Thursday, May 1st, 2025 Posted by Jim Thacker Lumion releases Lumion View for SketchUp html PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd" Originally posted on 30 March 2025, and updated for the final release. Lumion has released Lumion View, its new real-time visualization plugin for SketchUp. The plugin provides users of the architectural modeling software with a ray traced view of their scenes, making it possible to visualize changes to materials and lighting more accurately. The add-on is described as a “lightweight design companion” to Lumion, the standalone edition of the software, and is aimed at early-stage design workflows. Lumion: an easy-to-use near-real-time tool for architectural visualisation work One of the most widely used renderers for arch viz work, the standalone edition of Lumion creates renders of architectural scenes in near-real time. It is intended to provide architects with little background in visualization with a more straightforward way to create realistic stills and animations than standard DCC apps. Users import building models from other software, either in standard 3D file formats like FBX, OBJ, SKP and MAX, or via built-in live links to major CAD applications. They can then edit materials inside Lumion, set up lighting and weather effects, dress the scene from the accompanying asset library, and create simple camera animations. Lumion View: a new plugin integrating the rendering technology directly into SketchUp The new Lumion View plugin cuts out the import and setup work, enabling users to visualize buildings directly inside SketchUp, Trimble’s architectural sketching software. Lumion pitches it as a tool for initial design work – for exploring ideas and adjusting massing, lighting and materials – that works alongside Lumion itself “without duplication of effort”. It provides a ray traced view of a SketchUp scene, with changes to materials and lights displayed in real time. As well as adjusting the properties of the new PBR materials introduced in SketchUp 2025.0, users can assign Lumion’s Standard, Glass and Water materials to objects. It is also possible to assign readymade ‘conceptual styles’, which replace all of the existing materials in a scene, aside from glass and water, with clay, wood or styrofoam. To light the scene, users can place and edit native SketchUp Spotlights, Point Lights, Area Lights and Line Lights, or adjust the direction and brightness of the Lumion sun light. The completed scene can be rendered inside SketchUp at up to 4K resolution, or transferred to Lumion for editing: exported SKP files retain material settings and lights created in Lumion View. Lumion plans to make Lumion View compatible with Revit and other CAD software in future: you can register online to be notified of the wider release. Price, licensing and system requirements Lumion View 2025.0 is compatible with SketchUp 2025.0 running on Windows only. It is compatible with NVIDIA, AMD and Intel GPUs, including iGPUs: you can find full system requirements here. The software is available rental only, and licensed on a named-user basis, with subscriptions having a MSRP of $229/year. You can find pricing details for Lumion itself in this story. Read an overview of Lumion View on the product website Find more technical details in the online release notes for Lumion View 2025.0 Have your say on this story by following CG Channel on Facebook, Instagram and X (formerly Twitter). As well as being able to comment on stories, followers of our social media accounts can see videos we don’t post on the site itself, including making-ofs for the latest VFX movies, animations, games cinematics and motion graphics projects. Latest News Lumion releases Lumion 2025.0 Significant update to the real-time visualization software adds AI render upscaling, ray traced water and fog, and new pricing and licensing. Thursday, May 1st, 2025 Lumion releases Lumion View for SketchUp New real-time design exploration and visualization plugin lets users generate ray traced 4K renders inside the 3D modeling software. Thursday, May 1st, 2025 Kiri Engine 3.14 adds AI-enhanced Lidar scanning Mobile 3D scanning app uses new machine learning techniques to refine Lidar environment scans captured on iPhone and iPad Pros. Wednesday, April 30th, 2025 CLO Virtual Fashion releases Marvelous Designer 2025.0 Big update to the 3D clothing design tool adds a new fur material, new keyframe animation capabilities, and a modular garment library. Wednesday, April 30th, 2025 Chaos releases Vantage 2.8 Real-time renderer gets support for light linking, lens effects, and Shader Execution Reordering on NVIDIA GPUs. See all the new features. Wednesday, April 30th, 2025 Ziva VFX and OIDN creators win Sci-Tech Academy Awards Soft tissue simulation and render denoising system win tech 'Oscars', due to be presented at a ceremony tonight. Check out the other winners. Tuesday, April 29th, 2025 More News Sparseal adds SDF modeling to Uniform 1.4 EZ-Tree is a lightweight free open-source tree generator Substance 3D creator Sébastien Deguy to leave Adobe Download free Houdini tools from SideFX's Project Grot Master FACS Rigging for Facial Motion Capture Unity 6.1: discover 5 key features for CG artists Adobe releases Photoshop 26.6 Chaos releases Corona 12 Update 2 for 3ds Max and Cinema 4D Get 260+ modular Wild West building assets for Unreal Engine Get VFX artist Thomas Marcos' free Blender clouds shader NVIDIA open-sources PhysX's GPU simulation code Hair Cinematic Tool 2.0 simplifies hair shading in Unreal Engine 5 Older Posts

The Qwen2.5-Omni-3B model is licensed for non-commercial use only under Alibaba Cloud’s Qwen Research License Agreement.Read More

Apple can no longer collect fees on purchases made outside the App Store in the US, court rules As a result, Fortnite will return to the App Store next week News by Samuel Roberts Editorial Director Published on May 1, 2025 Epic Games has claimed victory in its legal battle against Apple, after a US court enforced an injunction which prohibits the tech giant from collecting fees on app purchases made outside its App Store. In addition, Apple is no longer allowed to stop developers from encouraging US-based users to make payments via alternative methods – like web transactions – that circumvent its ecosystem. The ruling is effective immediately. As a result, Fortnite will return to the App Store in the US next week, Epic's Tim Sweeney revealed on Twitter. "The Court enjoins Apple from implementing its new anticompetitive acts to avoid compliance with the Injunction," reads the order. "Effective immediately Apple will no longer impede developers' ability to communicate with users nor will they levy or impose a new commission on off-app purchases." The full ruling by the Northern District of California Court also refers Apple and its vice president of finance Alex Roman to the local US attorney for investigation regarding criminal contempt. The Verge reports that Apple will appeal the order. Regarding Roman specifically, the order states he "outright lied under oath". The original injunction was ordered following a trial between Epic and Apple in 2021, which ruled that Apple's 30% commission was 'anticompetitive', and that it needed to allow developers to direct users to alternative payment options. Following this, however, Apple added a 27% commission to 'off-app' purchases. The resulting new ruling explains that Apple used methods like 'scare screens' to deter users from exiting its ecosystem to make payments. "Apple's goal: to dissuade customer usage of alternative purchase opportunities and maintain its anticompetitive revenue stream," reads the order on this point. "In the end, Apple sought to maintain a revenue stream worth billions in direct defiance of this Court's Injunction." The results of the order only apply to the US, but Sweeney pitched a larger compromise on Twitter. "Epic puts forth a peace proposal: If Apple extends the court's friction-free, Apple-tax-free framework worldwide, we'll return Fortnite to the App Store worldwide and drop current and future litigation on the topic." Last month, Apple was fined $568 million by the EU for deterring third-party payment methods that circumvent the App Store.

