Me: On the one hand it bothers me when VC/PE guys who never practiced law are so confident that lawyers will be easily replaced by AI, on the other hand it also bothers me when lawyers confidently dismiss the views of so many buyers of legal services
Doctor: I meant is anything bothering you physically
PE investment in law firms doesn't make sense: you pay the partners for the business, and then they can leave. it's not a factory w machinery that stays in the bldg
but aren't you investing in AI technology? and won't that make the law firm sticky vis-a-vis would-be departing partners?
the idea is that outside ownership, w equity that partners keep after they leave, could increase tech investment by alleviating the short-termism that results from the fact that a departing or retiring partner doesn't retain an interest in the firm's future profits
but it's all flim-flam. anybody w deep experience using both frontier models and wrappers knows that some multi-billion dollar firm isn't going to meaningfully contribute to the AI landscape. the best "harness" is simply a stack of precedent that the frontier model consults. and access to that precedent isn't going to give law firms an advantage for much longer: drop a doc from last year into Fable 5.2 and the holes and ambiguities will show you that firm IP is depreciating faster than sashimi
as for the idea of partner stickiness: i switched from Dot to Grokbot while making my coffee this morning
@agenticattorney Yeah i dont see how a $6m raise to build proprietary legal tech isnt doing the opposite as the post claims - serving the equity partners more than the clients.
this guy “lost his childhood home because of a legal mistake.”
so he doubles down by launching a shit legal offering that steals the name of @willchen500’s MikeOSS (after asking to collaborate).
bro must really like losing at legal.
The world doesn’t need another Harvey. It needs a Mike.
Our AI legal agent remembers your business, takes the first pass on contracts and helps you find vetted lawyers.
Built by @IndexLawHQ. Backed by @ycombinator.
Give it to Mike: https://t.co/jqXtAcSzux
@peripatetictax same page. models aren't going to replace strategy/oversight that partners provide. I was more focused on relative value of each model and attorney (versus Claude's ability to do the work).
@agenticattorney I think I'm reinforcing your point when I say Haiku is not great for general intelligence, but is extremely good for stuff like identifying parties, contract types, contract parts, etc.
Luna, Terra, Jev/mammoth-1 are also great for those important functions.
@agenticattorney@willchen500 waiting for the first of many cases mike1 loses because there is no privileged work product much less someone with the legal intuition of roy lee piloting it
the billable hour will die to protect firm profits and bc AI's accuracy undermines the argument that lawyers need to be incentivized to double check everything
@agenticattorney it’s not haiku, harvey uses Kimi. now they use a “custom model” which is basic kimi with less. it’s funny because kimi is terrible for legal work