Why do workers have to wait for 2-4 weeks to be paid, in the same economy where online transactions go quickly and securely?
A new draft πππ¦πππ¦-forthcoming @WashULaw-proposes that they shouldn't.
Daily, or at least weekly, pay can be a reality.
https://t.co/OXjd7ESYhJ
@davidmanheim I think this is urgently needed. Still, there's a different class of governance solutions. You may want to check out our a-corps, https://t.co/LITX18z99k, and I'm happy to recommend other lit
Contracts have a backward-compatibility problem, just like code, and there's a lot of legacy clauses still left in.
And it's not like drafters don't care about what they write. They do adapt to email and even DocuSign. So there's a real puzzle here
Read more:https://t.co/4yT6OZyYuO
Can you sign a multimillion-dollar deal by telex?
@HoffProf goes fossil-hunting on his Substack, exploring a puzzle in dealmaking.
He's testing the Gulati/Scott/Choi theory that obsolete terms get fossilized in contracts.
He brings really vivid evidence: M&A agreements in 2025 still list telex and telegram
@mattBernius The authors collected expert expectations about their study in advance, which is useful to measure surprise and avoid goalpost shifting after publication
There are distinct ideas about using prediction markets for studies, but you'd have to worry about affecting the study
My pre-registered prediction of this outcome was: 85%. AI is an amazing tool for pedagogy, and many educators not just "sleep on it" but actively resist it, Berkeley Law style
@ProfDBernstein ChatGPT is very good and Google's is also quite good ("nano bannana 2"). The trick is to guide its style carefully, otherwise it becomes unbearable AI slop. Happy to send along a small gallery of options for chapter 2
@ProfArbel@scaling01 The revealed vs. stated preferences distinction is a fascinating angle. I'm quite interested in seeing how this develops as models become more capable.
This news from Kirkland along with the Harvey legal agent benchmark this week raises a big question: will the next major improvements in legally useful AI come from law-specific model building, or from improvements in general-purpose foundation models? π§΅
FWIW, law firms are not start up hubs, much less AI labs, and -- for certain values of 'ai technology' -- this will end badly.
There are ways to make these things work, but they all involve getting as far away from the foundational model layer as possible
@r_tallarita Yes, that's why they need to stick to the epidermis of AI, the outermost layers, and get their hands as far off as possible from architecture and training.
On age verification techniques for children safety with AI:
βI would urge you to think about when each type of age verification method is appropriate and what tolerance do we have for mistakes,β Arbel said.