Keep pulling the thread on Kevin Frazier & Alan Rosenstein.
Frontier AI models are already intellectually superior to the median lawyer.
Law firms are considering hiring fewer junior and summer associates due to the increasing capabilities of AI.
Anthropic's Claude Constitution is based on virtue ethics, prioritizing high-level principles and contextual judgment over a comprehensive set of detailed rules, reflecting an Aristotelian approach.
A "right to compute" has been enacted into law in Montana, and similar bills are being considered in Ohio and New Hampshire.
Anthropic has indicated it is no longer looking to hire junior employees in most departments, a trend potentially driven by AI productivity gains.
Arizona became the first U.S. state to allow non-lawyers to own and operate law firms.
According to Amanda Askell, Anthropic avoids giving Claude a long list of rigid rules because of concerns that forcing the model to follow a rule when it knows a better outcome is possible could lead to problematic behavioral generalizations.
According to its own statistics, the AI legal tool Harvey is used by 70% of the top 100 US law firms.
The billable hour compensation structure in the legal profession disincentivizes the adoption and efficient use of AI tools.
On the GDP Val benchmark for lawyers, Claude Opus is the top-performing model, winning one in three head-to-head comparisons against human lawyers and winning or tying in 70% of cases.
Some law firms now prefer hiring a mid-tier law school graduate with strong AI skills over a top graduate from a school like Harvard who lacks AI experience.
Unauthorized Practice of Law (UPL) statutes have been used by the legal profession to create significant hurdles for technology companies like LegalZoom.