AI Weekly Digest #26: ChatGPT Announcement Leaks

ChatGPT knowledge update: Cutoff date - April 2023

AI Weekly Digest #26: ChatGPT Announcement Leaks

Hello, tech enthusiasts! This is Wassim Jouini and Welcome to my AI newsletter, where I bring you the latest advancements in Artificial Intelligence without the unnecessary hype.

You can find me on LinkedIn, Twitter and Medium! Let’s connect!

Now let's dive into this week's news and explore the practical applications of AI across various sectors.

Main Headlines

Here are the three main trends to keep in mind if you are working in AI today.

#1 ChatGPT Knowledge Base Update

OpenAI has been updating its models to enhance efficiency, speed, and reliability while maintaining a vast repository of knowledge. The initial version of ChatGPT, released at the end of November 2022, contained information up to September 2021.

The model has since been retrained and now encompasses data up to April 2023, allowing it to provide insights on a variety of topics, including the COVID-19 pandemic, advancements in AI, the war in Ukraine, among others.

The tech community is buzzing with anticipation as OpenAI is set to host its inaugural developer conference on November 6th.

Leaks have revealed a slew of exciting updates to the ChatGPT platform that are expected to be unveiled during the event. Here are the major announcements to look forward to:

  • Introduction of custom chatbots through a new tool called Gizmo, which will aid in creating, managing, and selecting custom chatbots.

  • A new enterprise subscription and workspace management feature named the Team Plan, aimed at fostering collaboration and management in a workspace setting.

  • The debut of "context connectors" that would link ChatGPT with Google Drive and Microsoft 365, enabling documents, spreadsheets, or presentations to be attached to chats or used as context for conversations.

  • Speculated price reduction for the GPT-4 API, alongside speed enhancements for GPT-4 and general access to an updated version of GPT-4.

Looking forward to learning more about OpenA’Is announcements tomorrow!

The White House has issued a new executive order aimed at managing the risks associated with Artificial Intelligence (AI), highlighting several key areas of focus:

  • AI Security: To mitigate AI risks, an advanced cybersecurity program will be established to develop AI tools that can identify and rectify vulnerabilities in critical software. This initiative builds on the Biden-Harris Administration's ongoing AI Cyber Challenge, aiming to utilize AI's capabilities to enhance cybersecurity across software and network infrastructures.

  • Privacy: The order calls for the development of guidelines for federal agencies to assess the effectiveness of privacy-preserving techniques, including those utilized in AI systems. This step is aimed at advancing efforts to protect data and privacy.

  • Advancing Equity and Civil Rights: Efforts will be made to address algorithmic discrimination by facilitating training, technical assistance, and coordination between the Department of Justice and Federal civil rights offices. The order seeks to ensure fairness in the criminal justice system through the development of best practices for AI applications in various judicial processes.

  • Education: The order highlights the potential of AI to transform education and calls for the creation of resources to support educators in deploying AI-enabled educational tools, such as personalized tutoring.

  • Employment: A framework will be developed to mitigate potential harms and maximize benefits of AI in the employment sector. This includes addressing job displacement, labor standards, workplace equity, health, and safety, and data collection to provide guidance aimed at preventing unfair practices and ensuring a fair work environment.

xAI has entered the competitive arena of Large Language Models with its inaugural model. According to standard LLM benchmarks (see image below), this model outperforms GPT-3.5, although it has not yet reached the performance level of GPT-4.

Remarkably, xAI's model boasts 33 billion parameters, which, while significant, is still considered small in comparison to the original GPT-3 model's 175 billion parameters

The model isn’t publicly available yet ; You can join the Grok waitlist here.

This is it for Today!

Until next time, this is Wassim Jouini, signing off. See you in the next edition!

Have a great Sunday and may AI always be on your side!