Tesla has hit back following a Wall Street Journal report that its board had initiated a search for a new CEO to replace Elon Musk, taking to Musk’s social platform X to call the claim “absolutely false.” According to the WSJ report, about a month ago Tesla’s board began the process of hunting for a successor to Musk, reaching out to “several” executive search firms and getting as far as narrowing down to one firm to lead the search.  Around the same time the board reportedly told Musk he needed to spend more time on Tesla, and less on DOGE, and say so publicly. In an earnings call last week he did exactly that, promising that from May he’d be “allocating far more” of his time to Tesla. The auto company has suffered from slumping sales, declining revenue, and a major brand crisis since Musk took on a closer role with Donald Trump and led efforts to slash government spending. There’s good reason to think Tesla and Musk’s other companies stand to benefit from that work too, though. Tesla moved quickly to reject the Journal’s report, publishing a rebuttal on X from board chair Robyn Denholm, who said the directors are “highly confident” in Musk’s ability to lead Tesla. She both denied that the board had initiated a CEO search, and claimed that “this was communicated to the media before the report was published.” The WSJ has not yet updated its story or responded to Denholm’s statement, though notes in its original report that it reached out to Musk for comment and didn’t receive any. Musk himself later tweeted his thoughts, accusing the paper of an “EXTREMELY BAD BREACH OF ETHICS,” and suggesting that it had been sent an “unequivocal denial beforehand.” Tesla’s board has eight members, including Musk himself along with his brother, Kimbal Musk, and James Murdoch, son of Rupert Murdoch. The group has been accused of being “full of sycophants,” and “overly beholden to CEO Musk,” especially in the wake of last year’s efforts to secure Musk a record-breaking pay package. As for Denholm? Despite that high confidence in Musk, she just sold over $32 million worth of Tesla shares, her third major sale in three months, and appears to be liquidating her position in the company.

Addressing the Challenges in Reasoning-Intensive Retrieval Despite notable progress in retrieval-augmented generation (RAG) systems, retrieving relevant information for complex, multi-step reasoning tasks remains a significant challenge. Most retrievers today are trained on datasets composed of short factual questions, which align well with document-level lexical or semantic overlaps. However, they fall short when faced with longer, abstract, or cross-domain queries that require synthesizing dispersed knowledge. In such cases, retrieval errors can propagate through the pipeline, impairing downstream reasoning by large language models (LLMs). While LLM-based rerankers can improve relevance, their substantial computational cost often renders them impractical in real-world deployments. Meta AI has released ReasonIR-8B, a retriever model designed explicitly for reasoning-intensive information retrieval. Trained from LLaMA3.1-8B, the model establishes new performance standards on the BRIGHT benchmark, achieving a normalized Discounted Cumulative Gain (nDCG@10) of 36.9 when used with a lightweight Qwen2.5 reranker. Notably, it surpasses leading reranking models such as Rank1-32B while offering 200× lower inference-time compute, making it significantly more practical for scaled RAG applications. ReasonIR-8B is trained using a novel data generation pipeline, ReasonIR-SYNTHESIZER, which constructs synthetic queries and document pairs that mirror the challenges posed by real-world reasoning tasks. The model is released open-source on Hugging Face, along with training code and synthetic data tools, enabling further research and reproducibility. Model Architecture, Training Pipeline, and Key Innovations ReasonIR-8B employs a bi-encoder architecture, where queries and documents are encoded independently into embeddings and scored via cosine similarity. The model’s training relies heavily on synthetically generated data tailored to reasoning scenarios. The ReasonIR-SYNTHESIZER pipeline produces two primary types of training instances: Varied-Length (VL) Queries: These are long, information-rich queries (up to 2000 tokens), paired with corresponding documents, encouraging the retriever to handle extended contexts effectively. Hard Queries (HQ): Derived from curated documents with high educational value, these queries are designed to require logical inference. Multi-turn prompts are used to construct hard negatives—documents that appear superficially relevant but do not contain the necessary reasoning pathways. This approach contrasts with conventional negative sampling methods, which often rely on lexical overlap and are less effective for abstract or multi-hop questions. Additionally, the model’s attention mask is modified from LLaMA’s causal configuration to a bi-directional one, allowing the encoder to consider the full query context symmetrically, which is beneficial for non-sequential semantic alignment. Empirical Results on IR and RAG Benchmarks ReasonIR-8B achieves strong performance across several benchmarks: BRIGHT Benchmark (Reasoning-Intensive Retrieval): 24.4 nDCG@10 on original queries 29.9 with GPT-4 rewritten queries 36.9 with Qwen2.5 reranking, outperforming larger LLM rerankers at a fraction of the cost Retrieval-Augmented Generation (RAG) Tasks: +6.4% improvement on MMLU over a closed-book baseline +22.6% improvement on GPQA These gains are consistent across both standard and rewritten queries, with further improvements observed when combining REASONIR-8B with a sparse retriever like BM25 or a lightweight reranker. Importantly, the model continues to improve as query lengths scale, unlike other retrievers whose performance plateaus or declines. This suggests that ReasonIR-8B can better exploit information-rich queries, making it particularly well-suited for test-time techniques such as query rewriting. Conclusion ReasonIR-8B addresses a key bottleneck in reasoning-focused information retrieval by introducing a retriever optimized not only for relevance but also for computational efficiency. Its design—rooted in synthetic training tailored for reasoning, coupled with architectural and data-centric improvements—enables consistent gains in both retrieval and RAG tasks. By releasing the model, codebase, and training data generation pipeline as open-source tools, Meta AI encourages the research community to extend this work toward more robust, multilingual, and multimodal retrievers. For applications requiring cost-effective and high-quality retrieval under reasoning constraints, ReasonIR-8B represents a compelling and practical solution. Check out the Twitter and join our Telegram Channel and LinkedIn Group. 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/Microsoft AI Released Phi-4-Reasoning: A 14B Parameter Open-Weight Reasoning Model that Achieves Strong Performance on Complex Reasoning TasksAsif Razzaqhttps://www.marktechpost.com/author/6flvq/Multimodal AI on Developer GPUs: Alibaba Releases Qwen2.5-Omni-3B with 50% Lower VRAM Usage and Nearly-7B Model PerformanceAsif Razzaqhttps://www.marktechpost.com/author/6flvq/Mem0: A Scalable Memory Architecture Enabling Persistent, Structured Recall for Long-Term AI Conversations Across SessionsAsif Razzaqhttps://www.marktechpost.com/author/6flvq/Diagnosing and Self- Correcting LLM Agent Failures: A Technical Deep Dive into τ-Bench Findings with Atla’s EvalToolbox

