Jamie Tso

Founder @legalquants | ex-Clifford Chance

[Jamie Tso @jamietsohn]

Spent the last two weeks in SF and NYC talking to AI native law firms, GCs and academics. Here are my observations:

1. Part of the AI native law firm thesis is to build proprietary software that beats Harvey/Legora. But their most cracked lawyers that I have talked to won't even use the internal tool. They prefer their own harness on Claude Code or Codex.

2. Some people are convinced that certain AI native law firms were built just to become RL environments that generate real dataset for the labs.

3. Ex-BigLaw lawyers at AI natives miss working on headline deals (the ones that get you into Legal500 and Chambers & Partners).

4. BigLaw made its money running billion-dollar cross-border deals with multiple parties that agents can't run unsupervised. No AI native firm is doing that sort of work. BigLaw solved it by sorting out multi-jurisdictional licensing and regulatory issues across offices, with precise staffing. AI can potentially do the staffing part really well, but not the licensing part.

5. No AI native firm has a hiring strategy for 10x lawyers. They prefer hiring partners as their salesforce. The best 10x lawyers I know are still in BigLaw. AI natives need software-like margins so they focus on commoditized work that a model + purpose built harness can finish autonomously. And 10x lawyers enjoy doing non-routine, non-automatable tasks in the fastest way possible using AI.

6. The next frontier, if I am right, is full stack lawyers, lawyers who spot a new area of emerging practice and build productized service around that to monopolize a segment quickly (essentially turn your advice into an API that clients can run by BYOA). It's cheaper and easier than hiring partners to buy their client relationships. Specialist practices like tax, ERISA and regulatory have best chance of winning here.

[Jamie Tso @jamietsohn]

@theviclaw Legal seems to be the only vertical where even the frontier labs need a dedicated GTM team to succeed.

[Jamie Tso @jamietsohn]

@felixringe Great work Felix

[Jamie Tso @jamietsohn]

@Ocean42668833 Yeah… but at least most of these programs won’t make you sign full on non compete so you can still ‘potentially’ double dip but you need to give them priority to execute the trades.

[Jamie Tso @jamietsohn]

RT @techie_ray: "Legal Quants-style group for builders"

Feels good when LegalQuants itself has become a standard that is followed by Anthr…

[Jamie Tso @jamietsohn]

Ex-hedge fund lawyer here. Helped set up these multi-manager multi-strategy hedge funds. There are a few distinctions here:

1. These platforms achieve economies of scale by maintaining a central ‘book’ that aggregates and offsets positions coming from all pods so they don’t pay duplicative fees. The best multi-strat funds have an algo that allocates capital very efficiently across multiple pods as a result. It’s similar to BigLaw’s ‘staffing model’ just that BigLaw never had a sophisticated algo to deploy associate capacity.

2. They prune extremely quickly if a PM doesn’t perform, usually by experiencing a sharp drawdown. It’s called a ‘stop-out’. The PM walks out the door quickly. BigLaw pruning isn’t as brutal once you become an EP.

3. They have strict non compete non solicitation on their PMs. In fact the platform’s anchor investors usually mandate the platform to include those terms into PMs’ contracts to protect the platform’s value, and if the PM sets up their own shop, they must let the platform take a percentage of the new shop’s equity. Whereas the US ABA bans non-compete (Rule5.6)

4. These platforms don’t always hire the PMs fully. Some platforms like Citadel and Point72 just buy strategies/signals from smaller managers (this is called ‘alpha capture’ and deploy their own capital to pursue those opportunities and split profit with those managers. That way they don’t maintain the overhead. We seldom hear BigLaw licensing some lawyer’s templates/knowhow without hiring them.

5. These platforms mostly scale by deploying capital efficiently so PMs don’t need to set up their own fund. In BigLaw, you can’t really deploy work or relationship the same way so eventually it becomes winners take all rather than fully performance based.

In short, these platforms have much more durable value than BigLaw and can survive the departure of rainmaker PMs.

[Will Chen @willchen500]

Once had a friend who worked in hedge funds describe to me the multi-manager “pod shop” structure.

