Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More It was a big week for AI announcements following events from Microsoft, Google, and Anthropic. But OpenAI is finishing things out with news of its own. And no, we’re not just talking about its $6.5 billion acquisition of Jony Ive’s design team to lead a new hardware effort, “io” at OpenAI. Today, the company upgraded its Operator autonomous web browsing and cursor controlling agent within ChatGPT from using the prior GPT-4o multimodal large language model to the newer and more powerful o3 reasoning model. The update, released globally today, May 23, 2025, is available as a “research preview” to paying subscribers of OpenAI’s $200 USD-monthly ChatGPT Pro plan. Basically, that is OpenAI’s way of saying it’s not a fully “sanded down” or perfected product yet — it may still have kinks and issues. But with rival Google offering its own top tier AI subscription bundle for a price of nearly $250 USD regularly (currently running a discount down to $125 for the first three months) to access its latest Gemini multimodal, Imagen image generation, and Veo video generation models, suddenly OpenAI’s ChatGPT Pro plan seems more affordable by comparison. What is OpenAI’s Operator and what is it for? Operator first debuted in January 2025 as OpenAI’s initial step into semi-autonomous agents, specifically Computer Using Agents (CUAs). The idea is to go beyond the chatbot interface of ChatGPT and allow OpenAI’s powerful AI models to start taking more actions on behalf of the user. Thus, Operator was designed to autonomously point, click, scroll, and type to complete web-based tasks such as booking dinner reservations, compiling shopping lists, or ordering event tickets. This agentic capability allows it to complete user tasks directly through a browser interface, from booking reservations to gathering online data. For safety, privacy and security purposes, Operator didn’t use any existing web browser on a user’s PC or Mac. Instead, it ran in a cloud-hosted virtual browser accessible via a standalone site—operator.chatgpt.com—where users could input requests and observe the agent perform tasks in real time. It combined vision, reasoning, and interaction capabilities based on GPT-4o, marking a new direction for OpenAI in agentic AI. The product was launched as a research preview for ChatGPT Pro subscribers and featured built-in safety measures like user confirmations, Watch Mode, and restrictions on high-risk web platforms. It was also being tested in enterprise contexts, including travel planning and civic services, demonstrating its potential across both consumer and business environments. o3 offers improved accuracy, structure, and success rates With this update, OpenAI aims to enhance performance across several key dimensions. The new o3-based Operator demonstrates improved persistence and accuracy during browser interactions. In practical terms, this means it is more likely to complete user tasks successfully and with less need for correction or repetition. Moreover, users can expect responses that are clearer, more structured, and more comprehensive. In comparative evaluations, the new model shows a distinct preference advantage over its predecessor. Human preference studies reveal that users favor the o3 model for its style, comprehensiveness, and clarity. It also performs strongly in instruction following and efficiency, though results for factual correctness are more balanced between versions. Performance on third-party evaluation benchmarks reflects these enhancements. On the OSWorld benchmark that measures completion of browser-based tasks, the o3 model scores 42.9 compared to 38.1 for the previous version. However, OpenAI notes that due to limitations in the automated grading system, the actual performance gain could be closer to 20 percentage points! On WebArena, the new model achieved a score of 62.9, up from 48.1. The most dramatic improvement appears on the GAIA benchmark, where the o3 model scores 62.2, vastly surpassing the prior model’s 12.3. Side-by-side task comparisons further illustrate these gains. In one example involving a restaurant booking request, the new model provided a clearer and more detailed list of available reservations, including locations, Michelin ratings, and seating notes, presented in a well-formatted table. The previous version, while functional, delivered less information in a less organized manner, according to an image included with the new o3 Operator release notes: The o3 model also inherits the safety measures introduced with earlier versions, with further fine-tuning for its role as an agentic system. OpenAI has integrated enhanced training against harmful task execution, prompt injection vulnerabilities, and mistakes involving user intent. Evaluations show that the model now confirms 94% of sensitive actions before executing them, with 100% confirmation in financial transactions. Prompt injection susceptibility has also decreased from 23% to 20%. Notably, the o3 Operator maintains a cautious boundary on certain high-risk web interactions, such as email or financial platforms, where it may require user supervision via Watch Mode or explicitly refuse to proceed. These measures are part of a layered approach to safety that combines model-level robustness with real-time monitoring. While the upgrade to Operator marks a technical improvement, it also reflects OpenAI’s ongoing commitment to responsible AI deployment. The system’s ability to take real-world actions introduces new risks, and the development team continues to refine its safety protocols accordingly. According to OpenAI’s updated o3 system card documentation, the model remains below high-risk capability thresholds in categories such as biological and chemical misuse and has no native coding environment or terminal access, further reducing potential misuse vectors. Operator remains a research preview and is accessible only to ChatGPT Pro users. The Responses API version of Operator will continue to be based on the GPT-4o model, at least for now. Implications for enterprise technical decision-makers The upgraded Operator stands to significantly enhance the workflows of professionals in AI engineering, orchestration, data management, and IT security. For those building or maintaining machine learning models, the model’s improved accuracy and structured outputs reduce the overhead of test validation and troubleshooting. In orchestration contexts, it offers a practical, reliable tool for automating browser-based components of complex pipelines. Data engineers can delegate manual web interactions—such as data verification and scraping—with more confidence, freeing time for higher-level optimization work. Security professionals, meanwhile, gain a safer way to simulate user behavior in audits and incident response exercises, thanks to the model’s layered safety mechanisms. Across these disciplines, the o3-based Operator introduces both a capability upgrade and a risk mitigation framework, making it a practical addition to the modern technical toolkit. Daily insights on business use cases with VB Daily If you want to impress your boss, VB Daily has you covered. We give you the inside scoop on what companies are doing with generative AI, from regulatory shifts to practical deployments, so you can share insights for maximum ROI. Read our Privacy Policy Thanks for subscribing. Check out more VB newsletters here. An error occured.

