5̷̤̚͝H̴̙͐0̷͕̀G̶̳͗G̴͎̎0̶̱͒T̵̼͑H̵̰̽ ̶̬̒5̵̞̔U̷͔͘R̸͍̈́F̶͉̓3̵̯̈́Ṟ̶̍ | AI x Law
Phoenix, AZ
Joined August 2019
- Tweets712
- Following2.4K
- Followers572
- Likes17.5K
Pinned Tweet
We're launching the Applied Legal AI Credential at ASU Law on December 18!
Hands-on AI training for attorneys: build tools, work with agents, evaluate vendors, and earn a digital credential.
Up to 5 AZ CLE hours, including ethics.
events.asucollegeoflaw.com/a…
Sean A. Harrington retweeted
my sister is a freshman in college
me: what do the other kids think about ai
my sister: oh everyone hates it
me: but are they using it…?
my sister: well duh everyone is. everyone is on chat/gemini/claude for their classes
stated preference ≠ revealed preference
Sean A. Harrington retweeted
TRACKED CHANGES: Ropes & Gray now lets all of its associates count up to 100 hours of experimenting with AI toward their billable targets, Reuters reports.
Sean A. Harrington retweeted
Gamifying legal work for lawyers with crackhead work ethic is diabolical. I already work 60-100hrs a week in some cases but turning it into a game just to flex on in house plebs?
Fuck it. Sign me up let's see where this goes.
introducing Promethee, an app that turns your work into a multiplayer game, to perform.
now available for everyone.
promethee.io
Sean A. Harrington retweeted
Confirmed: @Muse can sue people in small claims court
Sean A. Harrington retweeted
Talked to a Stanford prof last night who told me that AI has led to students all acing all the homeworks, never showing up to office hours, and then failing the final exams in-person.
So in response the department is changing the TA's purpose, by eliminating office hours and requiring students meet with TAs 1:1 and walking them through each coding assignment to demonstrate they actually understand what's going on in the code.
Sean A. Harrington retweeted
Ding ding! this was never really a great vehicle for the broader questions of Copyright and AI Training — the Opinion did a good job of cutting through the smokescreen … THOMSON REUTERS vs ROSS INTELLIGENCE - 3rd Circuit opinion is now available ... case should also not be given Cert by #SCOTUS - again a bad vehicle for considering the broader questions ...
chatgptiseatingtheworld.com/…
Sean A. Harrington retweeted
Last week, I wrote about what data center investment could mean for Arizona.
Another $600+ million a year could mean:
- $10,000 for every public school teacher.
- Roughly half the annual cost of the ESA program.
- No state income tax for people earning under $50,000.
- More funding to secure Arizona's water future.
I'd much rather debate what we should spend the money on than whether we should welcome the investment at all.
What could a larger data center industry help Arizona afford?
@PwCUS estimates the industry generated $607 million in state taxes in 2024.
@PewResearch counted 98 operating data centers here and nearly as many planned or under construction. What if a larger industry eventually doubled its tax contribution?
To put another $607 million in perspective:
- That's $10,000 for each of Arizona's roughly 60,000 public school teachers.
- We could cover roughly half the annual cost of the ESA program.
- We could eliminate state income tax for people earning less than $50,000 a year.
- We could fund roughly 173,000 acre-feet of additional water, assuming we could secure and deliver supplies at $3,500 per acre-foot. Nearly a quarter of Arizona's Colorado River reductions.
And this doesn't include the local revenue flowing to cities and counties.
Different priorities. Different possibilities.
None of this is a forecast or a spending plan. But it gives you a sense of scale.
We spend a lot of time asking what new investment will require of Arizona. What if we started asking what it could help Arizona afford?
Here's how we can ensure the benefits reach Arizonans. 🧵
Sean A. Harrington retweeted
every major company is arriving at roughly the same offering:
persistent memory, email/calendar/messages, browser + computer use, background tasks, proactive notifications, voice, app/tool execution, ambient context, & some notion of a personal agent sitting above everything.
remarkable levels of convergence with very little differentiation whatsoever.
Sean A. Harrington retweeted
"AI slop" is threatening courts' ability to function, appeals court says. ow.ly/NRZU50ZPKuc @amandasrobert
Sean A. Harrington retweeted
The world is not ready for the amount of people soon to be walking around with these types of devices
1/ Better robot models come from better data.
