News/Open-source
Hugging Face’s new open-source AI model lets robots follow verbal commands
Companies Hugging Face and Physical Intelligence have launched Pi0, a groundbreaking open-source foundational model that enables robots to translate natural language commands directly into physical actions. The breakthrough explained: Pi0 represents the first widely available foundation model for robots that can understand and execute verbal commands, similar to how ChatGPT processes text. The model operates on Hugging Face's LeRobot platform and can handle complex tasks like folding laundry, bussing tables, and packing groceries Pi0 was trained using data from seven different robotic platforms across 68 unique tasks The technology employs flow matching to generate smooth, real-time action trajectories at 50Hz,...
read Feb 7, 2025The real lesson of DeepSeek, according to The Atlantic
China's AI firm DeepSeek made headlines in January 2025 with an AI model that achieved impressive results using fewer resources than industry standards, but subsequent allegations of training data misuse have complicated the narrative. The breakthrough and initial reaction: DeepSeek's announcement of its new AI model sparked significant reactions across the global tech landscape. Wall Street responded with a downturn in tech stocks Chinese commentators celebrated it as evidence of China surpassing U.S. technological capabilities The development raised concerns in Washington about America's competitive position in AI The controversy unfolds: OpenAI launched an investigation into DeepSeek's alleged misuse of ChatGPT...
read Feb 6, 2025China is openly sharing its AI — should the West be locking it out?
The launch of Chinese firm DeepSeek's open-source AI model has sparked intense debate about AI development and national security, particularly regarding U.S.-China technological competition. Key developments: DeepSeek's latest AI model, which rivals U.S. closed-source systems, has become globally popular while simultaneously promoting Chinese Communist Party (CCP) viewpoints. The model has become the most downloaded app globally despite promoting One China policy and avoiding discussion of sensitive topics like Uyghur genocide The software is freely available and modifiable, allowing developers to study and alter its behavior Within days of release, developers created over 500 variations of the model, generating five times...
read Feb 5, 2025India builds first open-source audio language model using Llama
Sarvam AI has developed India's first open-source audio language model, Shuka v1, by integrating Meta's Llama model to process voice queries across multiple Indian languages. Project overview: Shuka v1 represents a significant breakthrough in multilingual audio comprehension, combining Llama's language processing capabilities with a custom audio encoder to handle voice interactions in ten Indian languages. The system utilizes Llama as a decoder to process audio tokens generated by Sarvam's proprietary audio encoder Shuka v1 can accurately interpret and respond to voice queries in languages including Gujarati, Hindi, Kannada, and Marathi The open-source nature of the model allows government departments and...
read Feb 4, 2025The EU is launching a major collaboration to develop open-source AI models
The European Union is launching a major collaborative project to develop open-source large language models, bringing together 20 leading research institutions and companies across the continent. Project Overview: The OpenEuroLLM initiative aims to create multilingual AI language models that are both performant and compliant with European regulations and values. The project is coordinated by Jan Hajič from Charles University (Czechia) and co-led by Peter Sarlin from AMD Silo AI (Finland) Work officially begins on February 1st, 2025, with funding from the European Commission's Digital Europe Programme The initiative has been awarded the STEP (Strategic Technologies for Europe Platform) seal Key...
read Feb 4, 2025Red Hat’s approach to open-source AI favors pragmatism over utopian dreams
Red Hat advocates for a balanced, practical approach to open-source AI that acknowledges both the opportunities and limitations of applying traditional open-source principles to artificial intelligence systems. Key context: Red Hat, a leading enterprise open-source software company, is tackling the complex challenge of defining and implementing open-source principles in artificial intelligence development. The company acknowledges significant ambiguity around what "open-source AI" actually means, particularly given the unique characteristics of AI systems compared to traditional software Traditional open-source concepts face new challenges when applied to AI systems, where the definition of "source code" becomes less clear Red Hat has taken a...