Développeur Back-end Senior - Senior Backend developer / Equipe RicardoTurbulentMontreal QC CA1 minute agoApplyQue ferez-vous ?En tant que membre de l’équipe Ricardo de Turbulent, vous participerez à la modernisation de plateformes écrites en PHP. Vous contribuerez à l’adoption d’un plan d’architecture et à son développement, tout en soulevant les risques possibles de la solution proposée et en trouvant les mitigations nécessaires. Vous voulez repousser les limites de la technologie ainsi que devenir une ressource clé pour votre équipe.Collaborer de près avec le Product Owner pour l’élaboration de nouveaux projets avec le clientMettre en place des architectures efficaces et des solutions robustes grâce à un processus itératifSoulever et documenter les risques inhérents aux solutions proposées, trouver les mitigations nécessairesEffectuer l’estimation complète du travail à accomplir au niveau du développement backendMettre en place des architectures efficaces et des solutions robustes grâce à un processus itératifLivraison réussie de solutions backend dans le respect des normes élevées de qualité, de performance et de sécurité.Excellente gestion du stress et sens de la relation client pour évoluer efficacement dans un environnement exigeant et dynamique.Qu'attendons-nous de notre Développeur Back-end Senior ?10 ans+ d’expérience en développement webDiplôme en programmation informatique ou tout autre domaine connexeBonne maîtrise du cycle de production de projets web et multimédiaCommuniquer et collaborer étroitement avec des collègues de différentes expertises. Vulgariser des concepts complexes est naturel pour vous.Compétences avancées en programmation dans des langages backend tels que PHP, Node.js, TypeScript.Excellente expérience en conception et implémentations API RestConnaissance approfondie des bases de données relationnelles et non relationnelles, ainsi que des outils de gestion de bases de données (MySQL, Redis).Expérience avec les services Cloud (AWS ou autre)Bonne expérience de travail avec DockerSolide expérience dans la conception, le développement et la maintenance de systèmes backend robustes et évolutifs à volume élevé.Prise en charge les systèmes de compilation et de déploiement applicatif des projets qui vous sont assignés.Expérience avec des plateformes web telles que Symfony, Laravel ou autres frameworks similairesFamiliarité avec les engins d’indexation tels que Solr et ElasticSearchBonne connaissance de la méthodologie Agile – un atoutMaîtrise complète du français et bonne connaissance de l'anglaisÀ propos de Turbulent, une entreprise de Cloud ImperiumCIG Turbulent est à la pointe du développement de jeux, créant des fonctionnalités de jeu, des environnements, des personnages et des technologies pour l'univers de Star Citizen. Située à Montréal, notre équipe talentueuse est dédiée à la création d'expériences immersives et innovantes qui captivent les joueur·se·s du monde entier.Depuis sa création en 2002, Turbulent est un leader dans le développement de plateformes technologiques innovantes à travers divers secteurs, y compris le divertissement, les médias et l'éducation. En 2012, nous avons commencé notre collaboration avec Cloud Imperium Games (CIG) pour développer la plateforme web de Star Citizen. Ce partenariat a évolué et, en 2020, nous avons fondé un studio de développement de jeux qui travaille avec les équipes de développement nord-américaines de Star Citizen. Notre studio à part entière couvre la technologie en ligne, l'art, les environnements, l'interface utilisateur et la création de personnages. Le 13 juillet 2023, après plus d'une décennie de collaboration réussie, nous avons fièrement annoncé notre fusion avec Cloud Imperium Games.Cloud Imperium Games est reconnu pour créer des jeux vidéo révolutionnaires qui repoussent les limites du possible. Nos projets phares incluent Star Citizen, une simulation spatiale multijoueur record, et Squadron 42, une aventure épique en solo se déroulant dans le même univers.Pour plus d'informations sur la vie chez Cloud Imperium Games, cliquez ici 🎮Turbulent s'engage à employer une main-d'œuvre diversifiée. Les candidats qualifiés seront considérés sans égard à la race, la couleur, la religion, le sexe, l'origine nationale, l'âge, l'orientation sexuelle, l'identité de genre, l'expression de genre ou le handicap .------What will you do?As a member of Turbulent’s Core team Ricardo, you will take part in modernizing platforms written in PHP. You will contribute to the adoption and development of an architectural plan, while identifying potential risks in the proposed solution and finding appropriate mitigations. You’re driven to push technological boundaries and become a key asset to your team.Work closely with the Product Owner to design new projects with the clientImplement effective architectures and robust solutions through an iterative processIdentify and document risks inherent to proposed solutions, and determine appropriate mitigationsProvide accurate estimates for backend development tasksSuccessfully deliver backend solutions that meet high standards of quality, performance, and securityDemonstrate strong stress management and client relationship skills to thrive in a demanding and fast-paced environmentWhat do we expect from ou r Senior Backend developer ?10+ years of experience in web developmentDegree in Computer Programming or any related fieldStrong understanding of the web and