Organizers for this year’s U.S. Pavilion at the 2025 Venice Architecture Biennale have shared new renderings detailing the presentation PORCH: An Architecture of Generosity and its exploration of the typology common to the architecture of the American South.  The exhibition was organized by the Fay Jones School of Architecture and Design at the University of Arkansas in partnership with DesignConnects and the Crystal Bridges Museum of American Art. A series of 54 individual 'porch windows' cabinets will showcase architectural designs, models, and other artifacts that allow visitors to explore their various meanings and expanded definitions in different settings across the country.  Rendering by Luxigon. Courtesy the Co-Commissioners of the U.S.PavilionThe New York Times called it an "eye-catching merger of the vernacular and the global" in a recent preview. Susan Chin, one of three co-commissioners for the U.S. Pavilion, told the outlet: "It’s about people trusting one another, and ...

Season 2 of HBO’s The Last of Us adaptation kicked off on Sunday with a long-awaited premiere, and for the next month and a half, we’re going to get new episodes every week. Though Season 2 is not expected to cover the entirety of The Last of Us Part 2, it’s still looking locked on to be a packed batch of episodes, and HBO has released a new trailer to offer a glimpse at what the weeks ahead will bring.  From tense encounters with different types of Infected and plenty of personal drama between Ellie, Joel, and others, to explosive conflict between the WLF and the Seraphites in Seattle and more, there’s plenty coming up that fans of the source material will be familiar with. As the season’s first episode indicated, however, it looks like the show will continue to expand on The Last of Us Part 2’s content with additions of its own. Check out the trailer below.  Season 2 of The Last of Us has six more episodes left in its ongoing second season, with the next set to premiere on April 20. A third season has also already been officially greenlit.

Xaba, a startup building synthetic brains for industrial robots with zero code, announced it has secured a $6 million seed investment led by Hitachi Ventures. The Toronto startup said the extension to its seed round will accelerate the deployment of AI-powered robotics and cognitive industrial control systems. Hitachi Ventures led the round using funds from its new $400 million fund, with participation from Hazelview Ventures, BDC, Exposition Ventures, and Impact Venture Capital. Xaba is pioneering the application of industrial artificial intelligence (AI) to remake manufacturing processes. Its flagship product, xCognition, empowers industrial robots and collaborative robots (cobots) with AI-driven cognition and awareness, enabling them to autonomously generate programs and execute complex tasks such as welding, drilling, assembling, and additive manufacturing. By integrating real-time intelligence into automation, Xaba’s solutions significantly reduce deployment costs and enhance the quality, consistency, and flexibility of manufacturing operations. The company said it is taking aim at a $9 trillion opportunity. The shortage of skilled robotics programmers and control engineers creates even more challenges for companies to scale automation effectively. Massimiliano Moruzzi, CEO of Xaba, said in an interview with GamesBeat that industrial automation remains highly inefficient, relying on outdated controllers, rigid programming, and extensive manual intervention. Programming and deploying industrial robots alone cost the industry $7 billion annually, with 80% of automation costs stemming from manually developing logic for industrial controllers.  “Our vision is to disrupt the giants. What we’re developing is what I call a synthetic brain for information, or cognitive control,” said Moruzzi. Similar to what Open AI is doing for natural language commands for AI, Moruzzi said that Xaba is remaking factory language so that it can enable better automation, with the result being not only better robots for industrial purposes but also better human supervision and human support. Xaba’s generative industrial AI equips machines with cognitive intelligence, allowing them to autonomously adapt, optimize, and execute tasks with precision. At its core is xCognition, which acts as a kind of “Open AI for Industrial Automation” — fully automating industrial robotics and machine programming for any task while automatically generating both part-programs and all the programmable logic controller (PLC) machine logic required