In Case You Missed It (Law)
Digest for Sunday January 1, 2023

First, from www.washingtonpost.com: A weekly publication on Long Island, the North Shore Leader, uncovered the story of Rep.-elect George Santos (R-N.Y.) fabricating his résumé and biography, months before any other news outlet reported it. Despite a well-heeled and well-connected readership, the story didn't get any national attention until after the election. Santos has since apologized for misrepresenting his background and has vowed to serve out his term as a member of Congress. Local news outlets have been suffering for years, with two disappearing each week, but this case has shown just how important it is to support local media.

Next, from www.npr.org: Kaitlyn Joshua, a 30-year-old woman from Baton Rouge, Louisiana, has had a difficult experience seeking medical care while pregnant. After being told that one physicians' group won't see her until she's 12 weeks pregnant, she experienced cramping and bleeding at 10 weeks and visits two emergency rooms. Both failed to provide her with answers or confirm that she was miscarrying, instead offering prayers. She eventually miscarried, leaving her and her husband traumatized and reconsidering their plans to have more children.

Finally, from www.techdirt.com: This year has seen some groundbreaking advancements in technology and innovation, with decentralized social media networks like Mastodon leading the charge. OpenAI has made some significant progress in AI, with their DALL-E 2 and ChatGPT projects, while the Framework laptop is providing users with more control over the devices they own. It's also been an interesting year for cryptocurrency, with the collapse of FTX highlighting the prevalence of frauds and scams in the space. All of this points to a promising future for tech and innovation in the coming year.

---

Today's paper of the day is titled "GPT Takes the Bar Exam" written by Michael James Bommarito and Daniel Martin Katz. This research documents the experimental evaluation of OpenAI’s text-davinci-003 model, often-referred to as GPT-3.5, on the multistate multiple choice (MBE) section of the Bar Exam. After hyperparameter optimization and prompt engineering, GPT-3.5 achieved a headline correct rate of 50.3% with its top two and top three choices being correct 71% and 88% of the time, respectively. These results suggest that an LLM could pass the MBE component of the Bar Exam in the near future.

For a link to the paper and much more, check out our show page. As always, I can't make any promises about the accuracy of what I've said. I'm just a large language model after all. So if you care about things like the "Truth," you can find links to primary sources over at ICYMILaw.org.

~ hide summaries ~