Author(s): Naveen Krishnan Originally published on Towards AI. Image Source — Author Massive large language models (LLMs) dominated headlines, showcasing impressive feats of text generation and comprehension. However, a new wave is cresting — the era of Small Language Models (SLMs). These compact yet surprisingly capable models are challenging the status quo, proving that exceptional performance doesn’t always require colossal size and resources. Microsoft has been at the forefront of this movement with its Phi family of models. Following the success of Phi-3, which demonstrated remarkable abilities for its size, Microsoft has taken another significant leap forward with the introduction of the Phi-4 reasoning models. These new additions Phi-4-reasoning, Phi-4-reasoning-plus, and Phi-4-mini-reasoning — are specifically engineered to excel at complex reasoning tasks, a domain previously thought to be the exclusive territory of much larger models. They represent a pivotal moment, bringing sophisticated reasoning capabilities within reach for a wider range of applications and hardware. The Phi-4 Reasoning Family: More Than Just Words Before getting into the specifics of the Phi-4 models, it’s helpful to understand what sets “reasoning models” apart. While standard language models excel at predicting the next word in a sequence, reasoning models are trained to go further. They leverage techniques like inference-time scaling to tackle complex problems that require breaking tasks down into multiple steps, performing internal checks, and logically connecting pieces of information. This often involves generating intermediate steps or a “chain of thought” to arrive at a final answer, a capability typically associated with much larger, resource-intensive frontier models. The Phi-reasoning models represent a significant advancement by bringing these sophisticated reasoning abilities into the more efficient framework of SLMs, achieved through meticulous data curation, distillation from larger models, reinforcement learning, and a focus on high-quality synthetic datasets. Phi-4-reasoning: The Foundation The cornerstone of this new lineup is Phi-4-reasoning. With 14 billion parameters, it’s significantly smaller than many models boasting similar reasoning prowess. This process specifically trains the model to generate detailed reasoning chains, effectively utilizing additional computational steps during inference to solve complex problems. The key achievement here is demonstrating that smaller models, when trained on exceptionally high-quality and targeted data (including synthetic datasets), can effectively compete with their much larger counterparts in demanding reasoning tasks, particularly in mathematics and science. Phi-4-reasoning-plus: Enhanced Accuracy Through Reinforcement Building directly upon the foundation laid by Phi-4-reasoning, the “plus” variant introduces an additional layer of refinement through reinforcement learning (RL). While Phi-4-reasoning is trained via SFT on reasoning examples, Phi-4-reasoning-plus undergoes further training using RL techniques. Phi-4-mini-reasoning: Compact Power for the Edge Addressing the growing need for capable AI on devices with limited resources, this 3.8-billion parameter model is specifically optimized for mathematical reasoning and step-by-step problem-solving in environments where computational power or network latency is a constraint. Phi-4-mini-reasoning strikes a crucial balance between efficiency and advanced reasoning ability, making it an ideal candidate for applications like embedded tutoring systems, on-device AI assistants (like those envisioned for Copilot+ PCs), and lightweight deployments on mobile or edge systems where complex reasoning is needed without relying on cloud connectivity. Benchmarks: Performance Highlights Microsoft’s technical reports and blog posts highlight impressive results across a range of demanding benchmarks, demonstrating that these SLMs punch well above their weight class. Phi-4-reasoning performance across representative reasoning benchmarks spanning mathematical and scientific reasoning Both the 14B parameter Phi-4-reasoning and Phi-4-reasoning-plus models showcase remarkable capabilities. They consistently outperform not only the base Phi-4 model but also significantly larger open-weight models like the 70-billion parameter DeepSeek-R1-Distill-Llama-70B across various reasoning benchmarks. This includes complex mathematical reasoning tasks, such as those found in the AIME 2025 (the qualifier for the USA Math Olympiad), where they reportedly achieve performance better than even the full 671-billion parameter DeepSeek-R1 Mixture-of-Experts model. They also show strong performance on Ph.D. level science questions and general capability benchmarks like MMLUPro (measuring knowledge and language understanding), HumanEvalPlus (coding), IFEval (instruction following), and ArenaHard (general skills). The graph compares the performance of various models on popular math benchmarks for long sentence generation. The compact Phi-4-mini-reasoning (3.8B parameters) also holds its own impressively. Designed for efficiency, it still manages to outperform its base model and several larger models (including some 7B and 8B parameter models like OpenThinker-7B and DeepSeek-R1-Distill-Llama-8B) on popular math benchmarks such as MATH and GPQA Diamond. Accessing Phi-4 The Phi-4-reasoning, Phi-4-reasoning-plus, and Phi-4-mini-reasoning models are available in the Azure AI model catalog. From there, they can often be deployed as serverless API endpoints. This pay-as-you-go approach simplifies access, allowing developers to integrate these powerful reasoning capabilities into their applications without managing the underlying