to bring any machine to life. In essence, this is automation driven by self-programming robots that can easily transition from “text to action,” Moruzzi said. “Traditional robotics systems require extensive programming, constant human supervision, and struggle with real-world variability, in geometry, process parameters, materials, and actual production KPIs,” said Moruzzi. “We’re redefining automation by enabling robots and machines to self-optimize and execute complex tasks with minimal programming. The result is a dramatic reduction in waste and up to a 10x reduction in costs.” Meet Xaba: The autonomous AI control system for industrial automation With Xaba, manufacturers can simply describe automation goals, production KPIs, or operational tasks in human-readable text or functional specifications. From there, xCognition and PLCfy autonomously generate the required code, enabling robots and production lines to operate independently with real-time adaptability. Digital twins are meant to perfect designs of factories before they are built in the physical world. But Moruzzi said that the concept ought to be renamed “automated reality.” he said that industrial managers need to have a machine that can synthesize the experience of a human and transfer that to a robot. “At Xaba, we are developing foundational AI for automation, which means that, for example, my synthetic brain captures the physical, electromechanical model of the machine. Why am I doing this? Because experience is not inside to the encyclopedia that Open AI is accessing in order to transform that text into the actual. What I’m doing is technology called data ontology. Data ontology is going to be the next big wave in AI in order to transform weak AI to strong AI. What is data ontology? He said data ontology is his own segment within neuroscience data. “It has the capacity to do what at the moment only the human can do, which is called abduction. Abduction means that the brain is capable to formulating scenarios. To augment the task that you were about to do, that you learned, or to do a new task based on the experience that you have assimilated before,” he said. “I’m leveraging legacy data from the factory now. What I did for the last few years is capturing knowledge from legacy that comes from operator machines. We’re using this to do the same disruption that Open AI did for creating an email or summarizing a book or, in my case, automating something.” The result is faster deployment, minimized downtime, and smarter, more resilient automation across industries through: Physics-Informed Machine Learning Model: Acting as a true digital twin, it accurately replicates real-world environments, adapting to different machines and motion platforms for precise, real-time optimization. Robotics & PLC AI Code Generation: Proprietary AI models autonomously generate both robotic programs and PLC code by understanding operational workflows and machine logic. This reduces deployment time by up to 80% and eliminates manual coding. Real-Time Process Learning Module: Powered by Data Ontology and Graph Neural Networks (GNNs), this module captures, maps, and understands complex relationships between machines, sensors, and processes. It ensures dynamic adaptation and continuous optimization. Cognitive Control Framework: A universal AI platform that integrates seamlessly with any robotic system, CNC machine, or PLC controller, supporting both legacy and modern equipment. “Industry 4.0 promised intelligent, autonomous factories—but it’s often been stuck in pilot purgatory, held back by rigid, code-heavy systems and legacy infrastructure,” said Gayathri Radhakrishnan, a partner at Hitachi Ventures, in a statement. “Xaba breaks that deadlock. By giving industrial machines the ability to self-learn and self-program through generative AI, Xaba is turning the vision of smart manufacturing into a scalable, reality today.” Xaba’s AI is already transforming aerospace, automotive, and high-precision manufacturing by eliminating costly rework and manual adjustments in areas such as automotive manufacturing. Xaba’s AI optimizes aluminum casting and forging, enabling robots to precisely machine metallic castings while adjusting for machining tolerances—dramatically reducing assembly costs, rework, and production time. It’s also doing large-scale robotic drilling. Manufacturers have achieved 10 times faster production rates while significantly lowering capital expenditures, Moruzzi said. Unlike traditional systems requiring rigid programming and manual adjustments, Xaba’s AI allows robots to