Close Menu
    DevStackTipsDevStackTips
    • Home
    • News & Updates
      1. Tech & Work
      2. View All

      Sunshine And March Vibes (2025 Wallpapers Edition)

      May 17, 2025

      The Case For Minimal WordPress Setups: A Contrarian View On Theme Frameworks

      May 17, 2025

      How To Fix Largest Contentful Paint Issues With Subpart Analysis

      May 17, 2025

      How To Prevent WordPress SQL Injection Attacks

      May 17, 2025

      Microsoft’s allegiance isn’t to OpenAI’s pricey models — Satya Nadella’s focus is selling any AI customers want for maximum profits

      May 17, 2025

      If you think you can do better than Xbox or PlayStation in the Console Wars, you may just want to try out this card game

      May 17, 2025

      Surviving a 10 year stint in dev hell, this retro-styled hack n’ slash has finally arrived on Xbox

      May 17, 2025

      Save $400 on the best Samsung TVs, laptops, tablets, and more when you sign up for Verizon 5G Home or Home Internet

      May 17, 2025
    • Development
      1. Algorithms & Data Structures
      2. Artificial Intelligence
      3. Back-End Development
      4. Databases
      5. Front-End Development
      6. Libraries & Frameworks
      7. Machine Learning
      8. Security
      9. Software Engineering
      10. Tools & IDEs
      11. Web Design
      12. Web Development
      13. Web Security
      14. Programming Languages
        • PHP
        • JavaScript
      Featured

      NodeSource N|Solid Runtime Release – May 2025: Performance, Stability & the Final Update for v18

      May 17, 2025
      Recent

      NodeSource N|Solid Runtime Release – May 2025: Performance, Stability & the Final Update for v18

      May 17, 2025

      Big Changes at Meteor Software: Our Next Chapter

      May 17, 2025

      Apps in Generative AI – Transforming the Digital Experience

      May 17, 2025
    • Operating Systems
      1. Windows
      2. Linux
      3. macOS
      Featured

      Microsoft’s allegiance isn’t to OpenAI’s pricey models — Satya Nadella’s focus is selling any AI customers want for maximum profits

      May 17, 2025
      Recent

      Microsoft’s allegiance isn’t to OpenAI’s pricey models — Satya Nadella’s focus is selling any AI customers want for maximum profits

      May 17, 2025

      If you think you can do better than Xbox or PlayStation in the Console Wars, you may just want to try out this card game

      May 17, 2025

      Surviving a 10 year stint in dev hell, this retro-styled hack n’ slash has finally arrived on Xbox

      May 17, 2025
    • Learning Resources
      • Books
      • Cheatsheets
      • Tutorials & Guides
    Home»Development»Artificial Intelligence»Last Week in AI #267: Meta releases Llama 3, AI flies real F-16 in dogfight, electric Atlas from Boston Dynamics, Intel’s neuromorphic chip, and more!

    Last Week in AI #267: Meta releases Llama 3, AI flies real F-16 in dogfight, electric Atlas from Boston Dynamics, Intel’s neuromorphic chip, and more!

    April 25, 2024

    Top News

    Meta releases Llama 3, claims it’s among the best open models available

    Meta has launched two new models in its Llama 3 series of open generative AI models, Llama 3 8B and Llama 3 70B, which contain 8 billion and 70 billion parameters respectively. The company claims these models, trained on two custom-built 24,000 GPU clusters, are among the best-performing generative AI models available today. The Llama 3 models have outperformed other models such as Mistral’s Mistral 7B and Google’s Gemma 7B on at least nine benchmarks. The larger Llama 3 model, Llama 3 70B, is also competitive with flagship generative AI models, including Google’s Gemini 1.5 Pro. Meta also claims that the new Llama models offer more “steerability,” a lower likelihood to refuse to answer questions, and higher accuracy on trivia questions, questions pertaining to history and STEM fields such as engineering and science and general coding recommendations, thanks to a much larger dataset of 15 trillion tokens.

    AI Flew X-62 VISTA During Simulated Dogfight Against Manned F-16

    The U.S. Air Force Test Pilot School and the Defense Advanced Research Projects Agency (DARPA) have successfully flown the X-62 VISTA, an aircraft controlled by an Artificial Intelligence (AI) agent, in a simulated dogfight against a human-piloted F-16. This breakthrough, part of DARPA’s Air Combat Evolution (ACE) program, took place in the airspace above Edwards Air Force Base, California, with the AI successfully conducting over 100,000 lines of flight-critical software changes across 21 test flights. The X-62A was flown with safety pilots onboard who had the ability to disengage the AI agent if necessary, but this was not required during the tests. This achievement marks a significant step in the use of AI in flight-critical systems, with potential applications in both commercial and defense aerospace sectors.