infrastructure. Image Source — Author Using the Azure AI Inference Python SDK One of the most straightforward ways to interact with deployed Phi-4 models on Azure is through the azure-ai-inference Python SDK. Here’s a breakdown of how to get started: Install the SDK: If you haven’t already, install the necessary Python package: pip install azure-ai-inference # Required importsimport osfrom azure.ai.inference import ChatCompletionsClientfrom azure.ai.inference.models import SystemMessage, UserMessagefrom azure.core.credentials import AzureKeyCredentialendpoint = "<<Your Model Endpoint>>"model_name = "<<Your Model Name>>"client = ChatCompletionsClient( endpoint=endpoint, credential=AzureKeyCredential("<API_KEY>"), )response = client.complete( messages=[ SystemMessage(content="You are a helpful assistant."), UserMessage(content="I am going to Paris, what should I see?"), ], max_tokens=4096, temperature=0.8, top_p=0.95, presence_penalty=0.0, frequency_penalty=0.0, model=model_name)print(response.choices[0].message.content) Run a multi-turn conversation: This sample demonstrates a multi-turn conversation with the chat completion API. When using the model for a chat application, you’ll need to manage the history of that conversation and send the latest messages to the model. import osfrom azure.ai.inference import ChatCompletionsClientfrom azure.ai.inference.models import AssistantMessage, SystemMessage, UserMessagefrom azure.core.credentials import AzureKeyCredentialendpoint = "<<Your Model Endpoint>>"model_name = "<<Your Model Name>>"client = ChatCompletionsClient( endpoint=endpoint, credential=AzureKeyCredential("<API_KEY>"), )response = client.complete( messages=[ SystemMessage(content="You are a helpful assistant."), UserMessage(content="I am going to Paris, what should I see?"), AssistantMessage(content="Paris, the capital of France, is known for its stunning architecture, art museums, historical landmarks, and romantic atmosphere. Here are some of the top attractions to see in Paris:\n \n 1. The Eiffel Tower: The iconic Eiffel Tower is one of the most recognizable landmarks in the world and offers breathtaking views of the city.\n 2. The Louvre Museum: The Louvre is one of the world's largest and most famous museums, housing an impressive collection of art and artifacts, including the Mona Lisa.\n 3. Notre-Dame Cathedral: This beautiful cathedral is one of the most famous landmarks in Paris and is known for its Gothic architecture and stunning stained glass windows.\n \n These are just a few of the many attractions that Paris has to offer. With so much to see and do, it's no wonder that Paris is one of the most popular tourist destinations in the world."), UserMessage(content="What is so great about #1?") ], max_tokens=4096, temperature=0.8, top_p=0.95, presence_penalty=0.0, frequency_penalty=0.0, model=model_name)print(response.choices[0].message.content) Stream the output: For a better user experience, you will want to stream the response of the model so that the first token shows up early and you avoid waiting for long responses. import osfrom azure.ai.inference import ChatCompletionsClientfrom azure.ai.inference.models import SystemMessage, UserMessagefrom azure.core.credentials import AzureKeyCredentialendpoint = "<<Your Model Endpoint>>"model_name = "<<Your Model Name>>"client = ChatCompletionsClient( endpoint=endpoint, credential=AzureKeyCredential("<API_KEY>"), )response = client.complete( stream=True, messages=[ SystemMessage(content="You are a helpful assistant."), UserMessage(content="I am going to Paris, what should I see?") ], max_tokens=4096, temperature=0.8, top_p=0.95, presence_penalty=0.0, frequency_penalty=0.0, model=model_name)for update in response: if update.choices: print(update.choices[0].delta.content or "", end="")client.close() Use Cases The enhanced reasoning capabilities packed into the efficient Phi-4 models unlock a diverse range of potential applications, pushing the boundaries of what AI can achieve, especially in resource-constrained scenarios. Agentic Applications: The ability to handle complex, multi-step tasks makes Phi-4-reasoning and reasoning-plus strong candidates for powering AI agents. These agents could potentially automate complex workflows, perform sophisticated research, or manage intricate planning tasks that require logical decomposition and execution. Educational Tools: Phi-4-mini-reasoning, with its strong mathematical focus and compact size, is particularly well-suited for educational applications. Imagine intelligent tutoring systems embedded directly into learning platforms or devices, providing step-by-step guidance and personalized feedback for subjects like math and logic. Coding and Development: The strong performance on coding benchmarks suggests utility as a powerful coding assistant, capable of understanding complex logic, generating code snippets, debugging, and potentially even assisting with algorithmic problem-solving. Mathematical & Scientific Problem Solving: The models’ demonstrated prowess in math and science reasoning opens doors for applications in research, data analysis, and engineering, helping professionals tackle complex calculations, simulations, or hypothesis generation. On-Device AI: Phi-4-mini-reasoning, along with optimizations like Phi Silica for Windows NPUs, paves the way for more powerful AI experiences directly on personal computers (like Copilot+ PCs) and mobile devices. This enables features like offline summarization, advanced text intelligence, and potentially more sophisticated local assistants that don’t rely constantly on cloud connectivity. The broader impact of the Phi-4 reasoning family lies in democratizing access to advanced AI reasoning. By delivering capabilities previously confined to massive models within a smaller footprint, Microsoft is enabling developers to build more intelligent applications that can run efficiently on a wider variety of hardware, fostering innovation across numerous fields. Conclusion: Small Models, Big Reasoning Microsoft’s Phi-4 reasoning models mark a significant stride in the evolution of artificial intelligence. By packing state-of-the-art reasoning capabilities into remarkably efficient Small Language Models, they challenge the notion that only massive parameter counts can yield sophisticated cognitive abilities. From the foundational Phi-4-reasoning, enhanced by supervised fine-tuning, to the accuracy-boosted Phi-4-reasoning-plus refined with reinforcement learning, and the compact Phi-4-mini-reasoning designed for edge deployment, this family offers tailored solutions for diverse needs. Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor. Published via Towards AI

