^ArchiveArticles

Here is a quick overview of how ChatGPT operates under the hood. This will make it easier to spot suspicious behavior. The following is at a very high level.  For the gory details, see the following two guides: – The Illustrated GPT-2 (Visualizing Transformer Language Models) – Jay Alammar – Visualizing machine learning one concept at a time. (jalammar.github.io) – The GPT-3 Architecture, on a Napkin (dugas.ch) What is ChatGPT? Let’s start with what ChatGPT is. ChatGPT is a kind of machine learning algorithm known as a neural network. To understand what a neural network is, recall your algebra classes. You remember being given a set of equations and being told to solve for some variables. Then you learned you Read More ›

We’ve covered ChatGPT, OpenAI’s new text generator, quite a bit the last few weeks here at Mind Matters, but lest we forget, AI has made great strides in image generation, too. Another explosive AI tool is Lensa, an app that uses Midjourney’s Stable Diffusion model. Lensa was first developed by Prisma Labs in 2018 as a photo editing tool, but in late 2022, it added the “Magic Avatar” feature, which allows users to recreate their portraits artistically. Since then, it has skyrocketed to popularity alongside its AI cousin, ChatGPT. Despite its reputation as a “fun app,” some users have complained that the app produced sexualized and racialized avatars of themselves. In an Insider article, Bethany Biron writes, Melissa Heikkilä, a Read More ›