Blog
Essays, field notes, legal-tech analysis, and how FlowCounsel™ builds and delivers.
Legal Tech
Legal AI keeps forcing a false choice between approachable tools for cautious firms and raw flexibility for builder-heavy teams. Real firms need infrastructure that can serve both.
April 24, 2026
Legal Tech
The lawyer-developer era is real. What it needs next is not less control. It needs the shared infrastructure that makes control durable.
April 24, 2026
Legal Tech
Lawyers should touch the frontier models and tools. They should also understand the failure modes those tools naturally produce in legal work.
April 24, 2026
Legal Tech
A new Stanford paper quantifies a legal-AI failure mode the market still understates: systems that decide when the right answer is that more facts are needed.
April 23, 2026
Legal Tech
Heppner and Fortis make different parts of the legal-AI risk picture visible. Together they show why legal AI confidentiality cannot be evaluated apart from system design.
April 23, 2026
Legal Tech
Sullivan & Cromwell's AI hallucination apology is not just a story about elite-firm embarrassment. It is a systems lesson about tool selection, review states, provenance, and legal AI design.
April 22, 2026
Legal Tech
Human names made early AI agents easier to understand. Legal work needs something stricter: function-named specialists tied to context, source, and review.
April 21, 2026
Legal Tech
AI-assisted coding lets lawyers and other non-engineers ship working tools. Working is not the same as safe. Dependencies, secrets, patch cadence, prompt injection, and threat models still need an owner.
April 19, 2026
Legal Tech
Lawyers are starting to ship AI-assisted tools. The capability is real. The professional floor that applies when legal software touches real work is also real.
April 18, 2026