Pokémon TCG’s next big release, Destined Rivals, is out this month, and preorders are starting to appear again at retailers like Target and Walmart. We've collected all the retailers where you can buy the new expansion at, just below. Keep an eye on the listings for new restocks, and follow @IGNDeals for the latest updates.Destined Rivals US PreordersPokémon - Trading Card Game: Scarlet & Violet Destined Rivals Elite Trainer BoxPokémon - Trading Card Game: Scarlet & Violet Destined Rivals 6Pk Booster BundlePokémon - Trading Card Game: Scarlet & Violet Destined Rivals 3Pk BoosterSee it at WalmartPokémon - Trading Card Game: Scarlet & Violet Destined Rivals Booster Box (36 Packs)Pokémon - Trading Card Game: Scarlet & Violet Destined Rivals Sleeved BoosterSee it at Best BuyDestined Rivals Brings Back Classic Villains and Stunning New CardsTeam Rocket returns to stir up trouble once again, Trainers' signature Pokémon are back in the spotlight, and the card art? Some of the slickest in years. Whether you're chasing eye-catching collectibles or just addicted to the thrill of cracking a fresh pack, Destined Rivals is engineered to tempt you.Beware shady third-party resellers when preordering Destined Rivals. Here's the MSRP for each set so you don't get caught out.ETB: $49.99Booster Bundle: $26.99Booster Display Box: $159.991pk Booster: $4.993pk Booster: $12.99When Does Destined Rivals Release?Mark your calendar: the full set launches on May 30, 2025. That’s when sealed products officially hit shelves—stock willing. The Pokémon Company seems to be improving on shortages, but if history’s any guide, don’t count on leftovers.Between May 17–25, pre-release events will pop up at select stores, offering early access via Build & Battle boxes and small-scale tournaments. Want in? Talk to your local league store yesterday. And maybe bring snacks—you’re negotiating with gatekeepers now.What’s in the Destined Rivals Lineup?We all tell ourselves we’ll buy just one item. That’s adorable. Here’s everything that’ll test your willpower on launch day:Booster PacksBooster Boxes (36 packs)Elite Trainer BoxPokémon Center Exclusive Elite Trainer BoxBooster Bundle (6 packs)Triple-Pack BlistersBuild & Battle BoxBuild & Battle StadiumExpect premium collection boxes too—likely with alternate art promos starring fan-favorites like Misty, Cynthia, Ethan, or Marnie. You technically don’t need them. You’ll get them anyway.Alert: Amazon Massive TCG Restock Now LiveThere's been a massive restock of Pokémon TCG thanks to Amazon pooling it's stock levels internationally. You won't find sets like Prismatic Evolutions, but previous Scarlet and Violet / Sword and Shield sets are in stock right now.Pokemon TCG: Paradox Fury Premium Collection - 2024 Amazon Holiday ExclusivePokémon TCG: Scarlet & Violet—Twilight Masquerade Elite Trainer Box$59.87 at Amazon Pokémon TCG: Scarlet & Violet—Shrouded Fable Booster Bundle (6 Booster Packs)$29.87 at AmazonPokemon TCG: Mimikyu ex Box Pokémon TCG: Roaring Moon ex Box$31.22 at AmazonPokémon TCG: Divergent Powers Tin – Hisuian Typhlosion V$28.49 at AmazonPokémon TCG: Azure Legends Tin – Dialga ex$29.77 at AmazonPokemon TCG Annihilape EX Box$35.92 at AmazonPokémon TCG - Paradox Destinies Tin - Raging Bolt$29.92 at AmazonPokémon TCG: Collector Chest (Fall 2023): Meowscarada, Skeledirge & Quaquaval$42.74 at AmazonPokémon TCG: Azure Legends Tin – Xerneas ex $33.47 at Amazon Pokémon Kingambit Illustration Rare Box$29.29 at AmazonPokémon TCG: Scarlet & Violet—Shrouded Fable Mini Tin – Zoroark$20.24 at AmazonPokémon TCG: GO Tin - SnorlaxPOKEMON TCG: Combined Powers Premium Collection$61.07 at AmazonPokémon TCG: Terapagos ex Ultra-Premium Collection$123.76 at Amazon Pokemon TCG: Morpeko V-Union Special Collection$45.80 at Amazon Pokemon TCG Iron Valiant ex Box$44.29 at Amazon Pokemon TCG: Pokemon GO Premier Deck Holder Collection Dragonite VSTAR$67.61 at AmazonPokémon TCG: Darkness Stacking Tin - Umbreon$25.33 at AmazonPokémon TCG: Paradox Clash Tin: Walking Wake EX$31.38 at AmazonPokémon TCG: Shrouded Fable Triple Blister$24.41 at AmazonPokémon TCG: Fighting Stacking Tin - LycanrocPokémon TCG: Paradox Destinies Tin - Gouging Fire ex$33.35 at AmazonPokemon TCG: Trainer’s Toolkit 2023$47.75 at AmazonPokémon TCG: Houndstone ex Box$32.48 at AmazonPokémon TCG: Palafain ex Box$35.11 at AmazonPokémon TCG: Dragon Stacking Tin $25.02 at Amazon Pokémon TCG: Psychic Stacking Tin$49.29 at AmazonPokémon TCG: Collector Chest – Scream Tail & Iron Valiant $42.72 at Amazon Pokémon TCG: Scarlet & Violet—Twilight Masquerade Three-Booster Blister$31.31 at AmazonPokémon TCG: Vibrant Paldea Mini Tin - Arboliva w/LeafeonPokémon TCG: Vibrant Paldea Mini Tin – Palafin & Pachirisu$21.42 at AmazonPokémon TCG: Vibrant Paldea Mini Tin – Espathra & Ampharos$20.04 at AmazonPokémon TCG: Pokémon GO Pin Collection - Squirtle$27.80 at AmazonPokémon TCG: Pokémon GO Pin Collection – BulbasaurPokémon TCG: Vibrant Paldea Mini Tin – Dachsbun & Oricorio$20.39 at AmazonPokémon TCG: Holiday Calendar$60.54 at Amazon Pokémon TCG: Bloodmoon Ursaluna ex Box$32.14 at AmazonPokémon TCG: Back to School Eraser Blister – Bellibolt$24.20 at AmazonPokemon 2023 Holiday Calendar$73.45 at AmazonPokemon Sword and Shield Brilliant Stars Build and Battle Box$63.34 at AmazonPokémon Trading Card Game Classic$404.43 at AmazonPokémon TCG: Metal Stacking Tin$26.10 at AmazonPokémon TCG: Back to School Eraser Blister – Gengar$21.34 at AmazonPokemon TCG: Hisuian Electrode V Box$42.45 at AmazonPokemon TCG: Scarlet & Violet Build & Battle Stadium$84.00 at AmazonPokémon TCG: Paldea Partners Tin — Quaquaval ex$44.45 at AmazonPokémon TCG: Paldea Friends Mini Tins—Pawmi$21.26 at AmazonPokémon TCG: Poké Ball Tin Bundle 2024 Version—Poké Ball, Great Ball & Ultra Ball$65.16 at AmazonPokemon TCG: Sword & Shield—Astral Radiance Build & Battle Stadium$82.41 at AmazonPokemon TCG: 2024 Trainer’s Toolkit$34.98 at AmazonPokemon TCG: Scarlet & Violet - Surging Sparks Booster Bundle$45.02 at AmazonPokemon TCG: Shining Fates Collection Pikachu V Box$49.96 at AmazonPokemon TCG: Scarlet and Violet Shrouded Fable Elite Trainer Box$55.00 at Amazon Pokemon TCG: Scarlet and Violet Elite Trainer Box - Koraidon RedPokemon TCG: Scarlet & Violet Stellar Crown Elite Trainer Box$57.69 at AmazonCards That Should Be In Destined RivalsHere’s the full expected card list by Japanese set origin.Heat Wave ArenaEthan’s Ho-Oh exCynthia’s Garchomp exCynthia’s RoseradeMisty’s Psyduck, Staryu, Starmie, Magikarp, Gyarados, LaprasEthan’s Cyndaquil, Quilava, TyphlosionEthan’s Slugma, MagcargoHydrapple lineYanmega exZeraora, Electivire ex, Rotom, ManectricSteven’s Metang (alt print)Arven’s