seamlessly reconfigure different parts and processes without costly downtime. Xaba is also doing robotic welding. Xaba’s AI automates MIG (Metal Inert Gas) welding and TIG (Tungsten Inert Gas) laser welding, and laser welding, ensuring consistent, high-quality output across production lines while accelerating production timelines. And Xaba is handling large-scale 3D printing. Xaba’s xTrude system optimizes Fused Deposition Modeling (FDM), preventing delamination, collapse, and distortion. It allows manufacturers to fine-tune print parameters in real-time, improving reliability and reducing material waste. “Max and his team have created a bold new blueprint for the future of robotics and industrial automation,” said Marco Andriano, CEO Fives Cinetic Corp, in a statement. “Together, with Xaba’s xCognition, we’re delivering intelligent systems that transform decades of inefficiency into agile, scalable manufacturing environments—finally solving the most persistent programming and production challenges the industry faces today.” The company has 24 people working for it. The team includes AI scientists, mathematicians and mechatronics experts. They’re running an AI applied automation lab. Moruzzi believes that large language models (LLMs) are not the right technology for automation because the foundational model behind the LLM is based fundamentally on a set of weight factors. It’s like using an encyclopedia to answer a question about whether a robot should turn left or right. On top of that, LLMs are prone to hallucinations, which is bad in industrial settings. That means scale that may or may not synthesize the semantics in the way that gives you the right response, he said. Moruzzi designed his AI to be completely different, building a system that can create synthetic data on its own. “Your brain is not an LLM,” he said. Moruzzi said his company is entering production in the coming months. He noted there are only about 4.4 million industrial robots in the field now. That’s virtually nothing, considering how many humans there are. And the reason, he said, is that their brains — little more than industrial controllers — aren’t good enough. They’re like “empty boxes,” he said. “That’s why I’m building a cognitive brain,” he said. “That is the way to talk with the physical world.” GB Daily Stay in the know! Get the latest news in your inbox daily Read our Privacy Policy Thanks for subscribing. Check out more VB newsletters here. An error occured.

“Your honor, the FTC calls Mark Zuckerberg.” Flanked by two bodyguards, Meta’s CEO solemnly strode into a Washington, DC courtroom. Despite his last-ditch efforts to avoid a trial, he was there, jaw clenched, to defend his company from being broken up by the US government.Shortly after he was sworn in, the Federal Trade Commission’s lead attorney for the case, Daniel Matheson, asked Zuckerberg to reflect back on when Facebook was the underdog.“In hindsight, you’re glad you didn’t sell to MySpace?” Matheson asked. “Yes,” Zuckerberg responded.Over the next several hours of questioning, Matheson walked Zuckerberg down memory lane to the period just before Facebook’s $1 billion acquisition of Instagram in 2012, which the FTC claims was the first in a series of anti-competitive steps that locked out other companies. In a lawsuit that was initially filed five years ago and went to trial this week, the agency argues that Meta should be forced to spin off both Instagram and WhatsApp, which it later acquired for roughly $19 billion in 2014. While on the stand, Zuckerberg seemed to slowly relax as he recounted major moments from Facebook’s early history, from the launch of the News Feed to the company’s rocky transition to mobile phones in 2012. Considerable time was spent asking him about Facebook’s founding mission to connect friends and family, and how early rivals like Path and Google Plus challenged that use case. When asked to confirm that he has been Meta’s “sole decision maker” and controlling shareholder since 2006, he quickly nodded his head twice and said, “Yes.” While Matheson’s line of questioning at times felt monotonous, it seemed at least partly intended to provide historical context for Chief Judge James Boasberg, who admitted during pre-trial that he’d never used a Meta service. (At one point, Boasberg asked the Meta CEO for a crash course on what “native code” meant. Zuckerberg eagerly obliged.)Later in the day, the FTC started to hone in on the Instagram acquisition. Matheson showed internal emails in which Zuckerberg warned colleagues that Instagram’s early rise was “really scary” for Facebook. In other emails, he complained about the slow pace of development of Facebook’s own photos app, Facebook Camera, and described members of the team as “checked out.”