By mid-2025 China had become a global leader in open-source large language models (LLMs). According to Chinese state media, by July 2025 China accounted for 1,509 of the world’s ~3,755 publicly released LLMs, far more than any other country. This explosion reflects heavy state and industry investment in domestic AI, open licensing (often Apache- or MIT-style), and a strategic pivot by Chinese tech giants and startups toward publicly shared models. The result is a "revival" of open-source AI, with dozens of Chinese LLMs now available for download or use via Hugging Face, GitHub, or cloud APIs. These range from general-purpose foundation models dozens of billions of parameters in size to specialized chatbots and domain experts, many built on Mixture-of-Experts (MoE) architectures.
Researchers at MIT’s CSAIL are charting a more "modular" path ahead for software development, breaking systems into "concepts" and "synchronizations" to make code clearer, safer, and easier for LLMs to generate.
MIT researchers are proposing a new software development approach centered around "concepts" and "synchronizations" to address issues of complexity, safety, and LLM compatibility in modern software.
Concepts are self-contained units of functionality (like "sharing" or "liking") with their own state and actions, whereas synchronizations are explicit rules defining how these concepts interact, expressed in a simple, LLM-friendly language.
The benefits include ncreased modularity, transparency, easier understanding for both humans and AI, improved safety, and potential for automated software development. Real-world application: has been demonstrated by successfully restructuring features (liking, commenting, sharing) to be more modular and legible.
Future includes concept catalogs, a shift in software architecture, and improved collaboration through shared, well-tested concepts.
MIT researchers are proposing a new software development approach centered around "concepts" and "synchronizations" to address issues of complexity, safety, and LLM compatibility in modern software.
Concepts are self-contained units of functionality (like "sharing" or "liking") with their own state and actions, whereas synchronizations are explicit rules defining how these concepts interact, expressed in a simple, LLM-friendly language.
The benefits include ncreased modularity, transparency, easier understanding for both humans and AI, improved safety, and potential for automated software development. Real-world application: has been demonstrated by successfully restructuring features (liking, commenting, sharing) to be more modular and legible.
Future includes concept catalogs, a shift in software architecture, and improved collaboration through shared, well-tested concepts.
Hierarchical Reasoning Model (HRM) is a novel approach using two small neural networks recursing at different frequencies. This biologically inspired method beats Large Language models (LLMs) on hard puzzle tasks such as Sudoku, Maze, and ARC-AGI while trained with small models (27M parameters) on small data (around 1000 examples). HRM holds great promise for solving hard problems with small networks, but it is not yet well understood and may be suboptimal. We propose Tiny Recursive Model (TRM), a much simpler recursive reasoning approach that achieves significantly higher generalization than HRM, while using a single tiny network with only 2 layers. With only 7M parameters, TRM obtains 45% test-accuracy on ARC-AGI-1 and 8% on ARC-AGI-2, higher than most LLMs (e.g., Deepseek R1, o3-mini, Gemini 2.5 Pro) with less than 0.01% of the parameters.
An Apple study shows that large language models (LLMs) can improve performance by using a checklist-based reinforcement learning scheme, similar to a simple productivity trick of checking one's work.
This blog post details the training of 'Chess Llama', a small Llama model designed to play chess. It covers the inspiration behind the project (Chess GPT), the dataset used (Lichess Elite database), the training process using Huggingface Transformers, and the model's performance (Elo rating of 1350-1400). It also includes links to try the model and view the source code.
Andrej Karpathy discusses the transformative changes in software development driven by large language models (LLMs) and artificial intelligence, comparing the current era to the early days of computing. The article details Software 3.0 as the latest evolution in software development paradigms, where LLMs are programmable systems that interpret natural language prompts.
The article explores the concept of consciousness, arguing for a reductionist approach by Alan J. McComas. It discusses how consciousness is a function of the brain, with neural activities often preceding awareness. The piece also examines the role of large language models in chatbots, highlighting potential manipulative techniques and the need for vigilance in interactions. Additionally, it addresses the integration of AI in research and evolving journal standards.
Creativity and a Jetson Orin Nano Super can help hobbyists build accessible robots that can reason and interact with the world. The article discusses building a robot using accessible hardware like Arduino and Raspberry Pi, eventually upgrading to more capable hardware like the Jetson Orin Nano Super to run a large language model (LLM) onboard.
Researchers tested large language models (LLMs) and humans on a comprehensive battery of theory of mind tasks, revealing differences in their performance on tasks such as understanding false beliefs, recognizing irony, and identifying faux pas.
In this paper, the authors discuss the challenges faced in developing the knowledge stack for the Companion cognitive architecture and share the tools, representations, and practices they have developed to overcome these challenges. They also outline potential next steps to allow Companion agents to manage their own knowledge more effectively.