multimedia project lifecycleAble to communicate and collaborate closely with colleagues from various disciplines — explaining complex concepts comes naturally to youAdvanced programming skills in backend languages such as PHP, Node.js, and TypeScriptExtensive experience in designing and implementing REST APIsIn-depth knowledge of both relational and non-relational databases, as well as database management tools (MySQL, Redis)Experience with Cloud services (AWS or others)Solid experience working with DockerProven track record in designing, developing, and maintaining robust and scalable high-volume backend systemsAbility to manage build and deployment systems for assigned projectsExperience with web frameworks such as Symfony, Laravel, or similarFamiliarity with indexing engines like Solr and ElasticsearchGood knowledge of Agile methodology – an assetFluent in French and good command of EnglishAbout Turbulent, a Cloud Imperium CompanyCIG Turbulent is at the cutting edge of game development, creating game features, environments, characters, and technology for the Star Citizen universe. Located in Montreal, our talented team is dedicated to crafting immersive and innovative experiences that captivate players around the globe.Since its inception in 2002, Turbulent has been a leader in developing innovative technological platforms across diverse sectors, including entertainment, media, and education. In 2012, we began our collaboration with Cloud Imperium Games (CIG) to develop the web platform for Star Citizen. This partnership evolved, and in 2020, we founded a game development studio that powers the North American development teams of Star Citizen. Our full-fledged studio reaches across online technology, art, environments, UI, and character creation. On July 13, 2023, after more than a decade of successful collaboration, we proudly announced our merger with Cloud Imperium Games.Cloud Imperium Games is renowned for creating cutting-edge video games that push the boundaries of what is possible. Our flagship projects include Star Citizen, a record-breaking multiplayer online space simulation, and Squadron 42, an epic single-player adventure set in the same expansive universe.For more information about life at Cloud Imperium Games, check this out 🎮Turbulent is committed to employing a diverse workforce. Qualified applicants will receive consideration without regard to race, color, religion, sex, national origin, age, sexual orientation, gender identity, gender expression, or disability. Create Your Profile — Game companies can contact you with their relevant job openings. Apply

Artificial intelligence keeps on taking the community at large by surprise, especially the large language models with the advanced capabilities they tend to offer, surpassing expectations. Tech giants like OpenAI are increasingly focused on bringing more efficient AI models and pushing the potential of the technology further to do many of the mundane tasks on consumers' behalf. While we have been hearing about the vast application of the tools in varied domains, a recent study evaluated how these models tend to perform on emotional intelligence, and the results are not something you would be expecting. A team of researchers conducted a study with findings suggesting that AI is capable of processing and understanding human emotions even better than humans Companies and researchers are increasingly invested in finding ways in which artificial intelligence can be used to bring more efficiency and different ways it can be used in institutions. OpenAI and many other tech giants are arduously working towards making their models feel more natural with capabilities like contextual understanding. A recent study has been conducted by University of Geneva and University of Bern researchers to find out about AI's empathetic capabilities. The study that has been published in Communications Psychology delivered some interesting findings that pointed towards generative AI models like ChatGPT not only demonstrating emotional intelligence but also outperforming humans in emotional intelligence tests. The study involved a series of tests with the six leading large language models, including ChatGPT-4, ChatGPT-o1, Gemini 1.5 Flash, Claude 3.5 Haiku, Copilot 365, and DeepSeek V3. The study further involved five emotional tests that were generally used in academics and professionally to see emotional understanding, regulation, and management. The situations presented were realistic and emotionally charged ones to see how the models would respond based on the emotional context provided. The results left the researchers baffled as all the LLM models outperformed the human participants significantly. The researchers even went a step ahead by asking ChatGPT-4 to create new EI test items, which were validated by human participants as well, and the results remarkably remained the same, with the AI models demonstrating a high level of contextual understanding. One of the Senior Researchers, Marcello Mortillaro,  had the following to say on the findings: LLMs are therefore not only capable of finding the best answer among the various available options, but also of generating new scenarios adapted to a desired context. This reinforces the idea that LLMs, such as ChatGPT, have emotional knowledge and can reason about emotions. These findings are vital, especially if we see how the technology is increasingly exceeding expectations in domains previously exclusive to humans. This could have great broader implications in terms of augmenting human skills in sensitive fields such as conflict management or coaching. Deal of the Day

