut their time by an hour or so. Again, make sure to A/B test the amount of time you’re reducing for players - since it’ll depend on the game.5. Daily reward multiplierFor games that have a daily reward, you can add a placement that multiplies it - giving users an extra reason to engage. During the first week and month of a player’s life in the game, daily rewards are really valuable. You’re essentially giving players a taste of what your game is like and are drawing them in. From the start, if players need 100 coins in a game, giving them 80 coins for free in the daily check-in and 20 through rewarded video is an easy way to motivate players to continue. It’ll help to A/B test these rewards to find the right ratio whether that’s 50/50, 80/20, or so on.Make sure this isn’t your only placement. Strategy players only get one shot per day to use it - meaning you can’t monetize players more than once per day.Learn more about ironSource’s game design consultancy. Get in touch.

Semiconductor giant AMD followed through with its plan to spin out ZT Systems’ server-manufacturing business. AMD announced on Monday that it was selling ZT Systems’ server-manufacturing business to electronic manufacturing services company Sanmina. The $3 billion deal is a mix of cash and stock: $2.25 billion in cash; a $300 million premium, including 50% cash and 50% equity; and a $450 million contingent payment based on financial performance over the next three years, according to Reuters. The deal is expected to close by the end of 2025, subject to regulatory approval. After this divestiture, AMD will maintain control of ZT Systems’ rack-scale AI solutions design business. This announcement isn’t a shock. When AMD announced its intent to acquire ZT Systems, an AI and cloud infrastructure company, for $4.9 billion in August 2024, the company said at the time that it planned to divest that part of ZT Systems’ business after the deal formally closed. AMD’s acquisition of ZT Systems officially closed in March 2025, according to Reuters. Alongside this announcement, AMD said that Sanmina will become a “preferred” new product introduction manufacturing partner for AMD cloud rack and cluster-scale AI solutions. ​​“By combining the deep experience of our AI systems design team with our new preferred NPI partnership with Sanmina, we expect to strengthen our U.S-based manufacturing capabilities for rack and cluster-scale AI systems and accelerate quality and time-to-market for our cloud customers,” said Forrest Norrod, executive vice president and general manager, data center solutions business unit at AMD, in the company’s announcement. TechCrunch reached out to AMD for comment. Techcrunch event Join us at TechCrunch Sessions: AI Secure your spot for our leading AI industry event with speakers from OpenAI, Anthropic, and Cohere. For a limited time, tickets are just $292 for an entire day of expert talks, workshops, and potent networking. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 REGISTER NOW

With its rebrand and product expansion, Raindrop is betting that the next generation of software observability will be AI-first by design.Read More

With its rebrand and product expansion, Raindrop is betting that the next generation of software observability will be AI-first by design.Read More

Google’s NotebookLM mobile app is out now for Android on Google Play and iOS and iPadOS via the App Store, Google Labs product manager Biao Wang says in a blog post. A mobile app has been “one of the most frequent requests” for the team. The app appears to offer similar functionality to the desktop version of NotebookLM, including the ability to upload sources of information that the app can summarize. It can also make AI-generated, podcast-like Audio Overviews. With the app, you can listen to those Audio Overviews in the background while you’re doing other things on your phone or while offline, Google says. Might be a handy way to get caught up on your performance review. The company teased the launch of the mobile app for the AI-powered tool last month. Its official arrival is happening just ahead of Google I/O, which kicks off with the opening keynote at 1PM ET on Tuesday. It’s probably going to be an AI show. Update, May 19th: The iOS app is now available, too.

"Des centaines de milliers de clients utilisent Microsoft 365 Copilot pour les aider à faire des recherches, réfléchir et élaborer des...

