Something is Happening

Little by little and then all at once

Over the past couple of months it feels like the ground has shifted, or is starting to shift quickly. AI is far better at things than I anticipated it would be by this point. I am honestly caught a little off-guard.

Consider this, by AI Investor Matt Shumer:

Let me give you an example so you can understand what this actually looks like in practice. I’ll tell the AI: “I want to build this app. Here’s what it should do, here’s roughly what it should look like. Figure out the user flow, the design, all of it.” And it does. It writes tens of thousands of lines of code. Then, and this is the part that would have been unthinkable a year ago, it opens the app itself. It clicks through the buttons. It tests the features. It uses the app the way a person would. If it doesn’t like how something looks or feels, it goes back and changes it, on its own. It iterates, like a developer would, fixing and refining until it’s satisfied. Only once it has decided the app meets its own standards does it come back to me and say: “It’s ready for you to test.” And when I test it, it’s usually perfect.

I’m not exaggerating. That is what my Monday looked like this week.

But it was the model that was released last week (GPT-5.3 Codex) that shook me the most. It wasn’t just executing my instructions. It was making intelligent decisions. It had something that felt, for the first time, like judgment. Like taste. The inexplicable sense of knowing what the right call is that people always said AI would never have. This model has it, or something close enough that the distinction is starting not to matter.

Something Big is Happening

Or this, by Paul Ford in the New York Times:

What might the future look like if 100 million, or a billion, people can make any software they desire? Could this be a moment of unparalleled growth and opportunity as people gain access to tech industry power for themselves?

According to the market, the answer is no. Recently, software stocks — Monday.com, Salesforce, Adobe and many others — plummeted all at once; the Nasdaq 100 lost half a trillion dollars in two days. Legal software company stocks slumped recently because Anthropic released tools to automate some legal work. Financial services firms and real estate services — the market keeps devaluing them because traders expect there to be less need for humans at desks in an A.I.-automated future. Why will anyone need all that legacy software when A.I. can code anything up for you in two shakes of a robotic lamb’s tail?

Personally this all feels premature, but markets aren’t subtle thinkers. And I get it. When you watch a large language model slice through some horrible, expensive problem — like migrating data from an old platform to a modern one — you feel the earth shifting. I was the chief executive of a software services firm, which made me a professional software cost estimator. When I rebooted my messy personal website a few weeks ago, I realized: I would have paid $25,000 for someone else to do this. When a friend asked me to convert a large, thorny data set, I downloaded it, cleaned it up and made it pretty and easy to explore. In the past I would have charged $350,000.

…

All of the people I love hate this stuff, and all the people I hate love it. And yet, likely because of the same personality flaws that drew me to technology in the first place, I am annoyingly excited.

Here is why: I collect stories of software woe. I think of the friend at an immigration nonprofit who needs to click countless times, in mounting frustration, to generate critical reports. Or the small-business owners trying to operate everything with email and losing orders as a result. Or my doctor, whose time with patients is eaten up by having to tap furiously into the hospital’s electronic health record system.

The A.I. Disruption We’ve Been Waiting for Has Arrived (emphasis mine).

But what does any of this have to do with lawyers?

Lawyers have a reputation as perpetual laggards when it comes to adopting new technology. Many firms are still using local servers, paper calendars, and systems that look like a version of MS Access but built by the Soviets in 1987. It’s a stereotype, sure, but every tech-forward lawyer knows a lawyer who still uses [insert horribly inefficient thing] because they simply don’t have time to buy a MS365 license.

Legal technology is cheaper, more accessible, and better than it’s ever been. Moving to a cloud-based practice management system like _______ can cost less than $100 per month per seat. You can conceivably get a small firm onto really nice software pretty easily and economically.

But up till now, the benefit from changing from local servers, paper calendars, and the MS Access Red Scare Edition to better cloud-based software was incremental. The downstream effects felt incremental. An example: a firm that moves from using only individual paper calendars to MS365 and Outlook gains efficiency, but it’s incremental efficiency. Being able to view everyone’s calendars at once and easily schedule appointments and send invites is much better than the hassle of paper and playing phone tag to set a single meeting. But you still have the same people in the same offices for the same 12-hour workday (lawyers, amirite). While people would say it made them more productive, it’s not likely that they would say the gain was 10x productivity.

Three years ago it was impossible to say that AI adoption would 10x productivity with a straight face. Same for 2024, and most of 2025. But now I think it’s actually possible.

The models we have now are far different than the models in 2023. The issue is that many people, especially in law, tried out the free version of ChatGPT in 2023-2024, asked it a question about a legal concept, decided it sucked (it did, let’s be real), and moved on. It couldn’t really do that much that was actually useful. Even if it had produced an A+ legal memo, you’d have to copy and paste the text somewhere and spend so much time fixing the formatting it wasn’t worth it. But that’s not the current state of play.

