Technical Articles

Daniel
Dewhurst

Writing about AI systems, developer tools, Laravel, TypeScript, and the craft of building software.

Topics

GPT-5.6 Sol vs Kimi K3 vs Claude Opus 5: an early coding comparison

GPT-5.6 Sol, Kimi K3, and Claude Opus 5 trade coding wins. An early comparison of capability, cost, availability, and agent workflows.

26 Jul 2026 12 min read ai llm developer-tools

Do token-saving plugins make AI coding agents worse?

Skills like Caveman and tools like RTK save tokens, but their benchmarks rarely measure whether the agent still completes the work correctly.

24 Jul 2026 12 min read ai developer-tools claude-code

DiffusionGemma: Google's open model that writes text in parallel

Google DeepMind shipped an open-weight model that generates text by denoising a block of tokens instead of predicting one at a time. Here's how it works and where it fits.

12 Jun 2026 10 min read ai machine-learning llm

GPT-5.5 turns autonomous cyber capability into a trend

AISI's evaluation puts GPT-5.5 beside Claude Mythos on advanced cyber tasks. The response is faster patching, better logs, and stricter access.

12 Jun 2026 8 min read ai openai ai-safety

Axocoatl: multi-agent coordination without the orchestrator

A first look at Axocoatl, a new Rust runtime where persistent AI agents coordinate through pheromone-style signals instead of a central orchestrator.

11 Jun 2026 10 min read ai rust open-source

The AI coding workflow is getting simpler

The best AI coding results come from telling the model what you're trying to build, then steering the conversation until the shape is right.

31 May 2026 13 min read ai developer-tools openai

View all 34 articles →