“We really need to get our act together quickly on this since Instagram’s growing so fast,” Zuckerberg wrote in another internal email shown to the court. In a separate exchange with an engineering executive working on Facebook Camera, Zuckerberg tried to instill a sense of urgency: “If Instagram continues to kick ass on mobile or if Google buys them, then over the next few years they could easily add pieces of their service that copy what we’re doing now.”In court, Zuckerberg downplayed the threat Instagram posed to Facebook at the time. “Yeah, of course,” he said in response to Matheson asking if both apps were competing to connect friends with each other. “Was that the main thing that was going on? Not to my recollection.” The FTC’s case hinges on the argument that Meta has a monopoly in the US on “personal social networking services,” a market the agency says only includes Snapchat and MeWe, a self-described “privacy-first social media network” that claims to have “over 20 million users worldwide.” By including these two services, the FTC claims that Meta owns nearly 80 percent of active users in the market.During Meta’s opening arguments, the company’s lead lawyer Mark Hansen argued that the FTC’s market definition is artificially narrow by excluding TikTok, iMessage, and other services. He called the case “a grab bag of FTC theories at war with the facts and at war with the law.” A common strategy in antitrust cases is for a company to diminish its influence to appear less monopolistic. In Meta’s view, the market for user attention is much broader than the FTC’s definition. Hansen presented internal Meta data showing how Facebook and Instagram usage soared when TikTok was briefly offline in the US earlier this year. And when Facebook had a worldwide outage in 2021, he presented data showing that YouTube’s usage increased far more than Snapchat’s.A slide from Meta’s opening arguments. MetaEven if it can prove that Meta has monopoly power in a relevant market, the FTC will also have to show over the coming weeks that the company illegally acted to achieve or maintain its dominant position.To hear Meta retell it, the company saw opportunities where it could invest and grow fledgling products into now-massive apps used around the world. But the FTC argues that, like Zuckerberg’s early refusal to sell to MySpace, Instagram and WhatsApp would have been just fine on their own. At the end of the day, as the FTC’s Matheson was still quizzing Zuckerberg about his intent behind buying Instagram, Judge Boasberg asked to end the testimony. As Zuckerberg stepped down from the witness stand, one of his security guards motioned for them to leave the room before everyone else started filing out — another attempt for the CEO to get ahead of the fray.See More:

In the rapidly evolving landscape of large language models (LLMs), researchers and organizations face significant challenges. These include enhancing reasoning abilities, providing robust multilingual support, and efficiently managing complex, open-ended tasks. Although smaller models are often more accessible and cost-effective, they typically fall short in performance when compared to their larger counterparts. Hence, there is a growing emphasis on developing mid-sized models that effectively balance computational efficiency with strong reasoning and instruction-following capabilities. The recent release of GLM 4 from Tsinghua University, particularly the GLM-Z1-32B-0414 variant, addresses these challenges effectively. Trained on a substantial dataset of 15 trillion tokens, GLM 4 is designed to offer reliable multilingual capabilities and incorporates innovative reasoning strategies referred to as “thinking mode.” This release positions GLM 4 alongside other notable models like DeepSeek Distill, QwQ, and O1-mini, and is distributed under the widely respected MIT license. Notably, despite its relatively moderate parameter size of 32 billion, GLM 4 demonstrates performance comparable to much larger models such as GPT-4o and DeepSeek-V3, which contain up to 671 billion parameters, particularly in reasoning-centric benchmarks. On a technical level, GLM-Z1-32B-0414 leverages extensive high-quality training data, including synthetically generated reasoning tasks, to strengthen analytical capabilities. The model integrates sophisticated techniques such as rejection sampling and reinforcement learning (RL) to