🤝 AidfulAI Newsletter #9: Adopt the Future of Local AI and Voice-based Note-taking
Dear curious minds,
In this issue, we address an urgent security alert for ChatGPT users. We also discuss Googles recent advice to its employees about sharing data with chatbots and explore the potential of a new ecosystem, GPT4All, that allows for using AI locally without data sharing. Furthermore, we take a closer look at the trend of voice-based note-taking and its impact on personal knowledge management. Let's get started!
Major AI News
🔐🚨 ChatGPT Users: Change Your Passwords Now!
Your ChatGPT account might be in danger. A hacker has stolen over 100,000 ChatGPT account credentials and sold them on the dark web. This means that someone else might be able to log in to your ChatGPT account and see your chats, images, and code. If you added a credit card to your account, someone might generate new API keys and use them with your card covering the costs.
If you use ChatGPT, you should change your password right away, and you should turn on two-factor authentication (2FA). The latter will make your account more secure and harder to hack.
OpenAI states that this was the result of malware on peoples devices and not an OpenAI breach, as published in an update to the initial article on Hacker News.
Privacy-Friendly AI
📑 🔐 Google Employees Are Advised to Limit Data Shared With Chatbots, Even Their Bard
Alphabet, Google's parent company, has advised its employees to be cautious when sharing confidential information with AI chatbots, according to an article from Reuters.
Alphabet also clarified that its AI chatbot, Google Bard, collects conversations, location, usage information, and feedback to be used in developing machine-learning technologies. Human reviewers read, annotate and process selected conversations. Google claims that they disconnect these conversations from their Google accounts and advises its users in the Bard Privacy Notice to not include any information that can be used to identify you or others in your Bard conversations.
You can configure that Bard does not save conversations to your Google account, but it is not clear if this also excludes them from being manually processed or used for future training runs.
🔒💻 GPT4All: How to Use LLMs on Your Own Machine Without Sharing Your Data
If you value your privacy but still want to use LLMs, you might want to try GPT4All, a new ecosystem that lets you use large language models (LLMs) on your computer without relying on a cloud service.
GPT4All is a free and open-source project that aims to make LLMs accessible and customizable for everyone. You can choose from a variety of open-source models, each with its own capabilities and limitations. Some models can be used commercially, while others are for personal or educational use only. You don't need a GPU or an internet connection to use GPT4All. All you require is a desktop computer running Windows, OSX, or Ubuntu. GPT4All provides easy to use installers for each platform. The latest release increased the performance on Apple's M1/M2 Macs and added a model which is fine-tuned for coding by Replit, a collaborative software development platform, which can be used commercially.
With GPT4All, you can enjoy the benefits of LLMs without sharing your data. However, you should be aware that the quality of the replies is often not on the same level as you are used to from ChatGPT. But, you are controlling your data. This is only the starting point, and I am convinced that there will be more, powerful local models available soon.
PKM and AI
🗣️📝 Voices of the Future: Changing the Way We Take Notes
Say goodbye to traditional note-taking because the future is voice-based! Mustafa, joined our latest Twitter space conversation and opened our eyes to the innovative possibilities AI brings to the table to transform the way you take notes.
Thanks to the improved quality of audio transcriptions, e.g. by the open-source approach Whisper from OpenAI, you can simply speak your thoughts and let AI transcribe them flawlessly. You can even go one step further by adding a command which should be executed by the AI, e.g., “summarize my ideas into bullet points” or “create counterarguments to my thought”.
Some PKM tools like Reflect and Tana are already integrating voice-based capturing in their mobile apps. If you are not using one of these tools to take your notes, you can also use the official ChatGPT app from OpenAI or the Bing app from Microsoft, which also supports voice inputs. However, you should be aware that these are all cloud services and your data might be analyzed and used by the companies creating them.
This new way of note-taking can change the way you create and conserve your personal knowledge.
Podcasts
🤖 💡 Generative AI: The Future of ML/AI?
Generative AI is a type of artificial intelligence that can create new data, such as images, text, or music. It is different from traditional AI, which is typically used to classify or predict data. Generative AI is still in its early stages, but it has the potential to revolutionize many industries.
In the podcast episode “From ML to AI to Generative AI” of the Practical AI podcast, the hosts Chris and Daniel discuss the different approaches to generative AI. They address the risks and compliance implications of using it, and how it fits into the wider landscape of ML/AI.
🤖📚 Unleashing AI's Potential: Insights from Ethan Mollick
Ever thought about how talking to AI can spark new ideas? Ethan Mollick, a professor from the Wharton School, chats about this in the newest Infinite Loops Podcast episode.
Ethan shares how startups can better use AI and talks about the things we can do when humans and AI work together. But AI can bring challenges too, like the problem of fake pictures. Still, those who start using AI early might gain a lot. What does this mean for old ways of doing business? Listen to Episode 165 of Infinite Loops to find out!
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.