read Feb 2, 2025Shutterstock-Lightricks partnership offers example of how to ethically source training data for AI video
Shutterstock's innovative "research license" model with Lightricks marks a significant shift in how AI companies can legally and ethically access training data, potentially making high-quality datasets more accessible to startups and smaller developers. The groundbreaking partnership: Shutterstock and AI creative technology company Lightricks have established a new licensing framework that allows AI companies to access training data through a graduated approach. Lightricks will train its open-source video generation model LTXV using Shutterstock's HD and 4K video library The model enables companies to begin with a smaller research license for testing before upgrading to commercial licenses This approach directly addresses the...
read Feb 1, 2025Sam Altman admits OpenAI’s closed-source strategy may have been misguided
Sam Altman, OpenAI's CEO, acknowledged during a Reddit AMA that his company's closed-source AI strategy may have been misguided, particularly as competition from open-source models intensifies. Key developments: OpenAI's stance on open-source AI development appears to be shifting following market disruption from Chinese AI firm DeepSeek. Altman directly stated "we have been on the wrong side of history" regarding open source strategy, though he noted this view isn't universally shared within OpenAI The company is actively discussing the possibility of releasing model weights This represents a significant departure from OpenAI's recent proprietary approach to AI development Market impact and competitive...
read Feb 1, 2025Why you may be all wrong when it comes to DeepSeek
DeepSeek, a Chinese AI company, has captured widespread attention with its latest AI model release, reaching the top spot on the iPhone App Store and generating significant discussion in the tech industry. Key developments: DeepSeek's R1 model has gained remarkable popularity despite not being the highest-performing AI model currently available. The model ranks fourth on the chatbot arena leaderboard, behind Google's Gemini DeepSeek has achieved significant technical breakthroughs in model efficiency The model's rapid rise to prominence has sparked discussions about US-China AI competition User interface innovation: DeepSeek's success largely stems from its unique approach to displaying AI reasoning processes....
read Feb 1, 2025What the headlines on DeepSeek are missing, according to RAND
The DeepSeek AI company has achieved significant technical progress while operating under U.S. export controls on advanced AI chips to China, demonstrating both efficiency gains and limitations in the current regulatory landscape. Key developments: DeepSeek has managed to train advanced AI models using Nvidia H800 chips, which were specifically designed to comply with initial U.S. export controls. The company trained its V3 model using 2,000 H800 chips, showing impressive efficiency DeepSeek previously operated Asia's first 10,000 Nvidia A100 cluster and reportedly maintains 50,000 "Hopper" chips The timing of their R1 model release coincided with President Trump's inauguration, potentially for strategic...
read Jan 31, 2025Mistral AI launches small, local and open-source alternative to GPT-4o mini
Mistral AI has released Small 3, a 24B-parameter open-source language model designed to run locally while delivering performance comparable to larger proprietary models. Key features and capabilities; Small 3 represents a significant advancement in efficient, locally-deployable language models that can operate with minimal computing resources. The model can run on a MacBook with 32GB RAM, making it accessible for individual developers and small organizations Built with fewer layers than comparable models to optimize for speed and latency Achieved over 81% accuracy on the MMLU benchmark test without using reinforcement learning or synthetic data Released under the Apache 2.0 license, allowing...
read Jan 29, 2025Ex-Google, Apple engineers unveil Oumi AI, a truly open-source AI development platform
Oumi, a new AI platform developed by former Google and Apple engineers, has launched with $10 million in seed funding to provide fully open-source access to AI model development tools. The platform's core offering: Oumi provides comprehensive access to AI model code, weights, and training data, backed by a consortium of 13 leading research universities. The platform delivers a complete toolkit for building, evaluating, and deploying foundation models, supporting a wide range of parameters from 10M to 405B Advanced training capabilities include Support Fine-Tuning (SFT), Low-Rank Adaptation (LoRA), Quantized LoRA (QLoRA), and Direct Preference Optimization (DPO) The system accommodates both...
