News from the AI & ML world
DeeperML - #none
|
Ryan Daws@AI News
//
OpenAI is set to release its first open-weight language model since 2019, marking a strategic shift for the company. This move comes amidst growing competition in the AI landscape, with rivals like DeepSeek and Meta already offering open-source alternatives. Sam Altman, OpenAI's CEO, announced the upcoming model will feature reasoning capabilities and allow developers to run it on their own hardware, departing from OpenAI's traditional cloud-based approach.
This decision follows OpenAI securing a $40 billion funding round, although reports suggest a potential breakdown of $30 billion from SoftBank and $10 billion from Microsoft and venture capital funds. Despite the fresh funding, OpenAI also faces scrutiny over its training data. A recent study by the AI Disclosures Project suggests that OpenAI's GPT-4o model demonstrates "strong recognition" of copyrighted data, potentially accessed without consent. This raises ethical questions about the sources used to train OpenAI's large language models. Recommended read:
References :
alex.wawro@futurenet.com (Alex@tomsguide.com
//
Microsoft is actively integrating AI across its product lines to enhance both functionality and security. One significant development involves the use of AI-powered Security Copilot to identify vulnerabilities in open-source bootloaders. This expedited discovery process has revealed 20 previously unknown vulnerabilities in GRUB2, U-Boot, and Barebox, which could impact systems relying on Unified Extensible Firmware Interface (UEFI) Secure Boot. These vulnerabilities, particularly in GRUB2, could allow threat actors to bypass Secure Boot and install stealthy bootkits, potentially granting them complete control over affected devices.
Microsoft is also expanding AI capabilities on Copilot Plus PCs, particularly those powered by Intel and AMD processors. Features like Live Captions, which translates audio into English subtitles in real-time, as well as creative tools like Cocreator in Paint, Restyle Image, and Image Creator in Photos, are becoming more widely available. The company is additionally testing a new tool called Quick Machine Recovery, designed to remotely restore unbootable Windows 11 devices by automatically diagnosing and deploying fixes through Windows Update, preventing widespread outages similar to those experienced in the past. Recommended read:
References :
Alexey Shabanov@TestingCatalog
//
Microsoft is supercharging its Copilot assistant with new capabilities, transforming it into a companion for all. The company is equipping Copilot with new features designed to make it more responsive and helpful, including memory recall and personalization. This will allow the AI assistant to better understand and remember user preferences, complete tasks, analyze surroundings, and keep life organized. Microsoft aims to make AI work for everyone and wants Copilot to become the AI companion people want, tailored just for them.
Microsoft launched two AI reasoning agents for 365 Copilot: Researcher and Analyst. Researcher handles complex research using multiple sources, while Analyst functions as a data scientist to transform raw data into insights. These agents will roll out this month as part of a new program called "Frontier". The company is also adding new mobile and web features, personalization options, and exclusive tools for Surface devices. Recommended read:
References :
Noah Kravitz@NVIDIA Blog
//
NVIDIA is making strides in both agentic AI and open-source initiatives. Jacob Liberman, director of product management at NVIDIA, explains how agentic AI bridges the gap between powerful AI models and practical enterprise applications. Enterprises are now deploying AI agents to free human workers from time-consuming and error-prone tasks, allowing them to focus on high-value work that requires creativity and strategic thinking. NVIDIA AI Blueprints help enterprises build their own AI agents.
NVIDIA has announced the open-source release of the KAI Scheduler, a Kubernetes-native GPU scheduling solution, now available under the Apache 2.0 license. Originally developed within the Run:ai platform, the KAI Scheduler is now available to the community while also continuing to be packaged and delivered as part of the NVIDIA Run:ai platform. The KAI Scheduler is designed to optimize the scheduling of GPU resources and tackle challenges associated with managing AI workloads on GPUs and CPUs. Recommended read:
References :
Ellie Ramirez-Camara@Data Phoenix
//
Google has recently unveiled Gemini 2.5 Pro, hailed as its most intelligent AI model to date. This "thinking model" excels in reasoning and coding benchmarks, boasting a 1M token context window. It has already claimed the top position on the Chatbot Arena LLM Leaderboard, showcasing its superior performance in overall, math, instruction following, creative writing, and hard prompts. Gemini 2.5 Pro is now available for Gemini Advanced users, with Vertex AI integration planned for the near future.