Qualifications: Required: A minimum of 5 years of experience in capture management with a proven record of recent wins. In-depth knowledge of Army and/or Naval organizations and procurement behaviors. Strategic thinking skills to highlight unique differentiators in IT service delivery. Ability to translate past performance into compelling win strategies. Proficiency in project management with organizational and communication skills. Solid understanding of government contracts and contract vehicles. Extensive experience in the federal proposal process. Strong judgment in assessing bid/no-bid decisions. Preferred business development experience. Familiarity with CRM systems (e.g., Salesforce) and pipeline tools (e.g., TechnoMile) is advantageous. Desired: Training or certification in the Shipley Method for Proposal Writing & Capture. Bachelor's degree preferred; Master's degree is an advantage. Proven success in leading the capture of large federal contracts ($50M+), especially within DHA and DHS. Comprehensive knowledge of the federal acquisition lifecycle. Responsibilities: Lead efforts to secure major contracts ($50M+) within DoD organizations. Adhere to established business development and capture processes, ensuring compliance with Federal Procurement and Business Ethics Codes. Determine necessary resources for bids and estimate effort levels. Collaborate with various internal and external stakeholders (e.g., analysts, solution architects, proposal managers). Develop win themes and perform competitive analyses using available intelligence. Formulate price-to-win strategies in partnership with solution architects, program managers, and other team members. Implement best practices in capture management to improve success rates. Provide executive briefings on capture status and manage concurrent capture initiatives. Represent the organization externally and gather business intelligence to inform strategies. Cultivate positive relationships with internal teams to support seamless implementation of capture strategies. Build winning industry teams that address capability gaps and meet special requirements like small business utilization. Analyze performance requirements and competitor strengths to craft effective avoidance and mitigation strategies. Ensure the development of proposals that stand out and align with strategic goals. Maintain professional conduct and stewardship of resources while executing duties. Meet or exceed performance goals to contribute to overall organizational growth.Apply NowLet's start your dream job Apply now Meet JobCopilot: Your Personal AI Job HunterAutomatically Apply to Remote Full-Stack Programming JobsJust set your preferences and Job Copilot will do the rest-finding, filtering, and applying while you focus on what matters. Activate JobCopilot Executive 1st View company Jobs posted: 1 Tired of Applying to Jobs Manually?Let JobCopilot do it for you.No more spreadsheets. No more copy-pasting. Just set your preferences and let your Al copilot search, match, and apply to jobs while you sleep.Applies for jobs that actually match your skillsTailors your resume and cover letter automaticallyWorks 24/7-so you don't have to Activate JobCopilot

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More It was a big week for AI announcements following events from Microsoft, Google, and Anthropic. But OpenAI is finishing things out with news of its own. And no, we’re not just talking about its $6.5 billion acquisition of Jony Ive’s design team to lead a new hardware effort, “io” at OpenAI. Today, the company upgraded its Operator autonomous web browsing and cursor controlling agent within ChatGPT from using the prior GPT-4o multimodal large language model to the newer and more powerful o3 reasoning model. The update, released globally today, May 23, 2025, is available as a “research preview” to paying subscribers of OpenAI’s $200 USD-monthly ChatGPT Pro plan. Basically, that is OpenAI’s way of saying it’s not a fully “sanded down” or perfected product yet — it may still have kinks and issues. But with rival Google offering its own top tier AI subscription bundle for a price of nearly $250 USD regularly (currently running a discount down to $125 for the first three months) to access its latest Gemini multimodal, Imagen image generation, and Veo video generation models, suddenly OpenAI’s ChatGPT Pro plan seems more affordable by comparison. What is OpenAI’s Operator and what is it for? Operator first debuted in January 2025 as OpenAI’s initial step into semi-autonomous agents, specifically Computer Using Agents (CUAs). The idea is to go beyond the chatbot interface of ChatGPT and allow OpenAI’s powerful AI models to start taking more actions on behalf of the user. Thus, Operator was designed to autonomously point, click, scroll, and type to complete web-based tasks such as booking dinner reservations, compiling shopping lists, or ordering event tickets. This agentic capability allows it to complete user tasks directly through a browser interface, from booking reservations to gathering online data. For safety, privacy and security purposes, Operator didn’t use any existing web browser on a user’s PC or Mac. Instead, it ran in a cloud-hosted virtual