read Jan 29, 2025DeepSeek’s new image generator is another win for cost-effective AI
DeepSeek, a Chinese AI startup, has released Janus-Pro, a new open-source text-to-image AI model that claims to outperform established competitors like Stable Diffusion and DALL-E. Key Features and Capabilities: The Janus-Pro model family ranges from 1 billion to 7 billion parameters and operates using an autoregressive framework for image generation and analysis. The model is available under an MIT license, making it suitable for commercial use Users can download Janus-Pro through HuggingFace and GitHub platforms Smaller versions of the model are limited to analyzing images at 384 x 384 resolution Performance and Benchmarks: DeepSeek's internal testing shows promising results for...
read Jan 29, 2025AI architecture innovation: What’s really driving DeepSeek’s success
DeepSeek has made a remarkable advancement in artificial intelligence efficiency with their v3 model, achieving state-of-the-art performance while consuming only 2.8 million H800 hours of training time—dramatically less computational resources than comparable models. This achievement challenges the industry's typical approach of scaling up computational power to improve performance, demonstrating that strategic architectural innovations can deliver superior results with greater efficiency. Through sophisticated improvements like Multi-head Latent Attention (MLA) and enhanced expert systems, DeepSeek v3 represents a significant step forward in the field of language model development, suggesting that thoughtful design optimization may be more valuable than raw computational power in...
read Jan 29, 2025DeepSeek R1 vs DeepSeek V3: Which is better at coding?
Testing methodology and scope: A comprehensive evaluation of DeepSeek's V3 and R1 models was conducted by a journalist at ZDNET using four established coding challenges that have previously been used to benchmark other AI models. The testing framework included writing a WordPress plugin, rewriting a string function, debugging code, and creating a complex automation script Both V3 and R1 variants were evaluated against identical criteria to ensure consistent comparison The assessment focused on code accuracy, functionality, and practical implementation Performance breakdown: DeepSeek V3 emerged as the stronger performer, successfully completing three out of four challenges while R1 managed two successful...
read Jan 29, 2025SLIMA Kashif is a new open-source AI model designed specifically for Arabic
SILMA Kashif 2B Instruct v1.0 is a new bilingual AI model specifically designed for Arabic and English retrieval-augmented generation (RAG) tasks, with a primary focus on question answering and secondary capabilities in entity extraction. Core capabilities and architecture: The model is built on Google Gemma's foundation and operates within the 3-9 billion parameter range, featuring a 12,000-token context window for processing large amounts of text. The model excels at answering questions in both Arabic and English languages It processes both short snippets and lengthy passages effectively The system can provide both concise and detailed responses based on context Entity extraction...
read Jan 28, 2025The biggest winner of the DeepSeek shakeup may be open-source AI
The rise of DeepSeek-R1, an AI model created by Chinese company DeepSeek at a fraction of traditional costs, marks a significant shift toward open-source AI dominance in the technology landscape. The breakthrough explained: DeepSeek-R1 has achieved impressive performance while costing only $6 million to develop, compared to the billions spent by major tech companies on their proprietary models. The model builds upon open-source foundations, including Meta's Llama models and the PyTorch ecosystem Meta's chief AI scientist Yann LeCun emphasized that this development demonstrates the growing superiority of open-source models over proprietary ones The cost efficiency has sent shockwaves through the...
read Jan 28, 2025How to sign up for DeepSeek
DeepSeek, a Chinese AI startup, has rapidly risen to become a top-downloaded app on Apple's App Store in the US and UK, offering a competitive alternative to ChatGPT with advanced natural language processing and image generation capabilities. What you need to know: DeepSeek represents a new entrant in the AI chatbot space that has quickly gained prominence, particularly during recent ChatGPT outages. The platform offers both browser-based and mobile app access Recent additions include image generation features, positioning it as a competitor to Midjourney and DALL-E The service allows for document uploads and various integration options Browser-based access instructions: Accessing...