Mabostiff exMarnie’s Impidimp (alt print)Applin, DipplinOgerpon (Teal Mask, Hearthflame Mask, Wellspring Mask, Cornerstone Mask variants)Cynthia’s Milotic, FeebasBuizel, Floatzel, Dondozo exDwebble, CrustleShayminPonyta, RapidashArven’s Toedscool, ToedscruelArven’s Maschiff, Skwovet, GreedentMudbray, MudsdaleElectabuzzEthan’s Pinsir, Ethan’s PichuTrainer Cards: Judge, Ethan’s Adventure, Cynthia’s Power Weight, Sacred Ash, MC’s Hype Up, Spikemuth GymThe Glory of Team RocketTeam Rocket’s Mewtwo exTeam Rocket’s SpidopsTeam Rocket’s Meowth, Persian exTeam Rocket’s Porygon, Porygon2, Porygon-ZTeam Rocket’s TarountulaTrainer Cards: Team Rocket’s Giovanni, Archer, Ariana, ReceiverSpecial Energy: Team Rocket EnergySteven’s Starter DeckSteven’s Metagross exSteven’s Skarmory, Beldum, MetangSteven’s CarbinkSteven’s Claydol, BaltoyTrainer Card: Granite CaveMarnie’s Starter DeckMarnie’s Grimmsnarl exMarnie’s Impidimp, Morgrem, Liepard, ScraftyMarnie’s Purrloin, ScraggyTrainer Cards: Energy Recycler (reprint), Spikemuth GymPlayDestined Origins Cards I’ve Got My Eye OnThere’s a good chance I’ll end up with multiple binders full of these, but a few cards have already secured a permanent place in my mental wishlist.Cynthia’s Garchomp ex is a power move in every sense. Big damage, hand draw, and the smug energy of someone who always gets her turn one setup. This is peak Champion energy and I want three.Ethan’s Ho-Oh ex is what happens when utility and sparkle collide. The ability accelerates Fire Energy like it’s no big deal, and Shining Feather heals your whole team while smacking for 160. It’s absurd, and I’m obsessed.Team Rocket’s Mewtwo ex is pure drama. It refuses to attack unless you’ve built an entire Rocket-themed deck around it, but when it does, it throws energy around like a caffeinated Gengar. It’s also incredibly extra, which I respect.Misty’s Psyduck is here to make you laugh and then probably lose a game in the most lovable way possible. It has an ability that lets you discard it to the top of your deck, for... reasons. I don’t care. It’s perfect.Marnie’s Grimmsnarl ex brings the villain vibes with an aggressive Dark-type build that thrives on energy acceleration and being just annoying enough to work. It’s a disruption deck’s dream, and I plan to run it like it’s 2020 all over again.Pokémon TCG Single Card DealsThat's right, I'm hunting down single card deals now too. Hype it as an all time high right now, so that means prices are at their peak. That also means they're ready to bottom out at any moment, as ridiculous pricing turns buyers away eventually. The Pokémon TCG bubble is set to burst this year, either by scalpers getting too greedy or by massive reprints. Some prices are already dropping, with Journey Together slowly calming down:Surging Sparks Single CardsPikachu ex - 238/191 - SV08$431.99 at TCGPlayerLatias ex - 239/191 - SV08$499.95 at TCG PlayerMilotic ex - 237/191 - SV08$147.00 at TCG PlayerPikachu ex - 247/191 - SV08$120.92 at TCG PlayerHydreigon ex - 240/191 - SV08$84.69 at TCG PlayerJourney Together Single CardsLillie's Clefairy ex - 184/159$224.74 at TCGPlayerN's Zoroark ex - 185/159$249.97 at TCGPlayerIono's Bellibolt ex - 183/159$149.99 at TCGPlayerHop's Zacian ex - 186/159$100.39 at TCGPlayerN's Zoroark ex - 189/159$62.95 at TCGPlayerShrouded Fable Single CardsPersian - 078/064 - SV: Shrouded Fable$67.61 at TCGPlayerCassiopeia - 094/064 - SV: Shrouded Fable$59.98 at TCGPlayerFezandipiti ex - 092/064 - SV: Shrouded FableSee it at TCGPlayerHoundoom - 066/064 - SV: Shrouded Fable$49.99 at TCGPlayerDuskull - 068/064 - SV: Shrouded Fable$49.88 at TCGPlayerTemporal Forces Single CardsRaging Bolt ex - 208/162 - SV05$112.94 at TCGPlayerIron Crown ex - 206/162 - SV05$64.95 at TCGPlayerWalking Wake ex - 205/162 - SV05$149.95 at TCGPlayerGouging Fire ex - 204/162 - SV05$57.73 at TCGPlayerGastly - 177/162 - SV05$66.24 at TCGPlayerStellar Crown Single CardsSquirtle - SV07: Stellar Crown$91.24 at TCGPlayerTerapagos ex - 170/142 - SV07$69.70 at TCGPlayerBulbasaur - SV07: Stellar Crown$86.24 at TCGPlayerDachsbun ex - 169/142 - SV07: Stellar Crown$38.00 at TCGPlayerHydrapple ex - 167/142 - SV07: Stellar Crown$37.69 at TCGPlayerScarlet and Violet Base Single CardsGardevoir ex - 245/198 - SV01: Scarlet & Violet Base$54.99 at TCGPlayerDrowzee - 210/198 - SV01: Scarlet & Violet Base$43.74 at TCGPlayerRalts - 211/198 - SV01: Scarlet & Violet Base$41.99 at TCGPlayerKirlia - 212/198 - SV01: Scarlet & Violet Base$40.98 at AmazonMiriam - 251/198 - SV01: Scarlet & Violet Base$33.54 at TCGPlayerTwilight Masquerade Single CardsGreninja ex - 214/167 - SV06: Twilight Masquerade$219.99 at TCGPlayerPerrin - 220/167 - SV06: Twilight Masquerade$93.56 at TCGPlayerCarmine - 217/167 - SV06: Twilight Masquerade$71.74 at TCGPlayerEevee - 188/167 - SV06: Twilight Masquerade$32.48 at TCGPlayerBloodmoon Ursaluna ex - 216/167 - SV06: Twilight Masquerade$26.99 at TCGPlayerPokemon TCG: Paradox Fury Premium Collection - 2024 Amazon Holiday ExclusivePokémon TCG: Scarlet & Violet—Twilight Masquerade Elite Trainer Box$59.87 at Amazon Pokémon TCG: Scarlet & Violet—Shrouded Fable Booster Bundle (6 Booster Packs)$29.87 at AmazonPokemon TCG: Mimikyu ex Box Pokémon TCG: Roaring Moon ex Box$31.22 at AmazonPokémon TCG: Divergent Powers Tin – Hisuian Typhlosion V$28.49 at AmazonPokémon TCG: Azure Legends Tin – Dialga ex$29.77 at AmazonPokemon TCG Annihilape EX Box$35.92 at AmazonPokémon TCG - Paradox Destinies Tin - Raging Bolt$29.92 at AmazonPokémon TCG: Collector Chest (Fall 2023): Meowscarada, Skeledirge & Quaquaval$42.74 at AmazonPokémon TCG: Azure Legends Tin – Xerneas ex $33.47 at Amazon Pokémon Kingambit Illustration Rare Box$29.29 at AmazonPokémon TCG: Scarlet & Violet—Shrouded Fable Mini Tin – Zoroark$20.24 at AmazonPokémon TCG: GO Tin - Snorlax