browser accessible via a standalone site—operator.chatgpt.com—where users could input requests and observe the agent perform tasks in real time. It combined vision, reasoning, and interaction capabilities based on GPT-4o, marking a new direction for OpenAI in agentic AI. The product was launched as a research preview for ChatGPT Pro subscribers and featured built-in safety measures like user confirmations, Watch Mode, and restrictions on high-risk web platforms. It was also being tested in enterprise contexts, including travel planning and civic services, demonstrating its potential across both consumer and business environments. o3 offers improved accuracy, structure, and success rates With this update, OpenAI aims to enhance performance across several key dimensions. The new o3-based Operator demonstrates improved persistence and accuracy during browser interactions. In practical terms, this means it is more likely to complete user tasks successfully and with less need for correction or repetition. Moreover, users can expect responses that are clearer, more structured, and more comprehensive. In comparative evaluations, the new model shows a distinct preference advantage over its predecessor. Human preference studies reveal that users favor the o3 model for its style, comprehensiveness, and clarity. It also performs strongly in instruction following and efficiency, though results for factual correctness are more balanced between versions. Performance on third-party evaluation benchmarks reflects these enhancements. On the OSWorld benchmark that measures completion of browser-based tasks, the o3 model scores 42.9 compared to 38.1 for the previous version. However, OpenAI notes that due to limitations in the automated grading system, the actual performance gain could be closer to 20 percentage points! On WebArena, the new model achieved a score of 62.9, up from 48.1. The most dramatic improvement appears on the GAIA benchmark, where the o3 model scores 62.2, vastly surpassing the prior model’s 12.3. Side-by-side task comparisons further illustrate these gains. In one example involving a restaurant booking request, the new model provided a clearer and more detailed list of available reservations, including locations, Michelin ratings, and seating notes, presented in a well-formatted table. The previous version, while functional, delivered less information in a less organized manner, according to an image included with the new o3 Operator release notes: The o3 model also inherits the safety measures introduced with earlier versions, with further fine-tuning for its role as an agentic system. OpenAI has integrated enhanced training against harmful task execution, prompt injection vulnerabilities, and mistakes involving user intent. Evaluations show that the model now confirms 94% of sensitive actions before executing them, with 100% confirmation in financial transactions. Prompt injection susceptibility has also decreased from 23% to 20%. Notably, the o3 Operator maintains a cautious boundary on certain high-risk web interactions, such as email or financial platforms, where it may require user supervision via Watch Mode or explicitly refuse to proceed. These measures are part of a layered approach to safety that combines model-level robustness with real-time monitoring. While the upgrade to Operator marks a technical improvement, it also reflects OpenAI’s ongoing commitment to responsible AI deployment. The system’s ability to take real-world actions introduces new risks, and the development team continues to refine its safety protocols accordingly. According to OpenAI’s updated o3 system card documentation, the model remains below high-risk capability thresholds in categories such as biological and chemical misuse and has no native coding environment or terminal access, further reducing potential misuse vectors. Operator remains a research preview and is accessible only to ChatGPT Pro users. The Responses API version of Operator will continue to be based on the GPT-4o model, at least for now. Implications for enterprise technical decision-makers The upgraded Operator stands to significantly enhance the workflows of professionals in AI engineering, orchestration, data management, and IT security. For those building or maintaining machine learning models, the model’s improved accuracy and structured outputs reduce the overhead of test validation and troubleshooting. In orchestration contexts, it offers a practical, reliable tool for automating browser-based components of complex pipelines. Data engineers can delegate manual web interactions—such as data verification and scraping—with more confidence, freeing time for higher-level optimization work. Security professionals, meanwhile, gain a safer way to simulate user behavior in audits and incident response exercises, thanks to the model’s layered safety mechanisms. Across these disciplines, the o3-based Operator introduces both a capability upgrade and a risk mitigation framework, making it a practical addition to the modern technical toolkit. Daily insights on business use cases with VB Daily If you want to impress your boss, VB Daily has you covered. We give you the inside scoop on what companies are doing with generative AI, from regulatory shifts to practical deployments, so you can share insights for maximum ROI. Read our Privacy Policy Thanks for subscribing. Check out more VB newsletters here. An error occured.