    Boston Dynamics’ Robert Playter on the New Atlas

    In a recent interview, Boston Dynamics CEO Robert Playter unveiled the new electric Atlas robot, marking a significant shift from their previous hydraulic models towards a commercially viable product. This new Atlas is designed for diverse industrial applications, including handling heavy objects with complex geometries in manufacturing settings. Playter discussed the extensive development process, emphasizing the necessity of understanding the robot’s potential industrial applications to ensure its commercial success. Unlike earlier models, this Atlas features advanced electric actuators and an extraordinary range of motion, surpassing human capabilities. Playter highlighted the robot’s potential to revolutionize tasks through its strength, flexibility, and efficiency. He also mentioned the design considerations for Atlas’s head, focusing on non-humanlike features to make it appear more approachable and communicative. The interview also touched on Boston Dynamics’ strategic approach to product development and commercialization, leveraging their partnership with Hyundai to optimize manufacturing and market introduction.

    Intel Builds World’s Largest Neuromorphic System to Enable More Sustainable AI

    Intel has developed Hala Point, the world’s largest neuromorphic system, capable of supporting up to 20 quadrillion operations per second with an efficiency exceeding 15 trillion 8-bit operations per second per watt. This system outperforms architectures built on graphics processing units (GPUs) and central processing units (CPUs), and is expected to enable real-time continuous learning for AI applications in various fields, including scientific and engineering problem-solving, logistics, smart city infrastructure management, and large language models (LLMs). Researchers at Sandia National Laboratories plan to use Hala Point for advanced brain-scale computing research, focusing on scientific computing problems in device physics, computer architecture, computer science, and informatics. The development of Hala Point is significant due to the increasing need for sustainable AI solutions, as the scaling up of deep learning models has exposed sustainability challenges in AI, necessitating innovation in hardware architecture.

    Other News

    Tools

    Adobe Premiere Pro is getting generative AI video tools — and hopefully OpenAI’s Sora – Adobe is developing generative AI video tools for Premiere Pro, including integrations with third-party AI models like OpenAI’s Sora, to allow users to generate and manipulate video content using text prompts and extend video clip lengths.

    Slack rolls out its AI tools to all paying customers – Slack has rolled out AI tools to all paying customers, including thread summaries, conversational search, and language support, with the potential to integrate third-party apps in the future.

    Amazon Music’s Maestro lets listeners make AI playlists – Amazon Music’s new AI playlist maker, Maestro, allows users to create playlists using text prompts, including emojis, and share them with others.

    Snap plans to add watermarks to images created with its AI-powered tools – Snap plans to add watermarks to AI-generated images on its platform, using a translucent version of its logo with a sparkle emoji, to ensure transparency and discourage removal.

    Adobe’s PDF-reading AI Assistant starts at $4.99/month – here’s how to try it for free – Adobe’s AI assistant for PDF reading aims to efficiently parse and extract data from PDFs, offering a solution for time-consuming tasks.

    Business

    Baidu claims 200 million users for Ernie chatbot after only 13 months – Baidu’s Ernie chatbot has gained over 200 million users and 85,000 enterprise clients in just 13 months, with the company introducing new tools for developers at its AI developer conference.

    Introducing OpenAI Japan – OpenAI expands to Japan with a new office in Tokyo and a custom GPT-4 model optimized for the Japanese language, aiming to collaborate with local businesses and governments to leverage AI for societal challenges.

    Microsoft Makes High-Stakes Play in Tech Cold War With Emirati A.I. Deal – Microsoft announces a $1.5 billion investment in an AI giant in the UAE, in a deal orchestrated by the Biden administration to counter China’s technological influence, with detailed security arrangements negotiated with the US government.

    Chip Startup Rivos Raises More Than $250 Million to Tap AI Boom – Rivos Inc. secures over $250 million in funding, with Matrix Capital Management as the largest investor, to develop new hardware for artificial intelligence, with support from Intel Capital and MediaTek Inc.

    Collaborative Robotics Locks Up $100M, Latest Robot Startup To Raise Big – Collaborative Robotics, a startup based in Santa Clara, California, raised $100 million in a Series B funding round led by General Catalyst, to develop practical collaborative robots for various industries, contributing to the rise of robotics and AI in the market.

    Adobe’s ‘Ethical’ Firefly AI Was Trained on Midjourney Images – Adobe’s ‘Ethical’ Firefly AI was trained on Midjourney images, despite claims of being a “commercially safe” alternative.

    Google is combining its Android and hardware teams — and it’s all about AI – Google is reorganizing its teams to integrate AI across all its products, with a focus on combining hardware, software, and AI to improve user experiences and stay competitive in the AI revolution.

    Meta, in Its Biggest A.I. Push, Places Smart Assistants Across Its Apps – Meta introduces A.I.-powered smart assistant software across its apps, allowing users to ask for help and information in completing tasks and getting recommendations.

    Google Consolidates Teams With Aim to Create AI Products Faster – Google restructures its teams to accelerate the development of AI products and services.

    Google will provide AI to the military for disaster response – Google is providing AI tools to the National Guard to analyze images of disaster areas and help prioritize its response, marking a shift in the tech industry’s relationship with the military.