The company states the enhanced base model and improved post-training techniques allows Gemini 2.5 Pro to go beyond classification and prediction. In addition to its achievements in the Chatbot Arena LLM Leaderboard, Gemini 2.5 Pro has demonstrated state-of-the-art capabilities across academic benchmarks, achieving leading scores in math and science benchmarks, including GPQA and AIME 2025, as well as a score of 18.8% on Humanity's Last Exam. Recommended read:
References :
Kara Sherrer@eWEEK
//
Runway AI Inc. has launched Gen-4, its latest AI video generation model, addressing the significant challenge of maintaining consistent characters and objects across different scenes. This new model represents a considerable advancement in AI video technology and improves the realism and usability of AI-generated videos. Gen-4 allows users to upload a reference image of an object to be included in a video, along with design instructions, and ensures that the object maintains a consistent look throughout the entire clip.
The Gen-4 model empowers users to place any object or subject in different locations while maintaining consistency, and even allows for modifications such as changing camera angles or lighting conditions. The model combines visual references with text instructions to preserve styles throughout videos. Gen-4 is currently available to paying subscribers and Enterprise customers, with additional features planned for future updates. Recommended read:
References :
Chris McKay@Maginative
//
Anthropic has unveiled Claude for Education, a specialized AI assistant designed to cultivate critical thinking skills in students. Unlike conventional AI tools that simply provide answers, Claude employs a Socratic-based "Learning Mode" that prompts students with guiding questions, encouraging them to engage in deeper reasoning and problem-solving. This innovative approach aims to address concerns about AI potentially hindering intellectual development by promoting shortcut thinking.
Partnerships with Northeastern University, the London School of Economics, and Champlain College will integrate Claude across multiple campuses, reaching tens of thousands of students. These institutions are making a significant investment in AI, betting that it can improve the learning process. Faculty can use Claude to generate rubrics aligned with learning outcomes and create chemistry equations, while administrative staff can analyze enrollment trends and simplify policy documents. These institutions are testing the system across teaching, research, and administrative workflows. Recommended read:
References :
Megan Crouse@eWEEK
//
Cloudflare has launched AI Labyrinth, a new tool designed to combat web scraping bots that steal website content for AI training. Instead of simply blocking these crawlers, AI Labyrinth lures them into a maze of AI-generated content. This approach aims to waste the bots' time and resources, providing a more effective defense than traditional blocking methods which can trigger attackers to adapt their tactics. The AI Labyrinth is available as a free, opt-in tool for all Cloudflare customers, even those on the free tier.
The system works by embedding hidden links within a protected website. When suspicious bot behavior is detected, such as ignoring robots.txt rules, the crawler is redirected to a series of AI-generated pages. This content is "real looking" and based on scientific facts, diverting the bot from the original website's content. Because no human would deliberately explore deep into a maze of AI-generated nonsense, anyone who does can be identified as a bot with high confidence. Cloudflare emphasizes that AI Labyrinth also functions as a honeypot, allowing them to identify new bot patterns and improve their overall bot detection capabilities, all while increasing the cost for unauthorized web scraping. Recommended read:
References :
Ken Yeung@Ken Yeung
//
Google has launched a new feature called "Discover Sources" for NotebookLM, its AI-powered tool designed to organize and analyze information. Rolling out to all users starting April 2, 2025, the new feature automatically curates relevant websites on a specified topic, recommending up to ten sources accompanied by AI-generated summaries. This enhancement streamlines research by allowing users to quickly surface relevant content from the internet.
