Nikki Shaver

@askperplexity.bsky.social what is the recent news on AI that we should all know today?

This was a great conversation on #DeepSeek - looking forward to hearing more Office Hours discussions from Lega!

It’s a really eerie feeling when someone’s AI assistant joins a meeting before they do or anyone else does and then it’s just you and the AI on the call…

Harvard and Google to release 1 million public-domain books as AI training dataset | TechCrunch AI training data has a big price tag, one best-suited for deep-pocketed tech firms. This is why Harvard University plans to release a dataset that

Harvard and Google to release 1 million public-domain books as AI training dataset
By Paul Sawers
techcrunch.com/2024/12/12/h...
#books #publicdomain

OpenAI’s 12 days of ‘ship-mas:’ all the new announcements Follow along with OpenAI’s launches.

OpenAI announced “12 Days of Shipmas” - starting today, they will release product launches and demos every day for 12 days.

Today’s release:
🎄a new tier of ChatGPT subscription, ChatGPT Pro, and…
🎄the full version of o1!

#AI

www.theverge.com/24314146/ope...

I’m a sucker for a good holiday movie but #ChatGPT could write better plot lines and dialogue than a lot of the Hallmark ones flooding Netflix at the moment…

AI can bring benefits but we must find the appropriate use cases. We can’t do that without having some understanding of the tech. Firm wide AI literacy, mandated by the EU AI Act is a start. Art 4 is in force 2 February 2025

#irishresearch #legaltech #aigovernance

See you there, Carla!

Did you know that over 200 state and federal judges in the U.S. have issued orders or amended orders to address #AI use and misuse in their courtrooms?

Judicial activism is alive & well. Talk about a complex landscape for litigators: no uniformity or consistency, just a Wild West of court orders.

In one of my talks on change, I used a slide with a picture of a rotary dial phone to depict “the olden days.”

Except these are photos from the last two hotels I’ve stayed at, one in Detroit and one in Amsterdam.

#changemanagement

“Remove the squid”

How Claude responds when you push it to the limit.

Why I’m Thankful for GenAI (and Happy Birthday, ChatGPT!) To paraphrase the immortal words of classic rock band Styx, domo arigato, Mr. Roboto, for helping make these past two years so fun to write about. This week, most Americans will be celebrating Thank...

My weekly column for Legaltech Hub. Come for the GenAI and legal tech updates, stay for the robot kidnapping stories. www.legaltechnologyhub.com/contents/why...

Yet another case where an AI-generated brief was filed by a lawyer without verification. Sure enough, hallucinated citation. 🤦🏻‍♀️

Court orders fine but also that lawyer complete course on #GenAI.

Big day for Anthropic tomorrow. Court will hear argument on Concord Music’s motion for preliminary injunction. I’m skeptical but we’ll see

I love that when I say "the good place" and "the bad place" everyone knows exactly what I'm talking about.

Bluesky feels like the IRC channel for #legaltech geeks, but now we’ve all got better wifi. #geeksunite

Xmas gift idea: banned books.

Another case in which false AI generated content was submitted to the court. In this case, a Stanford professor (Hancock) serving as an expert submitted a sworn declaration citing a study that does not exist, likely a hallucination generated by an LLM.

#AI #GenAI #legalAI #legal

This is a wonderful start to the shake-up of education financing that the US needs.

😂 Oh we got you, girl!

Never too early

Another copyright case against an AI giant. #AI #GenAI

Interesting measure of “AI vibrancy”: the Global AI power Rankings, measuring AI use and activity in 36 countries.

No surprise to see the US and the UK high up there. China is in second place, but according to the report, increasingly falling behind the US.

#AI #GenAI

😂

I know! That’s so true. I’m always unfailingly polite to it so this feels very weird.

Welcome, Greg!

Learned something interesting about prompting yesterday: if you put something in ALL CAPS, the model will pay greater attention to it. (Probably learned from US contracts…)

Next time you’re prompting, if you’re having trouble getting one aspect across, try capitalizing it.

#AI #GenAI

YouTube video by Yannic Kilcher GSM-Symbolic: Understanding the Limitations of Mathematical Reasoning in Large Language Models

Apple released a paper in early October that has been widely shared in “pop AI circles” (ie non-data scientists interested in AI) as concluding that LLMs cannot / do not “reason.”

It’s worth watching this video where Yannic Kilcher disputes that.

#AI #GenAI

Welcome, Ron!

You too, it’s fun to find people on here after the mass migration…