Le président américain, Donald Trump, aux côtés de la première dame, Melania Trump, signe la loi « Take It Down Act », dans la roseraie de la Maison Blanche à Washington, lundi 19 mai 2025. JIM WATSON / AFP Face à un phénomène en pleine croissance, le président américain, Donald Trump, a promulgué, lundi 19 mai, une loi criminalisant le partage sans consentement d’images pornographiques réelles ou créées par l’intelligence artificielle (IA), notamment à des fins de vengeance (le « revenge porn »). Lire aussi | Article réservé à nos abonnés Le nombre de « deepfakes » pornographiques explose du fait des progrès de l’intelligence artificielle Ce texte nommé « Take It Down Act », soutenu par la première dame, Melania Trump, avait été précédemment adopté par le Sénat américain et la Chambre des représentants, où il avait recueilli un large soutien des deux partis. « Il s’agira de la toute première loi fédérale visant à lutter contre la diffusion d’images explicites et fictives sans le consentement de la personne », s’est félicité le président républicain. « Quiconque diffusera intentionnellement des images explicites sans le consentement d’une personne encourra jusqu’à trois ans de prison », a-t-il détaillé, évoquant également des « responsabilités civiles » pour les réseaux sociaux qui hébergent ces images mais « refusent de retirer ces images rapidement ». Les jeunes filles et femmes particulièrement touchées La diffusion non consentie d’images à caractère sexuel, parfois créées grâce au concours de l’IA, est un phénomène qui touche majoritairement les jeunes filles et femmes – dont certaines célébrités comme la chanteuse Taylor Swift ou l’élue du parti démocrate, Alexandria Ocasio-Cortez. Lire aussi | Taylor Swift victime de fausses images pornographiques, l’intelligence artificielle en accusation Avec la popularisation des outils d’IA, la fabrication de ces montages photo ou vidéo hyperréalistes s’est simplifiée et a ouvert la voie à leur utilisation massive à des fins de harcèlement ou d’humiliation, prenant de court les législateurs du monde entier. « Cette législation constitue une étape décisive dans nos efforts pour garantir que chaque Américain, en particulier les jeunes, puisse se sentir mieux protégé contre les atteintes à leur image ou à leur identité par le biais d’images non consenties », a salué Melania Trump aux côtés de son époux, lors d’une rare apparition publique. La première dame, mobilisée sur ce dossier, avait pour l’occasion invité à la Maison Blanche plusieurs jeunes filles victimes de telles pratiques. Parmi elles, une adolescente texane qui avait découvert fin 2023 des montages hyperréalistes (deepfakes) à caractère sexuel d’elle et de plusieurs de ses camarades, fabriqués sans leur consentement par un élève de son collège puis partagés sur le réseau Snapchat. « Les filles ne faisaient que pleurer, et pleurer à ne plus en finir, elles avaient honte », avait à l’époque confié à l’Agence France-Presse (AFP) sa mère, Anna Berry McAdams, qui s’inquiétait que ces images ne resurgissent dans le futur. « Cela pourrait les affecter toute leur vie », alertait-elle, pointant du doigt l’impuissance des victimes à agir et faire défendre leurs droits. Lire aussi | Article réservé à nos abonnés « J’étais sous antidépresseurs, et lui simplement auditionné » : pour les victimes de « revenge porn », un long combat pour obtenir justice Risque de censure Seuls quelques Etats américains, dont la Californie et la Floride, disposaient jusqu’à présent de lois criminalisant la fabrication et la publication de fausses images à caractère sexuel. Cette loi fédérale représente donc un « pas significatif » dans la lutte contre ce phénomène, estime, auprès de l’AFP, Renée Cummings, criminologue et chercheuse en intelligence artificielle à l’université de Virginie. « Mais son efficacité dépendra de sa mise en œuvre rapide et fiable, de sanctions sévères à l’égard des auteurs de ces actes et de son adaptabilité en temps réel face aux menaces numériques émergentes », a-t-elle insisté. Et de prévenir : « Nous devons veiller à ce que les mesures législatives protègent la société et punissent les auteurs, sans pour autant éroder par inadvertance le cryptage ou étouffer l’expression légitime » sur Internet. Lire le décryptage | Article réservé à nos abonnés « Deepfake », un terme imparfait pour une réalité appelée à durer Un risque pointé par plusieurs associations de défense des libertés publiques, qui se sont inquiétées d’un risque de censure. « Si la protection des victimes face à ces invasions odieuses de l’intimité est un but légitime, les bonnes intentions seules ne suffisent pas à faire une bonne loi », a ainsi estimé l’Electronic Frontier Foundation, dénonçant des « définitions vagues et le manque de garde-fous » qui pourraient mener à la censure. Le Monde avec AFP Réutiliser ce contenu