The travel series Conan O’Brien Must Go returns to Max for a second season on May 8, following the comedian as he travels across the globe, experiencing local cultures and connecting with fans he previously met on his podcast. The Sex and the City spinoff And Just Like That comes back for a third season on May 29. Carrie (Sarah Jessica Parker), Miranda (Cynthia Nixon), Charlotte (Kristin Davis) and their friends once again do their best to navigate love and friendship and life in NYC as 50-something women. The Oscar-winning film The Brutalist arrives on Max on May 16. This film follows visionary architect László Toth as he escapes post-war Europe and tries to rebuild his life, his career, and his marriage in America. Another exciting movie joining the Max library this month (May 31) is the new original film Mountainhead. Written and directed by Jesse Armstrong, the creator of Succession, Mountainhead follows a group of billionaire friends who meet up for a getaway amidst an ongoing global financial crisis. Here’s everything coming to HBO and Max in May. HBO and Max New Releases – May 2025 May 1 100 Foot Wave, Season 3 (HBO Original)A Haunted House (2013)A Haunted House 2 (2014)A Private War (2018)Badman’s Country (1958)Barricade (1950)Broadway Melody of 1940 (1940)Carson City (1952)Chronicle (2012)Chronicle: Director’s Cut (2012)Dallas (1950)Dylan’s Playtime Adventures, Season 1B (Max Original)Enter The Dragon (1973)Eternal Sunshine of the Spotless Mind (2004)Fort Dobbs (1958)Fort Worth (1951)Funny People (2009)Hannibal (2001)In the Fade (2017)Inception (2010)Legend of the Guardians The Owls of Ga’Hoole (2010)Lone Star (1952)Lost River (2015)Madagascar (2005)Madagascar 3: Europe’s Most Wanted (2012)Madagascar: Escape 2 Africa (2008)Megan Leavey (2017)Notting Hill (1999)Pee-wee’s Big Adventure (1985)Rachel and the Stranger (1948)Rancho Notorious (1952)Red Tails (2012)Ride, Vaquero! (1953)Riding Shotgun (1954)Rocky Mountain (1950)Royal Wedding (1951)San Antonio (1945)Santa Fe Trail (1940)Shoot-Out at Medicine Bend (1957)Son of Belle Starr (1953)Son of the Mask (2005)Springfield Rifle (1952)Stars in My Crown (1950)Tall in the Saddle (1944)Tall Man Riding (1955)The Band Wagon (1953)The Bounty Hunter (1954)The First Texan (1956)The Goonies (1985)The Kissing Bandit (1948)The Left-Handed Gun (1958)The Man Behind the Gun (1953)The Mortal Storm (1940)The Naked Spur (1953)The Oklahoma Kid (1939)The Oklahoman (1957)The Painted Hills (1951)The Princess Bride (1987)The Shining (1980)The Silence of the Lambs (1991)The Young Guns (1956)They Died with Their Boots On (1941)This Means War (2012)This Means War: Extended Edition (2012)Thunder Over the Plains (1953)Trail Street (1947)Vengeance Valley (1951)Vivacious Lady (1938)We Bought a Zoo (2011)Westbound (1959)Westward the Women (1951)Wichita (1955)Ziegfeld Girl (1941)May 2 Adult Best Friends (2024)Malditos, Season 1 (Max Original)Teen Titans Go!, Season 9A (Cartoon Network)May 3 Ready Player One (2018)May 6 Fareed Zakaria: The War on Government (CNN)Talking Pictures Podcast – Special Video Episode with Eli Roth Recorded At TCM Classic Film FestivalThe Playboy Murders, Season 3 (ID)Toxic, Season 1 (ID)May 8 Conan O’Brien Must Go, Season 2 (Max Original) May 9 Bloody Trophy (Krwawe Trofeum) (Max Original)The Whale (A24, 2022) Join our mailing list Get the best of Den of Geek delivered right to your inbox! May 11 Delicious Miss Brown, Season 10 (Food Network)The Bakersfield 3: A Tale of Murder and Motherhood, Season 1 (ID)May 12 Where the Crawdads Sing (2022)May 13 Betting on Paradise, Season 1 (HGTV)Contraband: Seized at the Border, Season 6 (Discovery)Dream Destination: A Surprise Trip, Season 1 (discovery+)May 14 Cutthroat Kitchen: Knives Out, Season 1 (Food Network)Mini Reni, Season 1 (Magnolia Network)May 15 Duster, Season 1 (Max Original)Mermicorno: Starfall, Season 2 (Max Original)United States of Scandal with Jake Tapper, Season 2 (CNN Original) May 16 The Brutalist (A24)May 17 Bugs Bunny Builders, Season 2E (Cartoon Network)May 18 Love & Marriage: Detroit, Season 2 (OWN)Say Yes to the Dress, Season 21 (TLC)May 19 American Monster: Abuse of Power, Season 1 (ID)May 20 Kong: Skull Island (2017)May 21 Homicide Hunter: American Detective, Season 5 (ID) May 22 Expedition Files, Season 2 (Discovery)May 23 Pee-wee as Himself (HBO Original)Restoring Galveston, Season 7 (Magnolia Network)Silly Sundays, Season 1B (Cartoon Network)May 25 Say Yes To The Dress With Tan France, Season 1 (TLC)May 28 Down Home Fab, Season 3 (HGTV)May 29 And Just Like That…, Season 3 (Max Original)May 31 Mountainhead (HBO Original)

Apple has just been handed perhaps the most bullet-proof injunction ever, affecting the App Store in the United States, as a result of the ongoing dispute between Apple and Epic Games. Apple has said it will appeal, but comply in the meantime — and the judgement explicitly spells out how to do so, with seemingly little wiggle room. There’s one small concession that app developers must conform to, but otherwise they get free reign … The latest developments date back to this November 2021 judgement, in which the judge presiding over Epic v. Apple declared that Apple must allow developers to freely link out to alternative payment methods on the web. Following a series of appeals, Apple finally implemented new terms in January 2024 that allowed developers to link out, but slapped a 27% commission on anything the user purchased outside of the app in a seven-day period. It also enforced specific language for the text of the link, guidelines on the visual presentation of the link (in a plain style, rather than a button), and required the app to show a full-screen modal “scare sheet” before opening the browser. Epic argued that these terms did not fulfil the free and open requirements of the 2021 judgment, and complained to the court. The result of this was finally handed down yesterday, hence the new injunction. The new injunction basically throws out all of the associated restrictions Apple put on the links, as well as removing any opportunity for Apple to collect commission on purchases made outside of the app. The judge used very clear language in the injunction document, presumably to prevent Apple from trying to weasel out of compliance again. Specifically, Apple must not: Impose any commission or any fee on purchases that consumers make outside an app Restrict developers’ style, language, formatting, quantity, flow or placement of links for purchases outside an app  Prohibiting or limiting the use of buttons or other calls to action Exclude certain categories of apps and developers from obtaining link access Interfere with consumers’ choice to proceed in or out of an app by using anything other than a neutral message apprising users that they are going to a third-party site Restrict a developer’s use of dynamic links that bring consumers to a specific product page in a logged-in state rather than to a statically defined page These seven points set out opposition to basically everything Apple required developers to do as of January 2024. The only mitigation is nestled in point 5. Apple is not allowed to interfere with the link-out process other than “a neutral message apprising users that they are going to a third-party site”. This means Apple is allowed to require a form of interstitial alert, that notifies the user that they are navigating outside of the walls of the native app. However, it is clearly not allowed to be as obstructive as the previously-enforced full-screen modal scare sheet. In fact, the judge goes further and lays out an approved example of a substitute, a simple pop-up dialog using simple language (shown in the header image of this post). Here’s a side-by-side of the old scare sheet and what the judge now recommends: What’s somewhat amusing is the newly-approved alert design was actually suggested during internal Apple design meetings as one possible option. The judge says Apple ultimately landed on most anticompetitive option, the full screen takeover. As such, now, the company doesn’t get to choose. 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

