Sorry for the delayed response to this, I just woke up:
The team that made dataroom has stated that they did not use any of papermark’s code and that dataroom was made from scratch with inspiration from existing document sharing softwares, and that this post’s allegations of us stealing code are false. We will do an audit of our code to see if anyone else’s code was used at all, we don’t stand for using open source code without attribution, open source is great and we’re supporters at Corgi. I also directly messaged Marc and even though he’s competitive we’re not exactly launching this mostly free product to make a lot of money, but based upon our team’s representations and the information we have on hand his allegations here are false; we will investigate further though and publish the results of our investigation on our website for everyone to see.
Related: we’re hiring a human writer @crosbylegal
LLMs can produce a lot of average writing.
We sell to a lot of not-average people. Founders of fast-scaling AI startups. General counsels at Fortune 500 companies. Chief Revenue Officers at trusted financial services firms.
We’re hiring a 99th percentile writer who can reach our 99th percentile customers.
There's speed. And then there's the speed of a @crosbylegal client.
Take @p0 Parallel Web Systems. The web was not built for agents, so Parallel invented an entirely new way to navigate the internet.
Not only are they ambitious in their mission, but they're relentless about their speed. They've scaled up in under a year, working with some of the world's most important companies.
I'm so excited to finally share the story of our work together 🚀
We built Claude for outbound sellers.
AEs & SDRs can harness GTM engineering through chat across 40+ data sources, no technical skills required.
We’ve had 57,548 queries in our first few weeks of beta, growing 45% w/w.
The teams winning outbound in 2026 are already on it. Self-serve now live here: https://t.co/pwdSGTG0Hq
To celebrate the launch of https://t.co/zLuwpQBoCs, we're hosting an event with NYC's top AI minds on July 1. @jsarihan will chat with @paraga. We're saving a few spots for "that cracked person we met online." If that might be you or someone you know, plz DM me!
Engram’s approach is so fresh it almost seems obvious (if it wasn’t so hard!). Models are forgetful and get overwhelmed by context. So just keep updating the models, ongoing improvement. Congrats to one of the most exceptional teams in the AI world.
Building RedlineBench with @micro1_ai was fascinating.
We asked a bunch of senior lawyers to negotiate against each other. At first they did similar things, but then things got weird. As the deals dragged on, each lawyer went rogue, relying on instinct to get things closed.
How can you possibly grade a model against that!? This is where @aliansarinik's team came in -- design an eval framework, review structures, pipeline design.
As always, loved discussing this with @TBPN@jordihays & @johncoogan
"It's not the same as a math problem where you can just verify it."
@johncoogan@jordihays & @ryanjdaniels break down benchmarks for non-verifiable domains and RedlineBench on @TBPN
LLMs are getting better at law, but on Micro1 and Crosby’s latest benchmark, even the top models can only accomplish 10-20% of what a lawyer can do. Can we train models to make sounder legal judgments, or will we always need a human (lawyer) in the loop?
CEOs @aliansarinik and @ryanjdaniels join a new TWiAI to talk about their new legal benchmark test, why AI still struggles in the courtroom, SpaceX buying Cursor, and more.
0:00 SpaceX acquires Cursor
19:43 Distillation vs. building your own model
30:21 Nadella's "Frontier Without an Ecosystem"
32:15 AI in the courtroom
1:05:38 Ando: the intriguing new workplace tool
1:07:14 Inside Micro1 and Crosby's new benchmark
🎥 Watch the full episode here 👇