NotebookLM, initially launched in 2023 as an AI-powered alternative to Evernote and Microsoft OneNote, previously relied on manual uploads of documents, articles, and notes. "Discover Sources" automates the process of pulling in information from the internet with a single click. The curated sources remain accessible within NotebookLM notebooks, allowing users to leverage them within Briefing Docs, FAQs, and Audio Overviews without repeatedly scouring the internet. This enhancement highlights the growing trend of AI-driven research tools shaping how we work and learn. Recommended read:
References :
@phys.org
//
Google's DeepMind has achieved a significant breakthrough in artificial intelligence with its Dreamer AI system. The AI has successfully mastered the complex task of mining diamonds in Minecraft without any explicit human instruction. This feat, accomplished through trial-and-error reinforcement learning, demonstrates the AI's ability to self-improve and generalize knowledge from one scenario to another, mimicking human-like learning processes. The achievement is particularly noteworthy because Minecraft's randomly generated worlds present a unique challenge, requiring the AI to adapt and understand its environment rather than relying on memorized strategies.
Mining diamonds in Minecraft is a complex, multi-step process that typically requires players to gather resources to build tools, dig to specific depths, and avoid hazards like lava. The Dreamer AI system tackled this challenge by exploring the game environment and identifying actions that would lead to rewards, such as finding diamonds. By repeating successful actions and avoiding less productive ones, the AI quickly learned to navigate the game and achieve its goal. According to Jeff Clune, a computer scientist at the University of British Columbia, this represents a major step forward for the field of AI. The Dreamer AI system, developed by Danijar Hafner, Jurgis Pasukonis, Timothy Lillicrap and Jimmy Ba, achieved expert status in Minecraft in just nine days, showcasing its rapid learning capabilities. One unique approach used during training was to restart the game with a new virtual universe every 30 minutes, forcing the algorithm to constantly adapt and improve. This innovative method allowed the AI to quickly master the game's mechanics and develop strategies for diamond mining without any prior training or human intervention, pushing the boundaries of what AI can achieve in dynamic and complex environments. Recommended read:
References :
Nitika Sharma@Analytics Vidhya
//
Google's DeepMind has achieved a significant milestone in artificial intelligence by developing an AI system, named Dreamer, that has mastered Minecraft without any human instruction or data. The Dreamer AI system successfully learned how to mine diamonds, a complex and multi-step process, entirely on its own through trial and error. This breakthrough highlights the potential for AI systems to generalize knowledge and transfer skills from one domain to another, marking a major step forward in the field of AI development.
Researchers programmed the Dreamer AI to play Minecraft by setting up a system of rewards, particularly for finding diamonds. The AI explores the game on its own, identifying actions that lead to in-game rewards and repeating those actions. The AI was able to reach an expert level within just nine days. The results are a good sign that AI apps can learn to improve its abilities over a short period of time, which could give robots the tools they need to perform well in the real world. Recommended read:
References :
Help Net@Help Net Security
//
Microsoft is enhancing Windows 11 roadmap transparency with new initiatives to better inform IT professionals and users about upcoming features. The company has launched a new Windows roadmap website designed to simplify the tracking of new Windows 11 features. This move addresses a key criticism regarding the lack of clarity around the testing and rollout phases of new functionalities. Microsoft aims to provide IT administrators with more insights, enabling them to effectively manage changes across their Windows estates.
The new roadmap consolidates information from various sources, including the Windows Insider Program and Microsoft's support site, offering a unified view of in-development features. Users can filter features based on platform, Windows versions, and rollout status, gaining insights into descriptions, release dates, and compatibility details. While the roadmap currently focuses on the client version of Windows 11, Microsoft plans to expand it to include other Windows versions in the future and is accepting feedback to further improve the tool's utility. Recommended read:
References :
staff@insideAI News
//
MLCommons has released the latest MLPerf Inference v5.0 benchmark results, highlighting the growing importance of generative AI in the machine learning landscape. The new benchmarks feature tests for large language models (LLMs) like Llama 3.1 405B and Llama 2 70B Interactive, designed to evaluate how well systems perform in real-world applications requiring agentic reasoning and low-latency responses. This shift reflects the industry's increasing focus on deploying generative AI and the need for hardware and software optimized for these demanding workloads.
