Discover more from Aidful News
🤝 AidfulAI Newsletter #17: AI For Your Private Daily Health Check
Dear curious minds,
Welcome to the ultimate newsletter for those interested in Artificial Intelligence (AI) and Personal Knowledge Management (PKM). In this week's newsletter, we'll dive into several intriguing developments:
OpenAI's acquisition of a Minecraft-like game creator
Azure ChatGPT, a private enterprise solution from Microsoft
Notion's new feature for saving and reusing AI prompts
DoctorGPT, an AI system for medical analysis and daily health checks
A podcast exploring human creativity in the age of AI
A podcast about the alignment effort from OpenAI
Stay tuned as we explore these trends, shedding light on the intersections of technology, creativity, and privacy in our rapidly advancing world.
Major AI News
🎮💰 OpenAI Buys a Minecraft Clone Maker
Global Illumination is the developer of Biomes, an open-source, web-based sandbox game similar to Minecraft, where players can explore, build, and collaborate in a 3D world.
The acquisition could signal OpenAI's interest in using games as a platform for developing and testing AI agents, as well as creating immersive and interactive experiences for users.
OpenAI did not disclose the terms of the deal or the future plans for Global Illumination but said that the team will work on its core products, including ChatGPT.
My take: Is mastering a Minecraft-like sandbox game with an AI the key to creating an AGI?
🔐💼 Azure ChatGPT: A Private and Secure ChatGPT for Enterprises
Microsoft has launched Azure ChatGPT, a service that allows enterprises to use ChatGPT without connecting to OpenAI. Azure ChatGPT can be deployed in your private Azure cloud.
Microsoft claims that Azure ChatGPT can protect intellectual property in contrast to using the public ChatGPT service.
My take: Not clear what this all means, but Microsoft and OpenAI have a complex relationship, as they collaborate on some projects but compete on others. Companies that are already subscribed to Microsoft services and especially Azure will like the new option. The recent news of the stolen Microsoft key which gave hackers access to many systems from Azure customers will likely not have a significant impact on this.
PKM and AI
🚀📝 Notion Enables Saving Favorite Prompts for Easy Reuse
Notion now allows you to save your favorite prompts for easy reusing them.
In the announcement video, the team shared their favorite AI prompts:
Identify action items
Make this more concise
My take: No revolution but nice to see more options in tools to tailor the AI to your own usage.
🩺💻 ChatGPT Passes Medical Exam, Aims for Daily Health Checks
DoctorGPT is a Llama2 7B model, fine-tuned with medical data. With that, it works offline and can be used on desktop and mobile platforms.
The AI system passes the US Medical Licensing Exam.
More info about DoctorGPT is shared in the Youtube video from its creator Siraj Raval.
Brian Roemmele shared in a 𝕏 post that he is working on using data from multiple sensors which are used as input for a daily health check created by DoctorGPT.
My take: Another use case that highlights the value of a local, and with that privacy-friendly, AI. Nevertheless, take any advice DoctorGPT gives you with caution, because as with any other LLM also this model might hallucinate.
🎙️✍️ Human Creativity in the Age of AI
Arvid Kahl talks in episode 241 of The Bootstrapped Founder podcast the future of human content creation.
The worth of spending time on creating content that might be absorbed by automated systems is questioned, highlighting the implications of AI's impact on human writers.
Differentiating between AI and human writing is becoming crucial, as people tend to trust human authors more, emphasizing the significance of author identity.
In social media communities, imperfect but human-written content is often preferred over AI-generated content, showcasing the value of authenticity and the human touch in writing.
My take: As Arvid states in this podcast episode, human content creation is not going to be replaced by AI anytime soon, but rather complemented and enhanced by it.
👩💻🤖 How OpenAI Plans to Create a Safe Superintelligence
In episode 159 of the 80,000 Hours Podcast, Rob Wiblin interviews Jan Leike, the head of alignment at OpenAI, about his ambitious project to make superintelligent AI systems aligned and safe to use within four years.
Superalignment is OpenAI's new initiative to tackle this problem, by using AI itself to help design, understand, and monitor AI systems. The idea is to create AI assistants that can help human researchers solve alignment challenges and keep up with the increasing complexity of AI models. The project is led by Jan Leike, who is optimistic that it can be done within four years.
The plan is to get AI to help us solve alignment and OpenAI is spending 20% of its computational resources for this endeavor.
Some critics are skeptical that this can be done, but Jan believes it is the most important research in AI.
My Take: It's good to see serious efforts like OpenAI's Superalignment team working to tackle the challenge of AI safety. With rapidly advancing capabilities, ensuring AI systems remain aligned with human values only grows more crucial.
Thanks for reading! If you are not already subscribed, enter your mail address below to receive new issues and support my work.
Disclaimer: This newsletter is written with the help of AI. I use AI as an assistant to generate and optimize the text. However, the amount of AI used varies depending on the topic and the content. I always curate and edit the text myself to ensure quality and accuracy. The opinions and views expressed in this newsletter are my own and do not necessarily reflect those of the sources or the AI models.