It’s been two years since we’ve hosted a Unite event to gather with our amazing community of game developers. Thanks to you, this year’s Unite is on track to become our most attended ever. If you haven’t had the chance to watch the keynote yet, we’ve got you covered with this spotlight on what was shared.“When we say the world is a better place with more creators in it, we mean that literally,” shared CEO John Riccitiello. “Because creators bring people together. Because we know the joy of creating, and we want more people to experience this joy.”“Now, we know creation is hard,” he continued. “Game creation is especially hard because it involves so many disciplines – from programming to design, art to lighting, character creation to animation, and some serious mastery of technology. Our role at Unity is to make the tools that make it easier for you, as creators, to realize your vision.”Bringing our community together to offer updates and insights into what we’re working on is so important to us. Today’s hybrid event – both virtual and in person within our local communities in Austin, Brighton, Copenhagen, Montréal, and San Francisco – is also an opportunity for us to listen to your feedback so we can ensure we’re building tools and features that meet your needs.Watch the full keynote, or keep reading for a recap of highlights from key areas of focus as we continue to heavily invest in our game engine to help you build, scale, and grow incredible games.DOTS, an acronym for Data-Oriented Technology Stack, is composed of three parts: Entity Component System (or ECS for short), Burst compiler, and the C# Job System. If your game requires massive scale, DOTS is a great alternative to an object-oriented coding framework because it lets you write safe, multithreaded code that can deliver performance gains. Simply put, DOTS enables you to build vast, complex worlds and deliver incredible experiences to your players.As Laurent Gibert, who oversees the development of DOTS and multiplayer technology at Unity, announced, “ECS is finally leaving its experimental phase. That means that with the coming 2022.2 Tech Stream – like any other feature – ECS will be supported for your projects.”Gibert emphasized that you don’t have to choose between data- and object-oriented development: The two can coexist. This means you can solve hard scaling problems while relying on your traditional workflows for everything else.To help you scale up your DOTS skill set more quickly, ECS will include graphics packages, physics capabilities, and Netcode, as well as full documentation, samples, and tutorials.“ECS is finally leaving its experimental phase. That means that with the coming 2022.2 Tech Stream – like any other feature – ECS will be supported for your projects.”Creating great visuals is time consuming, so our team spotlighted a few updates to help you achieve your vision more efficiently in both Universal Render Pipeline (URP) and High Definition Render Pipeline (HDRP).“We know that it’s hard to choose a rendering pipeline early in a project,” said Nancy LaRue, a marketing manager for technical artistry tools. “And it can be complicated to switch later on. So we’re working on modifying the render workflow so you’ll be able to use both URP and HDRP in the same project.”The segment highlighted some new tools for URP, including Forward+ rendering and a new Decals offering, that achieve parity and even eclipse what you could achieve with the Built-in Render Pipeline. To highlight URP tools’ effectiveness in production, Hannah Kennedy, art director from Obsidian Entertainment, took to the stage to talk about the studio’s new game Pentiment, launching November 15. Hannah’s team chose Unity and URP to deliver on their project because they needed a flexible approach, right from the start.“We were also able to use URP’s wide range of direct and indirect lighting solutions to quickly achieve the mood and enhance the dramatization of the game,” she said.Next, a demo illustrated an HDRP workflow to help you create dynamic environments using Physically Based Sky, Cloud Layers, Volumetric Fog, Adaptive Probe Volumes, the Volume System, and our incredible new Water System.We closed the graphics segment by celebrating some of the inspiring and jaw-droppingly gorgeous games that you’ve brought to life with Unity over the last few months. Catch the full sizzle below.“We’re working toward an end-to-end solution that covers every aspect of your multiplayer game – from creation to ongoing management of a live experience,” explained senior technical product manager Kiki Saintonge. She also shared what’s upcoming for multiplayer services.Developing and operating multiplayer games is incredibly challenging. We want to make it easier for you to design innovative multiplayer experiences, so we’re improving our Netcode solutions and bringing services like matchmaking and game hosting right into the Editor.To showcase what’s possible, Timothy “Timo” Vanherberghen, founder and CEO of Triangle Factory, joined us onstage. He spoke about the studio’s multiplayer VR games, Hyper Dash and Breachers, both of which leverage our Matchmaker and Game Server Hosting solutions.Timo had this to say about why his studio chose to select services from UGS: “We could have built a solution ourselves, but that would mean spending a lot of time building and maintaining features that have nothing to do with our core gameplay. We’re in this industry to build games, not the systems around them.”Triangle Factory’s next game, Breachers, will be in open alpha later this month.“We’re in this industry to build games, not the systems around them.”Today’s keynote also gave a preview of how we’re simplifying VR development with the XR Interaction Toolkit, including an example from Vinci Games in Blacktop Hoops. We shared the ways we’re making it easier for you to customize your Editor with updates to the UI Toolkit, which is now at full feature parity with IMGUI. The keynote even provided insight into how you can continuously fine-tune and test your live game content from a single platform to boost player satisfaction and reduce churn.Next, we announced the arrival of DirectX 12 in the upcoming Tech Stream 2022.2, as well as the work we’re doing to make our suite of art and VFX tools – including our digital human package, hair system, and puppet-based animation – widely available to you in real-time, so you can use these tools in your games.Unite 2022 continues for the remainder of today, with technical sessions, roadmap deep dives, and more. This content will also be available on-demand later this year, and you can watch the full keynote address below. Last but not least, we hope you’ll join us tomorrow for Unity for Humanity Summit 2022, an inspiring look at how creators are using real-time 3D technology to make the world a better place.