The v5.0 results reveal significant performance improvements driven by advancements in both hardware and software. The median submitted score for Llama 2 70B has doubled compared to a year ago, and the best score is 3.3 times faster than Inference v4.0. These gains are attributed to innovations like support for lower-precision computation formats such as FP4, which allows for more efficient processing of large models. The MLPerf Inference benchmark suite evaluates machine learning performance in a way that is architecture-neutral, reproducible, and representative of real-world workloads. Recommended read:
References :
@Google DeepMind Blog
//
References:
Google DeepMind Blog
, The Next Web
,
Google DeepMind is intensifying its focus on AI governance and security as it ventures further into artificial general intelligence (AGI). The company is exploring AI monitors to regulate hyperintelligent AI models, splitting potential threats into four categories, with the creation of a "monitor" AI being one proposed solution. This proactive approach includes prioritizing technical safety, conducting thorough risk assessments, and fostering collaboration within the broader AI community to navigate the development of AGI responsibly.
DeepMind's reported clampdown on sharing research will stifle AI innovation, warns the CEO of Iris.ai, one of Europe’s leading startups in the space, Anita Schjøll Abildgaard. Concerns are rising within the AI community that DeepMind's new research restrictions threaten AI innovation. The CEO of Iris.ai, a Norwegian startup developing an AI-powered engine for science, warns the drawbacks will far outweigh the benefits. She fears DeepMind's restrictions will hinder technological advances. Recommended read:
References :
@Latest from Laptop Mag
//
Microsoft is celebrating its 50th anniversary on April 4th, 2025, marking a significant milestone focused on its AI driven future. Top executives highlighted the company's vision for AI, centered around Copilot, which will be further integrated into Microsoft products. Mustafa Suleyman, CEO of Microsoft AI, emphasized that Copilot will understand users in the context of their lives and show up in the right way at the right time, calling it "a new kind of relationship with technology."
Microsoft is launching Researcher and Analyst AI agents in Microsoft 365 Copilot. Researcher integrates OpenAI's deep research model with Microsoft 365 Copilot's orchestration and search capabilities to perform complex, multi-step research workflows. Analyst, powered by OpenAI's o3-mini reasoning model, helps turn raw data into actionable insights within minutes, capable of running Python code to process complex data queries. These agents, accessible through the "Frontier" program, aim to provide on-demand assistance in data analysis and general research tasks, enhancing user capabilities across various applications. Recommended read:
References :
Ken Yeung@Ken Yeung
//
Microsoft is launching a 50-day AI Skills Fest, starting April 8, 2025, aimed at helping individuals advance their technical knowledge in artificial intelligence. The global event will feature a wide array of activities including hackathons, both live in-person and virtual events, self-paced tutorials, and community events. Microsoft hopes to set a Guinness World Record for the most participants completing an online multi-level AI lesson in 24 hours as part of the inaugural launch.
Microsoft is also rolling out updates to its Copilot+ PCs. AMD and Intel-powered Copilot+ PCs will gain access to Windows Copilot features like Cocreator, Image Creator, Live Captions and Restyle Image, previously available only on Snapdragon-powered devices. These features utilize the onboard Neural Processing Unit (NPU) and generative AI to enhance user experiences within Microsoft Paint and Photos, enabling image generation and modification. Recommended read:
References :
Matt Marshall@AI News | VentureBeat
//
References:
Microsoft Security Blog
, www.zdnet.com
Microsoft is enhancing its Copilot Studio platform with AI-driven improvements, introducing deep reasoning capabilities that enable agents to tackle intricate problems through methodical thinking and combining AI flexibility with deterministic business process automation. The company has also unveiled specialized deep reasoning agents for Microsoft 365 Copilot, named Researcher and Analyst, to help users achieve tasks more efficiently. These agents are designed to function like personal data scientists, processing diverse data sources and generating insights through code execution and visualization.
