Wikipedia sets the law. In an experiment in Ireland, a random set of cases were given detailed Wikipedia articles written by law students. The cases that were added became 25% more cited in actual legal opinions by busy judges than they had been before. https://
pubsonline.informs.org/doi/full/10.12
87/isre.2023.0034
…
@emollick
-
Wikipedia Boosts Legal Citations Through AI-Written Law Articles
By
–
-
Devin AI Coder Agent Creates Startup Dilution Economics Visualization
By
–
AI agents have real potential. I went back to Devin, the AI coder agent, after not using it for a day & completely forgot that I assigned it to create a visualization of the economics of startup dilution. It plugged away and I came back to a solid draft that could be iterated on
-
Scaling Trends: Establishing Testing Standards for Next-Gen AI Models
By
–
Also, sets up a good standard for testing when GPT-5, Gemini 2.0, etc. come out. We need to understand scaling trends to see what progress is actually being made, as the benchmarks out there are not very useful.
-
Testing GPT-4 Papers Against Gemini 1.5 and Claude 3
By
–
I want to see key GPT-4 papers re-tested with Gemini 1.5 and Claude 3 to see what generalizes across GPT-4 class LLMs. At a minimum, the papers on hallucination rates, Theory of Mind & Chain of Thought; as well as papers on performance on medical, legal & psychological questions
-
Regulators Must Define Legal AI Use in Regulated Industries
By
–
Talking to many companies in regulated industries, it is clear there is a burning need for regulators to define the ways in which AI can be legally used. Ethical experimentation will be the key to using AI well, now it is all just employees secretly using it without permission.
-
AI Accuracy Depends on Access to Quality Human Comparisons
By
–
But that doesn’t mean that AI can’t be more accurate than a human. It just depends on what humans you have access to.
-
Centaurs and Cyborgs: Human-AI Collaboration Framework
By
–
Write up of our paper: https://
oneusefulthing.org/p/centaurs-and
-cyborgs-on-the-jagged
… -
AI Quality Threshold Reduces Human Oversight and Error Detection
By
–
Lex is right, Sam was wrong. Our research shows that people don’t really check the work of AIs, once those AIs cross a certain threshold of quality. We found that consultants, for example, “fell asleep at the wheel” and did not look for AI errors. Training didn’t help, either.
-
Simon-Ehrlich Wager: Technology Resources and Economic Future
By
–
Context of the wager https://
en.wikipedia.org/wiki/Simon%
80Ehrlich_wager
… -
Invisible Productivity Gains from Automation Reshaping Employment
By
–
The massive productivity improvements from automation (and resulting changes in employment) that you don’t see, but on which the modern world depends. From Conway’s “The Material World”