Itâs Tuesday. Lotâs to cover.
Letâs dive in!
In todayâs Brainyacts:
Copyright loopholes
Googleâs NotbookLM available in US and gets huge update
GPT-4 is getting lazy and more AI model news
Law school launches GenAI course and other related content
đ to new subscribers!
To read previous editions, click here.
Lead Memo
Šī¸ đ¤ˇââī¸ Is this image protected by copyright?
It is based on these two images:
Well if you are keen to know, read this great summary and explanation. đ
Spotlight
đ§đŦ Google NotebookLM adds a dozen new features
Many of us have a place where we feel like we do our best thinking â maybe outside on a walk, or at the kitchen table sipping a cup of coffee. Itâs where we organize our thoughts to articulate messages, accomplish goals and make connections. And wherever we happen to do it, that thinking process almost always involves taking ideas from different tools and sources â whether quotes from books, instructions from websites or handwritten musings from notebooks â and combining them into new forms.
Large language models (LLMs) have given us an extraordinary opportunity to reimagine that process. With the advanced capabilities of LLMs, you can have more than just an organizational tool â you can have a personalized AI collaborator.
NotebookLM, an experimental product in Labs designed to help you do your best thinking, is now available in the U.S. to ages 18 and up. And itâs starting to use Gemini Pro, Googleâs best model for scaling across a wide range of tasks, to help with document understanding and reasoning.
Here are some of the recent updates:
Increased source limit: Notebooks can now include up to 20 sources.
Increased word count:Â Sources can now include up to 200,000 words.
Write individual notes: Adding a note will now create an independent new note instead of adding to a single notepad.
Noteboard: You can now see all your written and saved notes pinned to the Noteboard space above the chat box.
Save chat responses as notes. NotebookLM responses can be pinned to the noteboard for your reference later.
Citations saved to notes. When chat responses are saved as notes, it retains the original citations.
Jump to citation in source. Clicking on the citation number in a chat response or a saved note immediately takes you to the original quote in the source.
To see all the updates, click here.
For a 5 min refresh and walkthrough, check out this video:
AI Model Notables
âē OpenAI Admits That GPT-4 Has Become 'Lazier' For THIS Reason; Fix Coming Soon; more coverage here
âē French startup Mistral AI just released Mixtral, an open-source 45B parameter AI model using advanced âmixture of expertsâ architecture to match rivals with far lower cost and 6x faster
âē Reed Smith turns to AI for lawyer staffing solution
âē Anthropic leads charge against AI bias and discrimination with new research; link to paper here.
âē Anthropicâs latest tactic to stop racist AI: Asking it âreally really really reallyâ nicely
âē Meta releases demo product Audiobox - an AI tool to create AI-generated voice and noises. Think of a sound effects engine
âē ChatGPT passes neurology exam for first time
News You Can Use:
â Intellectual property attorney Jack Hicks encouraged members of the Class of 2023 to use artificial intelligence throughout their legal careers to enrich and enhance human development as AI creates "permanent, widespread, and disruptive change across society at large."
The legislation is set to take effect early next year and will be applicable two years later, with the move positioning the EU as the first major world power to enact comprehensive laws governing AI.
The new rules include transparency obligations for AI systems and foundation models, requiring technical documentation and compliance with EU copyright law.
Furthermore, high-impact AI models will face stricter evaluations and reporting requirements, and the use of real-time biometric surveillance by governments will be restricted to specific, serious cases.
Lastly, the legislation bans certain practices, such as cognitive behavioral manipulation and untargeted scraping of facial images, and consumers will have the right to lodge complaints and receive explanations, with significant fines for violations.
â OpenAI picks WilmerHale lawyers for Altman firing investigation
â Purdue Global Law School launches AI course
â Judges in England and Wales will be able to use ChatGPT to help write legal rulings despite warnings that GenAI can invent cases that never happened
â Microsoft and the AFL-CIO struck a deal, announced Monday, to train labor leaders in aspects of AI and make it easier for workers at Microsoft to unionize
â Liquid Neural Networks might be next-gen AI Model
â Out with prompt engineering, in with prompt tuning
Was this newsletter useful? Help me to improve!
Who is the author, Josh Kubicki?
Some of you know me. Others do not. Here is a short intro. I am a lawyer, entrepreneur, and teacher. I have transformed legal practices and built multi-million dollar businesses. Not a theorist, I am an applied researcher and former Chief Strategy Officer, recognized by Fast Company and Bloomberg Law for my unique work. Through this newsletter, I offer you pragmatic insights into leveraging AI to inform and improve your daily life in legal services.
DISCLAIMER: None of this is legal advice. This newsletter is strictly educational and is not legal advice or a solicitation to buy or sell any assets or to make any legal decisions. Please /be careful and do your own research.8