improve performance in agent-based tasks, coding, function calling, and search-driven question-answering tasks. Additionally, its “Deep Reasoning Model” variation further refines this by employing cold-start methods combined with extended RL training, specifically targeted at complex mathematical, logical, and coding tasks. Pairwise ranking feedback mechanisms are employed during training to enhance the model’s general reasoning effectiveness. An advanced variant, GLM-Z1-Rumination-32B-0414, introduces a novel approach termed “rumination,” enabling prolonged reflective reasoning for tackling open-ended, complex queries like comparative AI-driven urban analysis. This variant integrates advanced search tools with multi-objective reinforcement learning, significantly enhancing its utility in research-intensive tasks and complex retrieval-based scenarios. Complementing these larger models, the GLM-Z1-9B-0414 version, with its 9 billion parameters, provides strong mathematical and general reasoning capabilities, demonstrating the practicality of smaller-scale models. Performance data from benchmark evaluations emphasize the strengths of the GLM 4 series. Specifically, GLM-4-32B-0414 shows robust results compared to GPT-4o, DeepSeek-V3, and Qwen2.5-Max across multiple benchmarks. On the IFEval instruction-following benchmark, GLM 4 scores an impressive 87.6. In task automation benchmarks such as TAU-Bench, GLM 4 achieves strong scores in scenarios like retail (68.7) and airline (51.2). For search-augmented question-answering tasks, as evaluated by SimpleQA, the model records a high score of 88.1. Additionally, GLM 4 closely matches GPT-4o’s performance in function-calling tasks evaluated by the BFCL-v3 benchmark, securing an overall score of 69.6. In practical code repair scenarios tested through SWE-bench with the Moatless framework, GLM 4 achieves a success rate of 33.8%, underscoring its practical value. In summary, GLM 4 presents itself as an effective family of language models, successfully bridging the performance gap between smaller, more accessible models and the traditionally superior larger-scale counterparts. The GLM-Z1 series, especially the 32B variant, exemplifies this balanced approach by providing powerful reasoning capabilities while maintaining computational affordability. With the added advantage of its permissive MIT license, GLM 4 is positioned as a robust tool for research and enterprise applications requiring high-performance AI solutions without the extensive computational overhead traditionally associated with larger models. Check out GLM-4-Z1-32B-0414 Model and Other Models. All credit for this research goes to the researchers of this project. Also, feel free to follow us on Twitter and don’t forget to join our 90k+ ML SubReddit. Asif RazzaqWebsite |  + postsBioAsif Razzaq is the CEO of Marktechpost Media Inc.. As a visionary entrepreneur and engineer, Asif is committed to harnessing the potential of Artificial Intelligence for social good. His most recent endeavor is the launch of an Artificial Intelligence Media Platform, Marktechpost, which stands out for its in-depth coverage of machine learning and deep learning news that is both technically sound and easily understandable by a wide audience. The platform boasts of over 2 million monthly views, illustrating its popularity among audiences.Asif Razzaqhttps://www.marktechpost.com/author/6flvq/Small Models, Big Impact: ServiceNow AI Releases Apriel-5B to Outperform Larger LLMs with Fewer ResourcesAsif Razzaqhttps://www.marktechpost.com/author/6flvq/A Coding Implementation for Advanced Multi-Head Latent Attention and Fine-Grained Expert SegmentationAsif Razzaqhttps://www.marktechpost.com/author/6flvq/A Coding Implementation on Introduction to Weight Quantization: Key Aspect in Enhancing Efficiency in Deep Learning and LLMsAsif Razzaqhttps://www.marktechpost.com/author/6flvq/Allen Institute for AI (Ai2) Launches OLMoTrace: Real-Time Tracing of LLM Outputs Back to Training Data