Over the past 3-4 years, AI has gone from being a Pontiac Aztec with terrible alignment, a timing belt issue, and a rusty catalytic converter to a brand new Mercedes Benz G-Wagon. People see new model releases and think AI is still the same old Aztec after an oil change and maybe a trip to the repair shop. It’s not - the new models and agentic systems are profoundly different, much smarter and more powerful, very capable of actually doing things on a computer or a network, and have very strange abilities.

I am at a point of continual astonishment at the things these systems are capable of. Opus 4.6 is astonishingly good, and, aside from being annoyingly complimentary, has yet to really get things wrong. I’ve started to reach a point where I feel like I could do so much more using Claude Code and Cowork that I feel somewhat guilty not being able to find time to shove the inputs into the system. Is this what being “AI Pilled” feels like? “Better for you if you take me off.”

I’m having conversations with people who also feel that the past months were a palpable acceleration in intelligence and capability. But those same people acknowledge that people they know or work with have zero understanding of where we really are. Meanwhile my LinkedIn feed is now a non-stop AI chum fest of “Exposing the Truth about [insert AI provider here] ” or “AI is not X, it’s Y” engagement bait. The stock market is overreacting to things like a set of markdown files on GitHub (i.e. the Claude Cowork Legal Plugin), or even a Substack post that’s a work of complete fiction about a potential AI-led economic crash two years from now. Perhaps there is too much noise - we’re unable to process the signal.

I can’t say whether Claude Cowork or the Legal Plugin is or is not going to replace the big legal research companies - the real “moat” they have is their database of legal opinions combined with the KeyCite / Shepard’s system. That will continue to be a moat for as long as someone with very deep pockets doesn’t get involved, and I think the market dynamics are such that no one will. But if I was at a SAAS software company that sells a front end on a database, I would be taking a long hard look in the mirror. If I was at an “AI company” whose value proposition was some kind of special industry-related fine-tuning or secret algorithm, I’d want to read up on the Bitter Lesson. Compute always wins.

Moats

Are lawyers immune from this? I think it depends.

A little bird from… somewhere, in Sales, told me that all companies are asking variations of just the same two questions. They bluster and bluff and try to act informed, but they are all terrified. When you cluster their questions, they break down into, “Will everything be OK?” and “Will we be here in five years?”

The default answer, I’m afraid, is No. If you do nothing, you’re almost certainly going to get overrun. If you have an Atom Moat, then you stand a pretty good chance of weathering the storm, if you execute well. Just a chance, mind you: It’s a moat, not a force field. But atoms are a pretty good moat. If you make beer, or work with humans, or ship stuff, say, then you’ve got a bit more time to work with, maybe, to find your feet in the AI era.

The Anthropic Hive Mind - Steve Yegge

Lawyers that work with humans or navigate the court process in person have an “atom moat” - criminal defense, high-dollar litigation, high-stakes family law, or similar practice areas. Lawyers are also protected by a “regulatory moat” - one that says unless a person has a law degree and a bar license, they can’t practice law. But this regulatory moat is only so deep. It doesn’t protect lawyers from companies moving a huge percentage of legal spend in-house, or from insurance companies no longer agreeing to hourly rates engagements on civil defense files, or small businesses from just using Claude to solve their own problems.

I also think there’s a “reputation moat” where the sheer historical value of the name carries enough weight - nobody’s going to get fired for hiring Mayer Brown / Latham & Watkins / Kirkland & Ellis as their outside counsel. But past that, I’m honestly not sure. Susskind has been predicting the “End of Lawyers” for nearly 20 years now, but it still hasn’t happened. Did I mention that the American Arbitration Association is now deploying an AI-based arbitrator?

I do think that we should stop law from being a way of monetizing friction - the billable hour being the most salient example of this - and focus on how lawyers can create value for their clients. Perhaps we can dig a “value moat.” Or maybe the moat is inertia - realtors are still around, after all.

Back to Law Town

The talk about the “End of Lawyers” assumes, arguendo, that the amount of demand for available legal work is a fixed constant. It may not be - the Jevons Paradox is still alive and kicking. But even if the amount potential work is there’s still the issue of an unmet legal market that potentially values in the billions of dollars. Maybe we’re finally seeing the possibility of savvy lawyers creating their own “law factory” to meet previously unmet needs, provide value to those clients, and be able to make enough margin from this for it to be a good business decision.

If I were still practicing law I’d be spending nights and weekends figuring out how to set up Claude Cowork and create practice-specific plugins. I’d be delegating (responsibly) as much as I could to it and streamlining things so that I could drive new business. I’d be figuring out how to be the Law Town mayor.

Like I said in the title - it sure feels like something is happening. Exactly what is happening, well, only time will tell. But if things keep progressing this rapidly, the time might be next year, or even this one.