That takes two things: hardware that captures it fully, and tooling that turns it into training signal.
We built the first. Today, @GroundedSI launches the second: Grounded API. Access both hardware + API ↓
Sean A. Harrington retweeted
Also today: Claude Docs, Claude Slides, and Claude Design are in every conversation.
Ask Claude for a presentation and you get one you can open, edit, and export as PowerPoint or PDF. Same for a document or a design. There's no separate tool to navigate to. They're just in the chat.
Sean A. Harrington retweeted
I agree with Orin. And from my vantage, there seems to be a growing divide between two camps in the academy.
The first treats knowledge claims as a product and, so long as those claims are true and useful, the involvement of AI in their generation is largely immaterial. The second sees the value of knowledge claims as inseparable from provenance, with the worth of a claim inextricably bound up with its genesis.
My (admittedly armchair) sense is that the divide centers roughly on determinacy, both within and beyond the legal academy.
In determinate fields (or jurisprudential modalities), where the academy seeks objective, falsifiable findings, the first camp's norms dominate. For example, it makes little difference whether Copernicus or an AI proposes the heliocentric model of our solar system. The key is that it is correct, which carries meaningful implications for future knowledge discovery.
In indeterminate fields, where the academic pursuit is instead driven by judgment, reasoning, and persuasion, the second camp's norms predominate. In these spaces, the "right answer" to an issue isn't discerned by acquiring some heretofore missing knowledge, but rather by "making the case" for the normatively optimal path forward given the (fully explicable) tradeoffs at issue. Within this framework, having AI craft an argument feels much more uncomfortable, as it involves outsourcing considerations of morality and equity to a synthetic machine.
With apologies to Langdell, my sense is that most of the legal academy's work falls in the second camp, and I therefore expect that its norms will develop in an AI-skeptical direction as a result. At least for anything beyond ministerial tasks like cite-checking, proofreading, and editing.
Sean A. Harrington retweeted
📢 New policy brief: Dozens of U.S. jurisdictions still have unconstitutional and discriminatory laws on the books. But sifting through thousands of local codes by hand is infeasible. Our latest brief demonstrates how AI can make legal codes legible again, starting with discriminatory laws hiding in plain sight. hai.stanford.edu/policy/maki…
Sean A. Harrington retweeted
Astra has successfully beat all 48 levels of the “I’m Not a Robot” game:
Sean A. Harrington retweeted
Go to Costco, get yourself a 4 lb bag of cosmic crisp apples. Then get yourself a 2.7 year old who will take a bite out of each one when you’re not looking and throw them all behind the couch.
Sean A. Harrington retweeted
the 2nd prong of Ken Paxton's data center regulatory platform would enshrine into law just about the most open-ended and destructive theory of vicarious liability in history, stating that DATA CENTERS will be held CRIMINALLY liable for downstream uses of AI chatbots that "undermine children's safety" ( however that ends up being defined.)
This is like saying an electricity company will be held criminally liable for problems that develop with any appliance or product plugged into the grid. Or, if applied to gun manufacturers, it would mean a firearm or ammunition maker would be criminally liable for any act of violence committed by a person using their gun or ammo.
Let's be clear, when people propose outlandish theories of liability such as these, their real goal is to just destroy the underlying technology or industry in question. It is absolutely impossible for these sorts of theories of liability to work in the real world because, in our nation of endless frivolous lawsuits, the courts would just be filled with mountains of junk science-based claims as trial lawyers go in search of ways to bleed innovators dry to enrich themselves. All we would be left with was one or two extremely large big tech companies who would just constantly pay out to the trial lawyers to settle frivolous claims. Small innovators would never be able to bear this burden. Costs would go up for average Americans and they would be denied many important services and new options because of the new tort toll. Again, it's all just a trial lawyer enrichment racket.
Ken Paxton knows all this but he's on seek and destroy mission against the American digital technology sector and so he doesn't care about the consequences of such insanely destructive theories of liability.
Sean A. Harrington retweeted
I took the nearly 10k of synthetic data Harvey produced for its synthetic law firm, made it a bit more realistic, and then ran the same test Harvey had conducted on LLMs, directly against Harvey vaults.
thegeekinreview.substack.com…