read Jan 28, 2025DeepSeek is pretty good at coding, but here’s where it still falls short
In an increasingly crowded field of AI coding assistants, DeepSeek AI has emerged from China as a surprisingly capable contender, demonstrating strong programming abilities while operating with notably less computational overhead than its major competitors. The open-source chatbot's success in handling complex coding challenges - achieving a 75% success rate across rigorous tests - while maintaining efficient resource usage suggests a potential shift in how we think about the infrastructure requirements for advanced AI systems. Core performance assessment: DeepSeek R1 underwent four rigorous coding tests designed to evaluate its programming capabilities across different scenarios. The AI successfully completed a WordPress...
read Jan 28, 2025Block’s new open-source AI agent does everything from writing code to ordering dinner
Block's new open-source AI agent 'codename goose' has debuted with the ability to write code, handle daily tasks, and adapt capabilities mid-session through a flexible connection framework. Key features and capabilities; The new AI agent, released under Apache License 2.0, offers seamless interoperability between user interfaces, language models, and various systems through Anthropic's Model Context Protocol (MCP). Users can specify their preferred large language model (LLM) and add new tools during active sessions The agent can autonomously execute tasks including writing code, running tests, and managing dependencies Goose demonstrates impressive efficiency, with the ability to generate 70% of its own...
read Jan 28, 2025How DeepSeek capitalized on American ‘innovation debt’
DeepSeek, an artificial intelligence company, has gained attention for its cost-effective AI advancements built upon the foundational work of major tech companies. The innovation foundation: DeepSeek's development relies heavily on previous breakthroughs from industry giants like Google, OpenAI, Meta, and Nvidia, demonstrating how smaller companies can leverage existing technology to create refined solutions. Meta's open-source LLaMA models likely provided essential blueprints for DeepSeek's development approach Google's transformer architecture forms the basis of DeepSeek's Mixture-of-Experts model OpenAI's reinforcement learning advances enabled DeepSeek's reasoning capabilities Nvidia's GPU technology and low-precision training innovations made cost-efficient scaling possible Technical achievements: DeepSeek has made notable...
read Jan 27, 2025What is DeepSeek, the Chinese AI app taking the world by storm?
Chinese AI startup DeepSeek has launched a new AI model called R1 that has quickly risen to prominence, causing market volatility and raising questions about U.S. technological dominance. The breakthrough development: DeepSeek's R1 model, developed for just $5.6 million, has demonstrated capabilities comparable to leading U.S. AI models that cost billions to develop and maintain. The one-year-old startup's app surged to the #1 position on Apple's App Store, surpassing OpenAI's ChatGPT The company achieved its results using reduced capability Nvidia chips, demonstrating remarkable efficiency Meta CEO Mark Zuckerberg recently announced plans to invest $60-65 billion in AI development this year,...
read Jan 27, 2025DeepSeek claims its newest AI image model Janus-Pro surpasses Stable Diffusion, DALL-E 3
DeepSeek, a Chinese AI startup, has launched Janus-Pro, a new open-source AI image model that claims superior performance over established competitors like Stable Diffusion and DALL-E 3, despite resolution limitations. Key features and capabilities: Janus-Pro represents DeepSeek's latest entry into the competitive AI image generation market, with specific technical constraints and advantages. The model currently operates at a maximum resolution of 384x384 pixels, which is relatively small compared to other image generation tools The flagship version, Janus-Pro-7b, has demonstrated superior performance on two AI benchmark tests The model is being released as open-source, making it accessible to developers and researchers...
read Jan 27, 2025DeepSeek: What to know about the hot new open-source Chinese AI model everyone’s talking about
In a remarkable development that challenges conventional assumptions about AI development, Chinese technology DeepSeek has unveiled DeepSeek R1, an open-source AI model achieving performance levels comparable to industry leaders at a fraction of the typical cost and development time. Developed in just two months with a modest investment of $5.6 million, DeepSeek R1 demonstrates strong reasoning abilities and specialized capabilities, particularly in programming applications, while making its technology freely available to researchers and developers worldwide. This breakthrough raises important questions about the future landscape of AI development and the sustainability of premium-priced proprietary AI services. Key specifications and capabilities: DeepSeek...
read