Artificial intelligence (AI) relies on vast amounts of data. Enterprises that take on AI projects, especially for large language models (LLMs) and generative AI (GenAI), need to capture large volumes of data for model training as well as to store outputs from AI-enabled systems. That data, however, is unlikely to be in a single system or location. Customers will draw on multiple data sources, including structured data in databases and often unstructured data. Some of these information sources will be on-premises and others in the cloud. To deal with AI’s hunger for data, system architects need to look at storage across storage area networks (SAN), network attached storage (NAS), and, potentially, object storage. In this article, we look at the pros and cons of block, file and object storage for AI projects and the challenges of finding the right blend for organisations. The current generation of AI projects are rarely, if ever, characterised by a single source of data. Instead, generative AI models draw on a wide range of data, much of it unstructured. This includes documents, images, audio and video and computer code, to name a few. Everything about generative AI is about understanding relationships. You have the source data still in your unstructured data, either file or object, and your vectorised data sitting on block Patrick Smith, Pure Storage When it comes to training LLMs, the more data sources the better. But, at the same time, enterprises link LLMs to their own data sources, either directly or through retrieval augmented generation (RAG) that improves the accuracy and relevance of results. That data might be documents but can include enterprise applications that hold data in a relational database. “A lot of AI is driven by unstructured data, so applications point at files, images, video, audio – all unstructured data,” says Patrick Smith, field chief technology officer EMEA at storage supplier Pure Storage. “But people also look at their production datasets and want to tie them to their generative AI projects.” This, he adds, includes adding vectorisation to databases, which is commonly supported by the main relational database suppliers, such as Oracle. For system architects who support AI projects, this raises the question of where best to store data. The simplest option would be to leave data sources as they are, but this is not always possible. This could be because data needs further processing, the AI application needs to be isolated from production systems, or current storage systems lack the throughput the AI application requires. In addition, vectorisation usually leads to large increases in data volumes – a 10 times increase is not untypical – and this puts more demands on production storage. This means that storage needs to be flexible and needs to be able to scale, and AI project data handling requirements differ during each stage. Training demands large volumes of raw data, inference – running the model in production – might not require as much data but needs higher throughput and minimal latency. Enterprises tend to keep the bulk of their unstructured data on file access NAS storage. NAS has the advantages of being relatively low cost and easier to manage and scale than alternatives such as direct-attached storage (DAS) or block access SAN storage. Structured data is more likely to be block storage. Usually this will be on a SAN, although direct attached storage might be sufficient for smaller AI projects. Here, achieving the best performance – in terms of IOPS and throughput from the storage array – offsets the greater complexity of NAS. Enterprise production systems, such as enterprise resource planning (ERP) and customer relationship management (CRM), will use SAN or DAS to store their data in database files. So, in practice, data for AI is likely to be drawn data from SAN and NAS environments. “AI data can be stored either in NAS or SAN. It’s all about the way the AI tools want or need to access the data,” says Bruce Kornfeld, chief product officer at StorMagic. “You can store AI data on a SAN, but AI tools won’t typically read the blocks. They’ll use a type of file access protocol to get to the block data.” It is not necessarily the case that one protocol will better than the other. It depends very much on the nature of the data sources and on the output of the AI system For a primarily document or image-based AI system, NAS might be fast enough. For an application such as autonomous driving or surveillance, systems might use a SAN or even high-speed local storage. Again, data architects will also need to distinguish between training and inference phases of their projects and consider whether the overhead of moving data between storage systems outweighs performance benefits, especially in training. This has led some organisations to look at object storage as a way of unifying data sources for AI. Object storage is increasingly in use with enterprises, and not just in cloud storage – on-premise object stores are gaining market share too. Object has some advantages for AI, not least its flat structure and global name space, (relatively) low management overheads, ease of expansion and low cost. Performance, however, has not been a strength for object storage. This has tended to make it more suited to tasks such as archiving than applications that demand low latency and high levels of data throughput. Suppliers are working to close the performance gap, however. Pure Storage and NetApp sell storage systems that can handle file and object and, in some cases, block too. These include Pure’s FlashBlade, and hardware that runs NetApp’s OnTap storage operating system. These technologies give storage managers the flexibility to use the best data formats, without creating silos tied to specific hardware. Others, such as Hammerspace, with its Hyperscale NAS, aim to squeeze additional performance out of equipment that runs the network file system (NFS). This, they argue, prevents bottlenecks where storage fails to keep up with data-hungry graphics processing units (GPUs). But until better-performing object storage systems become more widely available, or more enterprises move to universal storage platforms, AI is likely to use NAS, SAN, object and even DAS in combination. That said, the balance between the elements is likely to change during the lifetime of an AI project, and as AI tools and their applications evolve. At Pure, Smith has seen requests for new hardware for unstructured data, while block and vector database requirements are being met for most customers on existing hardware. “Everything about generative AI is about understanding relationships,” he says. “You have the source data still in your unstructured data, either file or object, and your vectorised data sitting on block.” Read more about AI and storage Storage technology explained: AI and data storage: In this guide, we examine the data storage needs of artificial intelligence, the demands it places on data storage, the suitability of cloud and object storage for AI, and key AI storage products. Storage technology explained: Vector databases at the core of AI: We look at the use of vector data in AI and how vector databases work, plus vector embedding, the challenges for storage of vector data and the key suppliers of vector database products.