    AMD introduces AI chips for business laptops and desktops – AMD has unveiled a new series of semiconductors for AI-enabled business laptops and desktops, aiming to expand its share of the “AI PC” market, with the chips expected to be available in platforms from HP and Lenovo starting in the second quarter of 2024.

    A humanoid robot is on its way from Mobileye founder – Israeli firm Mentee Robotics, founded by the co-founder of Mobileye, is developing a prototype humanoid robot, Menteebot, which aims to integrate computer vision and generative AI for household tasks and learning through imitation, with plans to release a production-ready prototype in 2025.

    Research

    Introducing Idefics2: A Powerful 8B Vision-Language Model for the community – Idefics2 is a powerful 8B vision-language model that takes arbitrary sequences of texts and images as input, generating text responses, and it offers significant improvements over its predecessor, Idefics1.

    Scaling Laws for Fine-Grained Mixture of Experts – Fine-grained Mixture of Experts models outperform dense Transformers and their efficiency gap widens as model size and training budget scale up, with a new hyperparameter, granularity, enabling precise control over the size of the experts.

    🍷 FineWeb – 🍷 FineWeb is a dataset of over 15 trillion tokens of cleaned and deduplicated English web data from CommonCrawl, optimized for LLM performance, and outperforming other high-quality web datasets in model training.

    OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments – Benchmarking multimodal agents for open-ended tasks in real computer environments reveals significant deficiencies in their ability to serve as computer assistants, with the best model achieving only 12.24% success, primarily struggling with GUI grounding and operational knowledge.

    Concerns

    A.I. Has a Measurement Problem – Leading artificial intelligence tools lack standardized testing and evaluation, leaving uncertainty about their true capabilities.

    Microsoft AI creates scary real talkie videos from a single photo – Microsoft’s AI model, VASA-1, can create frighteningly realistic deepfake videos from a single still image and an audio track, raising concerns about the trustworthiness of online content.

    Meta’s Oversight Board probes explicit AI-generated images posted on Instagram and Facebook – Meta’s Oversight Board is investigating how Instagram and Facebook handle explicit, AI-generated images, focusing on cases in India and the U.S., and the broader issues of deepfake porn and online gender-based violence.

    SoA survey reveals a third of translators and quarter of illustrators losing work to AI – AI is impacting creative careers, with a third of translators and a quarter of illustrators losing work to generative AI, leading to concerns about devaluation of work and the need for ethical development and regulation.

    Policy

    NSA Publishes Guidance for Strengthening AI System Security – NSA releases guidance for deploying secure and resilient AI systems, aiming to improve confidentiality, integrity, and availability of AI systems and counter AI vulnerabilities.

    Medium bans AI-generated content from its paid Partner Program – Medium bans AI-generated content from its paid Partner Program, emphasizing the platform’s focus on human storytelling and implementing strict policies against fully AI-generated stories.

    Feds appoint “AI doomer” to run US AI safety institute – Feds appoint “AI doomer” to run US AI safety institute, sparking controversy and concerns about the focus on hypothetical AI risks over current ethical and societal issues.

    Copyright © 2024 Skynet Today, All rights reserved.

    Source: Read More 

    Facebook Twitter Reddit Email Copy Link
    Previous ArticleThis tiny chip can safeguard user data while enabling efficient computing on a smartphone
    Next Article Gosha Geogdzhayev and Sadhana Lolla named 2024 Gates Cambridge Scholars

    Related Posts

    Development

    February 2025 Baseline monthly digest

    May 17, 2025
    Development

    Learn A1 Level Spanish

    May 17, 2025
    Leave A Reply Cancel Reply

    Hostinger

    Continue Reading

    Friction – motion graphics software for vector and raster animations

    Development

    Razer’s Iskur V2 X gaming chair costs less than half the price of its premium sibling

    News & Updates

    Older games dominate Steam as 85% of playtime in 2024 was spent on titles released before this year

    Development

    Google Fixed Cloud Run Vulnerability Allowing Unauthorized Image Access via IAM Misuse

    Development

    Highlights

    CVE-2025-44110 – FluxBB Cross-Site Scripting (XSS) in Forum Description Field

    May 15, 2025

    CVE ID : CVE-2025-44110

    Published : May 15, 2025, 6:15 p.m. | 1 hour, 45 minutes ago

    Description : FluxBB 1.5.11 is vulnerable to Cross Site Scripting (XSS) in via the Forum Description Field in admin_forums.php.

    Severity: 0.0 | NA

    Visit the link for more details, such as CVSS details, affected products, timeline, and more…

    A Coding Implementation of Extracting Structured Data Using LangSmith, Pydantic, LangChain, and Claude 3.7 Sonnet

    March 25, 2025

    Enlightenment 0.27 Released with Bug Fixes, New Modules

    January 13, 2025

    Restic Backup GX offers a simple GUI for restic

    March 25, 2025
    © DevStackTips 2025. All rights reserved.
    • Contact
    • Privacy Policy

    Type above and press Enter to search. Press Esc to cancel.