Microsoft's focus includes securing AI and using it to bolster security measures, as demonstrated by the upcoming Microsoft Security Copilot agents and new security features. Microsoft aims to provide an AI-first, end-to-end security platform that helps organizations secure their future, one example being the AI agents designed to autonomously assist with phishing, data security, and identity management. The Security Copilot tool will automate routine tasks, allowing IT and security staff to focus on more complex issues, aiding in defense against cyberattacks. Recommended read:
References :
Michael Nuñez@AI News | VentureBeat
//
References:
venturebeat.com
, The Algorithmic Bridge
,
Anthropic has been at the forefront of investigating how AI models like Claude process information and make decisions. Their scientists developed interpretability techniques that have unveiled surprising behaviors within these systems. Research indicates that large language models (LLMs) are capable of planning ahead, as demonstrated when writing poetry or solving problems, and that they sometimes work backward from a desired conclusion rather than relying solely on provided facts.
Anthropic researchers also tested the "faithfulness" of CoT models' reasoning by giving them hints in their answers, and see if they will acknowledge it. The study found that reasoning models often avoided mentioning that they used hints in their responses. This raises concerns about the reliability of chains-of-thought (CoT) as a tool for monitoring AI systems for misaligned behaviors, especially as these models become more intelligent and integrated into society. The research emphasizes the need for ongoing efforts to enhance the transparency and trustworthiness of AI reasoning processes. Recommended read:
References :
Nazy Fouladirad@AI Accelerator Institute
//
References:
hiddenlayer.com
, AI Accelerator Institute
,
As generative AI adoption rapidly increases, securing investments in these technologies has become a paramount concern for organizations. Companies are beginning to understand the critical need to validate and secure the underlying large language models (LLMs) that power their Gen AI products. Failing to address these security vulnerabilities can expose systems to exploitation by malicious actors, emphasizing the importance of proactive security measures.
Microsoft is addressing these concerns through innovations in Microsoft Purview, which offers a comprehensive set of solutions aimed at helping customers seamlessly secure and confidently activate data in the AI era. Complementing these efforts, Fiddler AI is focusing on building trust into AI systems through its AI Observability platform. This platform emphasizes explainability and transparency. They are helping enterprise AI teams deliver responsible AI applications, and also ensure people interacting with AI receive fair, safe, and trustworthy responses. This involves continuous monitoring, robust security measures, and strong governance practices to establish long-term responsible AI strategies across all products. The emergence of agentic AI, which can plan, reason, and take autonomous action to achieve complex goals, further underscores the need for enhanced security measures. Agentic AI systems extend the capabilities of LLMs by adding memory, tool access, and task management, allowing them to operate more like intelligent agents than simple chatbots. Organizations must ensure security and oversight are essential to safe deployment. Gartner research indicates a significant portion of organizations plan to pursue agentic AI initiatives, making it crucial to address potential security risks associated with these systems. Recommended read:
References :
Stephen Ornes@Quanta Magazine
//
References:
Quanta Magazine
, medium.com
A novel quantum algorithm has demonstrated a speedup over classical computers for a significant class of optimization problems, according to a recent report. This breakthrough could represent a major advancement in harnessing the potential of quantum computers, which have long promised faster solutions to complex computational challenges. The new algorithm, known as decoded quantum interferometry (DQI), outperforms all known classical algorithms in finding good solutions to a wide range of optimization problems, which involve searching for the best possible solution from a vast number of choices.
Classical researchers have been struggling to keep up with this quantum advancement. Reports of quantum algorithms often spark excitement, partly because they can offer new perspectives on difficult problems. The DQI algorithm is considered a "breakthrough in quantum algorithms" by Gil Kalai, a mathematician at Reichman University. While quantum computers have generated considerable buzz, it has been challenging to identify specific problems where they can significantly outperform classical machines. This new algorithm demonstrates the potential for quantum computers to excel in optimization tasks, a development that could have broad implications across various fields. Recommended read:
References :
|
BenchmarksBlogsResearch Tools |