Amazon is currently offering a SteelSeries Arctis Nova Pro Wireless gaming headset starting at only $257.55 shipped. The least expensive model is the Xbox edition in White. The Xbox edition is the only one that will work with the PS5, Xbox Series X, and PC. The PS5 edition is also on sale, but for $269.99, and it won't work with the Xbox console. The Nova Pro is our top rated gaming headset of 2025; I've personally use headset on a daily basis for the past few years and love it.SteelSeries Nova Pro Wireless Gaming Headset for $258SteelSeries Arctis Nova Pro Wireless Gaming Headset (Xbox, PS5, PC)SteelSeries Arctis Nova Pro Wireless Gaming Headset (PS5, PC)The Arctis Nova Pro is the highest-end gaming headset in SteelSeries' repertoire. It combines excellent sound quality with comfortable ergonomics, a high-performance mic, and unique yet extremely practical features. One of the biggest standout features of the SteelSeries Nova Pro wireless headset is the ingenious hot-swappable battery system. SteelSeries was generous enough to includes two batteries: one that's in use in your headset and the other that's charging in the separate DAC controller. This way, you get wireless freedom with no waiting to recharge. Each battery lasts for up to 22 hours of continued use, so you won't be constantly swapping them out either.Other great features include an OLED base station with volume control knob, active noise cancellation (rare in a gaming headset), retractable boom microphone, all-day comfort thanks to the suspension headband and leatherette ear cushions, simultaneous 2.4GHz WiFi and Bluetooth pairing, onboard audio presets, and the versatile Sonar software for creating your own custom sound profiles (for PC gamers).SteelSeries Arctis Nova Pro review by Kevin Lee"SteelSeries’ Pro Wireless headsets have always been a bit of a unicorn in the gaming headset world thanks to their outstanding sound quality, uniquely removable batteries, simultaneous connection support, and other rare features. The Arctis Nova Pro Wireless truly elevates this line of gaming headsets to another level: spatial sound and audio quality are fantastic, batteries can be hot-swapped for uninterrupted use, and you can make your gaming audio experience even more nuanced thanks to very noticeable changes in the equalizer and Sonar settings. The revised design accommodates all heads – large and small – comfortably, even for extended sessions. And to top things off, the more premium design and hybrid active noise canceling make the headset even more usable as a daily pair of wireless headphones."PlayWhy Should You Trust IGN's Deals Team?IGN's deals team has a combined 30+ years of experience finding the best discounts in gaming, tech, and just about every other category. We don't try to trick our readers into buying things they don't need at prices that aren't worth buying something at. Our ultimate goal is to surface the best possible deals from brands we trust and our editorial team has personal experience with. You can check out our deals standards here for more information on our process, or keep up with the latest deals we find on IGN's Deals account on Twitter.Eric Song is the IGN commerce manager in charge of finding the best gaming and tech deals every day. When Eric isn't hunting for deals for other people at work, he's hunting for deals for himself during his free time.

I’ve used Apple’s Calendar app for years, despite compelling third-party alternatives on the App Store, and one key iOS feature is a great example of why I’d have a hard time switching. iOS provides convenient Apple Calendar event features in other apps When comparing calendar apps, there are plenty of third-party options that offer more features than Apple Calendar. Fantastical is a personal favorite. It excels at offering power user tools like scheduling, calendar sets, and more. Many iPhone users stick with Google Calendar or perhaps Outlook because they’re more custom-built for those calendar services. But in a similar way, Apple’s Calendar app is designed to pair especially well with the iPhone and iOS—and has an exclusive system feature I use a lot. Whenever someone sends a proposed meeting date and time, iOS recognizes that and Apple Calendar lets you take action on it. This uses the iOS ‘data detectors’ system, which tags certain types of text on your iPhone as actionable. When calendar dates are detected, you can tap the text to see a menu with the following options: Add to Calendar Add to Reminders Show in Calendar Copy Event I use this feature all the time to quickly, conveniently manage my schedule. My two go-to actions from the menu are ‘Show in Calendar’ and ‘Add to Calendar.’ The former jumps you into the Calendar app and shows you the specific day and time mentioned, that way you know exactly what else is going on at that time. The latter provides a full pop-up, without leaving your current app, to create a new event in Calendar. Both serve as shortcuts to help you quickly take action on any date and time you encounter—whether in the Messages app, in Apple Mail, or elsewhere across iOS. In theory, if Apple ever offers a ‘default app’ feature for calendaring, these shortcuts will work with third-party apps too. For now, one workaround is to add all your various calendars to Apple Calendar—solely to benefit from these shortcuts—but then still do all other calendaring in your preferred third-party app. I’ve done that at times in the past, but highly prefer the reduced mental clutter of keeping everything in Apple’s Calendar app for a seamless experience. Do you ever use this Apple Calendar feature? Why or why not? Let us know in the comments. Best iPhone accessories Add 9to5Mac to your Google News feed.  FTC: We use income earning auto affiliate links. More.You’re reading 9to5Mac — experts who break news about Apple and its surrounding ecosystem, day after day. Be sure to check out our homepage for all the latest news, and follow 9to5Mac on Twitter, Facebook, and LinkedIn to stay in the loop. Don’t know where to start? Check out our exclusive stories, reviews, how-tos, and subscribe to our YouTube channel

Image by Getty / FuturismDevelopmentsLast year, University of Washington School of Medicine professor of biochemistry David Baker won the Nobel Prize for his work on designing proteins that can be used in drugs, vaccines, materials, and sensors.But now that the Trump administration has begun to diminish the role of research and gut scientific funding, around 15 of Baker's graduate students and postdoctoral researchers are looking to leave the US, NBC News reports.A major funding squeeze is forcing Baker and his colleagues at the Institute for Protein Design to reevaluate and cut back."There’s so many amazing people who want to come in, and we can’t take them," he told NBC. "The Nobel Prize was just a little blip. But things have gotten quite bleak."Trump's war on science in the US has sparked concerns over a major brain drain, with a Nature poll of more than 1,200 scientists finding that a startling 75 percent are now considering leaving the country.The Trump administration has gutted federal agencies, with the National Institutes of Health ringing the alarm bells following massive layoffs and budget cuts. Billions of dollars worth of contracts have been ordered to be canceled by Elon Musk's Department of Government Efficiency."Right now, due to the funding cuts, we are unable to enroll any more participants into federally funded studies, or start new studies, or do really any new work," UW Medicine infectious disease researcher Rachel Bender Ignacio, who cut her own salary to distribute money to the rest of other staffers, told NBC.Even politically uncontroversial lines of research, including Alzheimer's and cancer, have been swept up in a major shrinking of funding, which could lead to significant slowdowns in progress toward treatments, cures, and other interventions."We’ve gone through a bunch of contingency planning," University of Washington’s Alzheimer’s Disease Research Center director Thomas Grabowski told NBC, referring to grant decisions slowing to a crawl. "When it starts to look like multiple, multiple, multiple months, then there’s not a good answer to your question."The university received about 1,2200 grants from the NIH, worth around $648 million, last year. This year that approval process ground to a halt, and more than 600 grants are still in limbo.Scientists are now in the dark, awaiting some much-needed clarity from the agency, which has spent much of its resources pointlessly chasing after president Donald Trump's number-one bogeyman: diversity, equity, and inclusion (DEI) initiatives."The fact that they’re cutting these things or putting them in limbo is really upsetting, and you know, I feel like they’re doing surgery with a chainsaw at the federal level," retired attorney Andrea Gilbert, who had undergone treatment for Alzheimer's disease under Grabowski's care, told NBC News.Share This Article

We are seeking a talented freelancer to develop a comprehensive brand kit and identity for Contra, a networking platform that empowers individuals to work independently. This role involves crafting a cohesive and visually appealing brand presence that resonates with our values and mission. What you’ll do:Develop a complete brand kit, including logo, typography, color palette, and other visual elements.Create a brand identity that aligns with Contra's mission and values.Ensure all deliverables are consistent and suitable for various platforms and mediums.Collaborate with our team to align the brand identity with our strategic goals. 

After almost a decade of tweaking the same formula, Overwatch 2 devs have cooked up something wildly new for season 16.