Thinking out loud.
Honest perspectives on software, AI, and technology decisions. No thought-leadership fluff. Just things we’ve learned building software for businesses that care about getting it right.
Open Source Does Not Mean Open Door
A plain English guide for lawyers on what open source really is, why it is already everywhere, why an open model can be safer than a closed one, and the better questions to ask any vendor.
Why Law Firms Still Need Custom Software Developers in the Age of AI
AI tools are changing software development fast, but law firms still need custom developers who understand their problems, not just their code requirements.
I am dictating this article on my phone. Here is why that matters for your business.
A personal account of how to get significantly more value from AI tools that most businesses already pay for, covering Projects, prompt libraries, voice input with Wispr Flow, phone camera document scanning, and connectors.
Claude Code VS Code Plugin Update
A rundown of the latest Claude Code extension update for VS Code, covering session history, branching, edit modes, effort controls, implementation plans as documents, and more.
AI Tools and Client Data: A Technical Note for Legal Practitioners
A technical follow-up to UKUT 81 (IAC). Explains what open source and public domain mean in their correct technical and legal contexts, sets out the consumer vs enterprise distinction that matters for compliance, covers UK GDPR obligations, legal privilege risk, and how ISO 42001 governance strengthens a firm's position. Co-authored by Naz Keceli and Khiliad Ltd.
AI is changing jobs. Here is where it is happening, where it is not, and what it means for law.
New Anthropic research shows where AI is actually changing work, not just where it theoretically could. The legal sector faces real exposure in document-heavy tasks, even as courtroom work remains protected.
Taalas HC1 Review: The AI Inference Chip That Changes What Fast Means
The Taalas HC1 hardwires Llama 3.1 8B into silicon and delivers up to 17,000 tokens per second. We tested it. Instant inference is real.
AI Readiness in 2026: What Enterprise Failures Teach SMEs About Getting the Basics Right
Enterprise AI failures in 2025 offer a masterclass for SMEs. The winners in 2026 won't move first—they'll move smart.
AI in Software Development: Balancing Innovation with Quality
Exploring how professional developers balance AI coding tools with human expertise to deliver faster custom software without compromising quality.
One email. Once a month. No noise.
The pieces we found genuinely useful, the ones we wrote, and the occasional sharp observation about where software is heading. Built for operators who’d rather skim than scroll.
