AI
Artificial intelligence: LLMs, agents, diffusion, vision, NLP and the latest from top labs.
593 published articles
Google DeepMind
Google DeepMind's Gemma 4 turns 26 billion parameters into a reasoning machine that fits on one GPU
Google DeepMind's Gemma 4 technical report details a family of open-weight models with mixture-of-experts, 1M-token context windows, and multi-modal vision. The release signals a strategic play to bring frontier-level reasoning to developers without the cost of proprietary APIs.
2026-07-09
Artificial intelligence
Aleph alpha's new megakernel library cuts moe inference latency by 200%
Alpha-MoE fuses multiple operations into a single persistent kernel to achieve up to 200% inference speed gains over Triton-based kernels in vLLM and SGLang, targeting FP8-precision MoE models.
2026-07-09
Artificial Intelligence
The editor's job is not dead. It just got harder
AI writing assistants are forcing a long-overdue redefinition of what editing means. The role isn't disappearing, it's evolving into a hybrid discipline that pairs human judgment with machine speed.
2026-07-09
Special Report
The specialization revolution: how smaller models are redefining ai's future
From Ornith 9B matching models four times its size to mathematical proofs that specialization is inevitable under finite resources, five recent breakthroughs challenge the dogma that bigger is always better in AI.
2026-07-09
Edge AI
Free GPU, no cloud: this developer just made Hindi LLMs work on a laptop
A developer successfully fine-tuned Google's Gemma 4 for Hindi on a free GPU and shipped the model to run on a CPU. The case study demonstrates practical techniques for low-resource language adaptation and edge deployment.
2026-07-09
Artificial Intelligence
The AI writing generator that made me rewrite this lede three times
AI writing tools are evolving from spell-checkers to co-authors, raising questions about originality, voice, and the value of the human rewrite.
2026-07-09
Artificial Intelligence
DeepSeek-OCR 2 Brings Visual Causal Flow to Open-Source Document Understanding
DeepSeek-OCR 2 introduces Visual Causal Flow for human-like visual encoding. Now open-source on GitHub, the model supports vLLM and Transformers, dynamic resolution (up to 1,216 visual tokens), and document-to-markdown conversion.
2026-07-09
AI Agents
Microsoft's bet on small models for agentic AI is about orchestration, not knowledge
Microsoft Research's MagenticLite project shows that small models can handle complex agentic tasks when orchestration, model design, and execution environment are codesigned. The release includes MagenticBrain, a 14B orchestrator, and Fara1.5, a computer-use model family that nearly doubles prior performance on web navigation benchmarks.
2026-07-09
Model Release
Anthropic Unveils Claude Sonnet 5: A Hybrid Reasoning Model for Real-Time Agents and High-Volume Work
Anthropic's Claude Sonnet 5 delivers hybrid reasoning with a 1M context window, excelling at coding, agentic workflows, and enterprise tasks. Available at an introductory price of $2 per million input tokens, promising multi-day coding projects compressed into hours.
2026-07-09
Web Performance
Chrome's new API could save you 177 MB by letting AI models share cache across websites
Google's proposed Cross-Origin Storage API aims to eliminate redundant downloads of AI models and Wasm runtimes by identifying resources via SHA-256 hash instead of URL. Transformers.js, WebLLM, and wllama are already experimenting with it, with Whisper alone saving over 177 MB per duplicate load.
2026-07-09
Product Launch
Anthropic launches Claude Tag on Slack for team-wide AI collaboration
Anthropic's Claude Tag brings proactive, context-aware AI collaboration to Slack. Teams can tag @Claude in channels and assign tasks ranging from code generation to data analysis. The tool builds shared memory, works asynchronously, and can take initiative, with 65% of Anthropic's product team code now created by the internal version.
2026-07-09
Enterprise AI
Anthropic and DXC Technology Launch Global Alliance to Embed Claude in Enterprise Systems
DXC Technology and Anthropic announce a global partnership to deploy Claude across DXC's managed environments for regulated industries. The deal includes certifying thousands of engineers and using Claude as the default AI in DXC's new OASIS platform.
2026-07-09