For over a decade, application security teams have faced a brutal irony: the more advanced the detection tools became, the less useful their results proved to be. As alerts from static analysis tools, scanners, and CVE databases surged, the promise of better security grew more distant. In its place, a new reality took hold—one defined by alert fatigue and overwhelmed teams. According to OX Security's 2025 Application Security Benchmark Report, a staggering 95–98% of AppSec alerts do not require action - and may, in fact, be harming organizations more than helping. Our research, spanning over 101 million security findings across 178 organizations, shines a spotlight on a fundamental inefficiency in modern AppSec operations. Of nearly 570,000 average alerts per organization, just 202 represented true, critical issues. It's a startling conclusion that's hard to ignore: security teams are chasing shadows, wasting time, burning through budgets, and straining relations with developers over vulnerabilities that pose no real threat. The worst part of it - is that security gets in the way of actual innovation. As Chris Hughes puts it in Resilient Cyber: "We do all this while masquerading as business enablers, actively burying our peers in toil, delaying development velocity, and ultimately impeding business outcomes. How We Got Here: Mountains of Issues, Zero Context Back in 2015, the application security challenge was simpler. That year, just 6,494 CVEs were publicly disclosed. Detection was king. Tools were measured by how many issues they found - not whether they mattered. Fast forward to 2025: Applications went cloud-native, development cycles accelerated, and attack surfaces ballooned. In just the past year, over 40,000 new CVEs were published, bringing the global total to over 200,000. Yet, despite these major changes, many AppSec tools have failed to evolve: they've doubled down on detection, flooding dashboards with unfiltered, context-free alerts. OX's benchmark confirms what practitioners have long suspected: 32% of reported issues have a low probability of exploitation 25% have no known public exploit 25% stem from unused or development-only dependencies This flood of irrelevant findings doesn't just slow security down - it actively impairs it. While most alerts can be disregarded, it is essential to accurately identify the 2-5% that require immediate attention. The report shows these rare alerts usually involve KEV issues, secrets management problems, and in some cases, posture management issues. The Need for A Holistic Prioritization Approach To combat this doom-spiral, organizations must adopt a more sophisticated approach to application security, based on evidence-driven prioritization. This requires a shift from generic alert handling to a comprehensive model that covers code from design stages to runtime, and includes multiple elements: Reachability: Is the vulnerable code used, and is it reachable? Exploitability: Are the conditions for exploitation present in this environment? Business Impact: Would a breach here cause real damage? Cloud-to-Code Mapping: Where in the SDLC did this issue originate? By implementing such a framework, organizations can effectively filter out the noise and focus their efforts on the small percentage of alerts that pose a genuine threat. This improves security effectiveness, frees up valuable resources, and enables more confident development practices. OX Security is addressing this challenge with Code Projection, an evidence-based security technology that maps cloud and runtime elements back to code origin, enabling contextual understanding and dynamic risk prioritization. Real-World Impact The data tells a powerful story: By using evidence-based prioritization, the alarming average of 569,354 total alerts per organization can be reduced to 11,836, of which only 202 require immediate action. Industry benchmarks reveal several key insights: Consistent Noise Thresholds: Baseline noise levels remain remarkably similar across different environments, whether enterprise or commercial, regardless of industry. Enterprise Security Complexity: Enterprise environments face significantly greater challenges due to their broader tool ecosystem, larger application footprint, higher volume of security events, more frequent incidents, and elevated overall risk exposure. Financial Sector Vulnerability: Financial institutions experience distinctively higher alert volumes. Their processing of financial transactions and sensitive data makes them high-value targets. As the Verizon Data Breach Investigations Report indicates, 95% of attackers are motivated primarily by financial gain rather than espionage or other reasons. Financial institutions' proximity to monetary assets creates direct profit opportunities for attackers. The findings have far-reaching implications. If less than 95% of application security fixes are critical to the organization, then all organizations invest enormous resources in triage, programming, and cybersecurity hours in vain. This waste extends to payments for bug-bounty programs, where white-hat hackers find vulnerabilities to fix, as well as the costs of complicated fixes for vulnerabilities that were not discovered early and reached production. The final significant cost is the tension created within organizations between development teams and security teams, who demand fixes for vulnerabilities that aren't relevant. Detection failed, Prioritization is the Way Forward As organizations face a projected 50,000 new vulnerabilities in 2025 alone, the stakes for effective security triage have never been higher. The old model of "detect everything, fix later" is not just outdated - it's dangerous. OX Security's Report makes a compelling case: The future of application security lies not in addressing every possible vulnerability but in intelligently identifying and focusing on the issues that pose real risk. Found this article interesting? This article is a contributed piece from one of our valued partners. Follow us on Twitter  and LinkedIn to read more exclusive content we post.