Tencent releases Hy3, a 295B-parameter MoE model outperforming larger open-source models, available for free on OpenRouter.
A developer explores automating coding tasks using AI agents to reduce personal involvement in routine work.
Improving LLM creative writing quality by modifying the sampling process using future entropy information.
Announcement of the release of 'Build a Reasoning Model (From Scratch)', a book on implementing modern reasoning techniques for AI.
A tech professional critiques anti-AI bias in coding, defending AI tools like LLMs while acknowledging their flaws and advocating for balanced adoption.
Tutorial on setting up a local coding agent using open-weight LLMs and open-source tools as an alternative to Claude Code and Codex.
An open source maintainer argues vulnerability reports are no longer special due to LLMs making security insight widely accessible.
Analysis of Graze's co/core AI cooperative and the niche future of LLMs, with skepticism about tech oligarchy.
Martin Fowler shares thoughts on LLMs making programming more fun, DDD Europe highlights, and conversation registers with AI.
Explains the key differences between microservices and AI agents, highlighting broken assumptions and security risks.
How LLMs like Claude inspired a developer to add a Talks page to his personal site after ChatGPT hallucinated a feature.
Explores how large language models automate reification, turning abstract concepts into perceived reality through AI-generated explanations.
A writer explains how AI-assisted writing provides access and voice, not slop or cheating, using personal essays as evidence.
Reflection on LLM-built projects and the question 'Now what?' after creation, focusing on purpose and ethics.
A reflective analysis on whether AI tools truly boost productivity or just create a sense of performative busywork in software development.
Analysis of Nemotron 3 Ultra, a 550B parameter LLM using Latent MoE scaling and hybrid Mamba-Transformer architecture.
Explores the argument that AI is 'just predicting tokens' and demonstrates understanding through novel murder mysteries with fake physics.
Microsoft announces MAI-Thinking-1 and MAI-Code-1-Flash LLMs, with details on parameters, licensing, and training data.
Microsoft announces MAI-Thinking-1 and MAI-Code-1-Flash LLMs, with low active parameters and claims of performance, but training data issues remain.
A plain English guide explaining how LLMs like ChatGPT actually work, covering prediction, vectors, embeddings, and common misconceptions.