Venture capitalists have always focused on investing in companies that leverage technology to either disrupt established industries or create entirely new business categories. But some VCs are starting to flip the script on their investing styles. Rather than funding startups, they are acquiring mature businesses –such as call centers, accounting firms, and other professional service firms—and optimizing them with artificial intelligence to serve more customers through automation. This strategy, often likened to private equity roll-ups, is being employed by firms such as General Catalyst, Thrive Capital, and solo VC Elad Gil. General Catalyst, touting this as a new asset class, has already backed seven such companies, including Long Lake, a startup that scoops up homeowners’ associations in an effort to make the management of communities more streamlined. Since its founding less than two years ago, Long Lake has secured $670 million in funding, according to PitchBook data. While the strategy is still new, a few other venture outfits have told TechCrunch that they are also considering trying out the investment model. Among them is Khosla Ventures, a firm known for making early bets on risky, unproven technologies with long development timelines. “I think we’ll look at a few of these types of opportunities,” Samir Kaul, general partner at Khosla Ventures, told TechCrunch. Interestingly, this PE-flavored approach could be a surprising benefit to the multitudes of AI startups VCs are backing. If a VC marries old businesses with new technology, AI startups wanting to serve these industries would essentially gain instant access to large, established clients. 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 According to Kaul, such access would be helpful when new startups have difficulties securing customers on their own. With the rapid rate of change in AI, the number of startups pouring into the market, and the historically long sales cycles involved in selling to enterprises, such difficulties apply to many AI startups.  But Khosla Ventures wants to proceed with caution. “The companies we’re looking at are very unlikely to lose money,” Kaul said, but he doesn’t want the strategy to ruin the firm’s strong return track record. “My biggest stress in life is I’m managing other people’s money, and I want to make sure that I continue to be a good steward of it.” While Khosla Ventures is starting to “dabble” in AI roll-up investments, Kaul explained that the firm wants to do a few deals to assess if such investments deliver strong returns for the firm before possibly raising money for some kind of vehicle specifically aimed at this investment strategy. If early bets pan out, Khosla would likely partner with a PE-style firm to help it with acquisitions rather than hire a team. “We wouldn’t do it alone, we don’t have that expertise,” he said.

Venture capitalists have always focused on investing in companies that leverage technology to either disrupt established industries or create entirely new business categories. But some VCs are starting to flip the script on their investing styles. Rather than funding startups, they are acquiring mature businesses –such as call centers, accounting firms, and other professional service firms—and optimizing them with artificial intelligence to serve more customers through automation. This strategy, often likened to private equity roll-ups, is being employed by firms such as General Catalyst, Thrive Capital, and solo VC Elad Gil. General Catalyst, touting this as a new asset class, has already backed seven such companies, including Long Lake, a startup that scoops up homeowners’ associations in an effort to make the management of communities more streamlined. Since its founding less than two years ago, Long Lake has secured $670 million in funding, according to PitchBook data. While the strategy is still new, a few other venture outfits have told TechCrunch that they are also considering trying out the investment model. Among them is Khosla Ventures, a firm known for making early bets on risky, unproven technologies with long development timelines. “I think we’ll look at a few of these types of opportunities,” Samir Kaul, general partner at Khosla Ventures, told TechCrunch. Interestingly, this PE-flavored approach could be a surprising benefit to the multitudes of AI startups VCs are backing. If a VC marries old businesses with new technology, AI startups wanting to serve these industries would essentially gain instant access to large, established clients. 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 According to Kaul, such access would be helpful when new startups have difficulties securing customers on their own. With the rapid rate of change in AI, the number of startups pouring into the market, and the historically long sales cycles involved in selling to enterprises, such difficulties apply to many AI startups.  But Khosla Ventures wants to proceed with caution. “The companies we’re looking at are very unlikely to lose money,” Kaul said, but he doesn’t want the strategy to ruin the firm’s strong return track record. “My biggest stress in life is I’m managing other people’s money, and I want to make sure that I continue to be a good steward of it.” While Khosla Ventures is starting to “dabble” in AI roll-up investments, Kaul explained that the firm wants to do a few deals to assess if such investments deliver strong returns for the firm before possibly raising money for some kind of vehicle specifically aimed at this investment strategy. If early bets pan out, Khosla would likely partner with a PE-style firm to help it with acquisitions rather than hire a team. “We wouldn’t do it alone, we don’t have that expertise,” he said.