ChatGPT

OpenAI, the creators of popular generative AI platforms ChatGPT and DALL-E 2, says the tool it developed to detect content written by AI only works 26% of the time: "While it is impossible to reliably detect all AI-written text, we believe good classifiers can inform mitigations for false claims that AI-generated text was written by a human." I would not classify a classifier that fails three-quarters of the time as a "good classifier." Just sayin'. Continue Reading →

Everyone is talking about generative AI products like ChatGPT, Midjourney, DALL-E 2, and Stable Diffusion (among dozens more), so it's clear that we're on the cusp of something exciting and new. What no one knows is where value will accrue. Continue Reading →

My Sunday essay was about Generative Synthetic Media. If you don't know what that is, have a quick look at the post, then go play around (pun intended) with some examples you'll find in this Google Research paper: MusicLM: Generating Music From Text. Continue Reading →

We know that OpenAI (the company behind ChatGPT) is working on a pricing plan, and some users are reporting that they’ve been granted access to a pro tier that costs $42 a month. Fans of "The Hitchhiker's Guide to the Galaxy" know this number well. In the book, the answer to the "Ultimate Question of Life, The Universe, and Everything" is 42, which makes me wonder if the final pricing will be announced on May 25. More importantly, I have been wondering if ChatGPT is, in fact, a pseudonym for Deep Thought. (Apologies to non-Douglas Adams fans.) Continue Reading →

What possible safeguards could be (or have been) put in place to prevent a bunch of mischievous teenagers (and I'm being careful to choose my words here) from crafting some very nasty code? Continue Reading →

At the 2023 World Economic Forum, Microsoft CEO Satya Nadella told the WSJ that the company will incorporate AI (such as ChatGPT) into all of its products. This raises a question: Do you wait to learn to use the version of ChatGPT that's built into Microsoft Office (Word, PowerPoint, Excel, etc.), or will you start learning how to use native ChatGPT right now? Continue Reading →

I understand why teachers are worried about students using ChatGPT, but let's not forget that students have always found ways to cheat, chatbots or not. I think it's more important to focus on how we can use these tools in a positive way in education. Continue Reading →

The article “What Are NSF Fees and Why Do Banks Charge Them?” on CNET is a serviceably written explanation of NSF fees. The byline is simply "CNET Money." If you click the byline, it says: "This article was created using an AI engine and reviewed, fact-checked and edited by our editorial staff." Continue Reading →

OpenAI has announced plans to begin charging for access to ChatGPT, a widely popular AI-powered chatbot capable of composing essays, emails, poems, and even computer code. In a statement on the company's official Discord server, OpenAI mentioned that they are considering monetizing ChatGPT as a means of securing the tool's long-term sustainability. Continue Reading →

My Sunday essay, "ChatGPT In Its Own Words," was mostly written by ChatGPT. It's impressive by every measure. I was remiss, though, in not reminding everyone that the system is designed as a research tool and that it is also in "free" pre-release mode. Continue Reading →