Three Pillar Articles: AI, Practice Evolution, and the Engineering of Law

Article 1: What AI Actually Changes About Being a Lawyer — and What It Doesn't

The question isn't whether AI will change law. It will. The sharper question is what it actually changes, and what stays fixed.

For decades, the lawyer's core task has been creation under constraint: drafting a motion, structuring a deal, building an argument from precedent and statute. You start with a blank page. You fill it with knowledge, judgment, and craft. The time you spend is the value you deliver. This is the industrial model of legal work—input hours, output documents.

AI doesn't eliminate that task. It collapses it.

Think of it as a phase transition in physics. You don't get a slightly faster version of the old process. You get a fundamentally different state. A first draft that once took six hours now takes six minutes. Not because the lawyer got faster. Because the lawyer didn't have to start from zero. The machine generated the scaffold. The lawyer verifies, edits, and refines.

This is the real shift: from creation to verification. Richard Susskind has been signaling this for years—the future of law isn't about doing legal work faster; it's about doing less of it, differently. AI accelerates us toward that future.

What changes:

• The time-value equation breaks. You can't bill six hours for work that now takes one hour to verify. Firms that cling to hourly billing will feel this acutely.
• The bottleneck moves from production to judgment. Can you spot what the AI missed? Can you know when to trust it and when to push back? That becomes the differentiator.
• Leverage inverts. A solo practitioner with AI and judgment becomes more competitive than a junior associate with neither.
• Knowledge work becomes more like quality assurance. You're checking, testing, validating—the same rigor an engineer applies to code review.

What doesn't change:

• The need for human judgment. AI generates plausible outputs. It doesn't know the judge, the opposing counsel, the client's risk tolerance, or the nuance of when a formal motion matters versus when a phone call does. Those calls remain human.
• Accountability. You still sign the brief. The AI doesn't.
• The craft of listening. Understanding what a client actually needs, beneath what they ask for, is still a human skill. AI can't do it.
• Strategic thinking. AI can help you see the options. You decide which path fits the client's goals and constraints.

Ethan Mollick's research on AI and expertise is instructive here: AI doesn't replace expertise; it changes what expertise means. An expert lawyer with AI isn't someone who can hold the entire tax code in their head. It's someone who knows what questions to ask, when to verify, and how to integrate AI output into sound judgment.

The gap between time spent and value delivered collapses. That's the real change. For a century, legal work scaled with hours. Now it doesn't. Firms that adapt their business model—moving away from hourly billing, toward value-based or fixed-fee work—will thrive. Firms that try to preserve the old model will find themselves caught between the economics of AI-assisted work and the psychology of clients who no longer accept six hours for a task that takes one to verify.

The lawyer's job doesn't disappear. It transforms. The craft shifts from making something from nothing to making sure what's made is sound. That's not a diminishment. It's a reorientation. And it rewards practitioners who can think like editors, quality auditors, and strategists—not just writers.

Article 2: The Virtual Firm Was the Rehearsal. AI Is the Performance.

In 2020, law firms had an unplanned experiment forced upon them. Remote work wasn't a choice; it was a mandate. Overnight, firms learned whether they could function without a physical office, whether client communication could happen over video, whether collaboration could survive asynchronous tools.

Some firms failed the test. Others thrived. The ones that thrived didn't just survive remote work—they engineered it. They rebuilt their workflows, their document systems, their client intake, their billing. They moved from paper to cloud. They standardized their processes. They learned to measure output instead of seat time.

That wasn't the performance. That was the rehearsal.

AI adoption requires the same infrastructure. You can't bolt AI onto a firm that still faxes documents, still bills in six-minute increments, still measures productivity by hours logged. The firms positioned to adopt AI aren't the ones with the biggest budgets. They're the ones that already did the hard work of digitization and process standardization during the remote transition.

This is where small and solo firms hold a decisive advantage.

A 200-person firm that went remote in 2020 had to retrain staff, migrate legacy systems, and convince partners that the old way wasn't coming back. That was painful and expensive. A 5-person firm made the same shift in a weekend. They already had less institutional inertia. They could change their billing model without a partner vote. They could adopt a new tool without IT approval committees. They could experiment.

Agility in practice structure is the real moat for AI adoption.

Here's the engineering parallel: in complex systems, the cost of change scales with the number of interdependencies. A small firm has fewer moving parts. Fewer people to retrain. Fewer legacy contracts to renegotiate. Fewer stakeholders who benefit from the status quo. That's not a weakness when the environment shifts. It's an advantage.

What this means in practice:

• Billing model flexibility. A solo practitioner can move from hourly to value-based work tomorrow. A 100-partner firm needs a task force. The small firm can price AI-assisted work competitively because they don't carry the overhead of the old model.
• Tool adoption spee