Each portfolio manager in the hedge fund is assigned capital and analysts and they pursue their own strategies independently. The fund handles the centralised risk management and operations, technology and analytics.

The structure enables the diversification of risk and pooling of shared resources.

Modern Biglaw seems pretty similar in practice.

Firms diversify across practices to reduce risk e.g. a restructuring practice is a great hedge for a downturn affecting PE and M&A. Each equity partner runs a team that they feed with their “book” that is semi-autonomous in a department. They do however cross collaborate and work with specialist lawyers who are a shared resource (and of course bring in their own work). And it’s much less brutal and volatile.

[Jamie Tso @jamietsohn]

@caliafp @willchen500 They also have very strict non compete and the platform also has first dip on equity if the PM is setting up their own shop.

[Jamie Tso @jamietsohn]

@john_law_kelly @dmandes1 Ha dw it’s a good discussion.

[Jamie Tso @jamietsohn]

Cannot agree more with Zack here. Just choose a frontier platform like Claude Code or Codex, enable browser/computer use, teach your lawyers how to max out on them by learning parallelisation, html and meta prompting and build a dozen of non-intrusive skills/plugins that help with their day to day work. Only when your team is staffed with 10x lawyers, you hire software engineers to distill the scalable components by running evals/ benchmarks that find out the best setup. I used to work in funds, something like this shouldn’t take more than a team of 3-4 to build. Bottleneck is still talent.

[Zack Shapiro @zackbshapiro]

There's a decent chance that law firms commissioning expensive, bespoke AI software today will spend more time negotiating the vendor contracts than their competitors will spend recreating the software a couple of years from now.

AI is already being used to reverse-engineer sophisticated commercial software that took years to develop. It's entirely plausible that by 2028, a reasonably competent associate could describe what one of these elaborate fund formation platforms does to Claude, point it at the firm's precedents, and have a functional substitute by the end of the week.

I'd be awfully nervous underwriting a major software development project right now when the cost of reproducing its functionality may be falling faster than you can finish building it.

Instead, law firms should be teaching their lawyers to build inexpensive, relatively simple automations for the problems they actually encounter. Have the lawyers use their own judgment, experience, and the firm's proprietary data to teach AI how to perform the work the way they would. And be perfectly happy to throw those tools away and rebuild them as the models improve.

The competitive advantage isn't going to be the software. It's going to be how effectively a firm's lawyers can teach AI to do what they do. Spending millions on bespoke software that might be trivial to replicate before it pays for itself seems like a particularly bad bet right now.

[Jamie Tso @jamietsohn]

@john_law_kelly @dmandes1 Totally. Power users are those who know the ceiling of the tool, know when to use AI and not not to and what to check.

[Jamie Tso @jamietsohn]

@BTenmann It’s not.

[Jamie Tso @jamietsohn]

@john_law_kelly @dmandes1 The average Harvey user can still breach SB574 if they don’t know when and what to verify. It’s a skill issue.

[Jamie Tso @jamietsohn]

@eddienasser Partner who knows exactly what the best possible client experience looks like, associate who has been through the grind and has a proven track record of building for their colleagues, and data engineer for system integration. All AI native by default.

[Jamie Tso @jamietsohn]

@dmandes1 @john_law_kelly OpenAI offers ZDR. It’s explicitly mentioned in the Astra for Law launch.

[Jamie Tso @jamietsohn]

@john_law_kelly Can you name the precise failure modes?

[Jamie Tso @jamietsohn]

@fantony_francis We teach it at the @LegalQuants Residency. https://t.co/OmRfSkShPD

[Jamie Tso @jamietsohn]

@shanethomas21 Which firms do you have in mind?

[Jamie Tso @jamietsohn]

Is revenue per lawyer a better benchmark for AI-native law firms than total revenue? Curious to know when we will see an AI-native law firm hit $2m in revenue per lawyer and how far they are from that figure now.

[Jamie Tso @jamietsohn]

@IWillNeverDieUS Why? Anthropic quotes and brags about their revenue per employee in their S1

[Jamie Tso @jamietsohn]

@Bot1796954 In that case, maybe revenue per employee matters more.