Ryanfilevine (@RyanM_Anderson) header
Ryanfilevine (@RyanM_Anderson) profile picture

Husband. Father of 6. Co-founder, CEO, @filevine. #verticalAI #biasedtowardaction

Salt Lake City, UT
Joined May 2012
Photo by Ryanfilevine (@RyanM_Anderson)Photo by Ryanfilevine (@RyanM_Anderson)Photo by Ryanfilevine (@RyanM_Anderson)Photo by Ryanfilevine (@RyanM_Anderson)Photo by Ryanfilevine (@RyanM_Anderson)Photo by Ryanfilevine (@RyanM_Anderson)Photo by Ryanfilevine (@RyanM_Anderson)Photo by Ryanfilevine (@RyanM_Anderson)Photo by Ryanfilevine (@RyanM_Anderson)Photo by Ryanfilevine (@RyanM_Anderson)
Product log: 10/30/26 Lex Summit was phenomenal. Biggest releases ever. Amazing community. Could not be prouder. But Lex doesn't mean we stopped cooking! The new MedChron is editable. Correct any errors or update the file with new documents, and MedChron rebuilds the whole case narrative around your changes. Also new in LOIS for Word: build exhibit packets and pick up your Console conversations right where you left off. And you can now move money between project funds sections in one transfer. No more debiting one account and crediting the other. More updates here: filevine.com/updates/?utm_so…
2
3
17
552
Fascinating. Anyone who has asked llms about themselves has noticed some bias. I don’t even blame them. It’s ok if they have a point of view. But not when deciding cases about themselves.
One interesting conversation we had in class was about how we felt about judges using AI when the case itself involves frontier models, such as the OpenAI copyright cases. I think there is a nonzero chance that the models have some bias toward defending themselves, especially when their parent company might be penalized in the lawsuit. As a result, I don't think judges in those cases should use AI—no matter the model—because some AI models might have bias against their competitors (e.g., using Claude when the case involves OpenAI; Claude might be biased against OpenAI since it's a competitor) So you would almost "recuse" the AI models from the case because they aren't impartial. Crazy to think about, but I think it makes sense.
7
442
When everything feels like the movies Yeah, you bleed just to know you’re alive
2
1
17
561
Ryanfilevine retweeted
“Pass up on the plastic greens for the golden grounds” - Matthew McConaughey on using LOIS AI as he sips coffee from his LOIS mug
2
7
558
“Don’t half ass it.” McConaughey at Lex
1
1
13
550
Ryanfilevine retweeted
LOIS for Outlook reads the email, its attachments and the Filevine project, then drafts the reply. You edit and send. New request? Create the Filevine project from the email and pick where each file lands. Now live: filevine.com/lois/?utm_mediu…
1
4
237
Frontier models will replace every industry that’s not the one I happen work in
hot take but if you can’t find mistakes when using frontier models in your field you’re either only asking easy questions or don’t have enough depth in what you do
7
408
Some companies ship products. Some companies ship companies.
Filevine Finance is the AI-native accounting platform built for law. Three-way trust reconciliation runs continuously. Outside counsel guidelines are enforced at time entry, before an invoice exists. Spend less time closing your books and more time running the business: filevine.com/finance/?utm_me…
2
5
1
31
3,542
You can just build hard things. Huge congratulations to @CainElliott and the F2 team on building the first ever ai native legal accounting platform.
2
3
1
24
785
When you’re heads down building, you often forget how much excellence has been already done This morning’s keynote sessions at LEX were…breathtaking! and with 1k+ attendees to celebrate Proud to be cheering in the crowd for @RyanM_Anderson @WasatchWatch @sonasulakian @JonasKratochvil & team Incredible story, incredible delivery 👏
3
2
11
489
In retrospect, this was a generational move by Big Pumpkin
2
1
472
Few know, they have a hand in everything
1
99
Opus 5.5 is extremely good. But I have a sense this isn’t reflective of broader ai adoption in the economy but rather fickle developer spend moving from OAI to Claude. We definitely saw that internally with our team.
New from OpenRouter: 7-day average token volume for Anthropic models is up 3.6x since the launch of Opus 5.5 Sonnet 5.5 and Haiku 5.5 have also been well received.
5
10
1,383
Directionally correct. Though a chat interface isn’t software’s final form either.
This quoted post is unavailable.
2
4
308
LOIS just saturated StakeBulletinBench
10
414
LOIS’ citation granularity is second to none. LOIS already checks citations for hallucinations as you draft. But we are picky. If the citation isn’t exactly right, we flag it out of caution. Usually though, the cases are real, just not perfectly cited. Now, you can click into an unverified citation (highlighted in light blue), and a case law lookup opens in the right panel. It shows full citations for opinions in our extensive library that might match. You can confirm you have the right opinion without leaving your draft. We also made Skills even easier to build. Now LOIS can create, edit, activate and deactivate Skills on your behalf, right in chat. Like the output you saw from LOIS? Just say, “make this a new skill.” That’s it. No copy and paste. Every user can easily build their own Skills to lock in and reuse their best prompts, and admins can build Skills for the whole firm to standardize processes and enforce AI guardrails. Try writing a Skill with LOIS this week and tell us how it goes. Everything else we shipped this week is here: filevine.com/updates/?utm_so…
2
16
598
I gave my daughter and friends tickets to Phoebe Bridgers They did this as a thank you
4
12
734
Imagine if a company did both (deterministic and AI…and blended) SOTA level, right now at scale. That’d be crazy
this is my issue with AI x law. the last thing I want as a lawyer is anything non-deterministic in my documents lol. it doesn't obey the skills even if you wrote the skills to try to make it behave give me a rigid docgen template and an AI that can't do anything but input variables into the docgen, over generalized legal AI, any day of the week . . .
2
4
411
😅
The highest achievers I know are incredibly effective professionally, usually at the same cost psychologically. The same instinct that drives you to keep raising the bar can make it nearly impossible to feel satisfied with what you’ve accomplished.
2
6
352