Who are you automating for?

Happy Monday and welcome to another issue of Artificial Insights, your weekly guide to augmented ideation.

I have been consumed by a research project mentioned earlier, the creation of an index of AI related concepts and terminology. A significant aspect of learning a subject is knowing the terms used to define the field. AI has evolved to include hundreds of ideas over the decades, and it seems we each know a small part of the whole. It turns out LLMs have tremendous self-knowledge, so I have been leveraging lots of GPT to develop and fact-check over 200 AI-related concepts and terms. Currently published as an Obsidian vault, and we are also working on non-linear methods for visualizing and navigating these ideas.

The primary reason I’m doing this is to help more people learn about the theory underpinning the largest technological revolution of our lifetimes, and maybe help us make sense of what we are experiencing.

This week’s edition is a small selection of what was featured on our new WhatsApp group 💬. While not for everyone, the group has quickly become my favorite place to share and comment on the news, so please join if this resonates.

Until next week,
MZ

Amazing lecture by Cassie Kozyrkov outlining the implications of an increasingly autonomous economy in great detail. Both technical and easy to understand. Highly recommended.

And this brief WSJ interview with OpenAI's Mira Murati talks about the underlying data used to create Sora with a few new videos.

Ethan Mollick from Wharton is arguably my favorite voice in the generative AI space. This post from a few hours ago talks about his approach to Centaurs (as opposed to Cyborgs), and explains how he used LLMs to write an upcoming book:

[

One Useful Thing

I, Cyborg: Using Co-Intelligence

I wrote a book about living and working with AI, Co-Intelligence, and it is coming out on April 2. If you like my Substack, you will like it, or at least that is what early readers are telling me. And, given that my publisher has told me that pre-orders are really important to a book’s success, if you want to get it, now would be a great time…

Read more

2 years ago · 211 likes · 50 comments · Ethan Mollick

](https://www.oneusefulthing.org/p/i-cyborg-using-co-intelligence?utm_source=substack&utm_campaign=post_embed&utm_medium=web)

In the WhatsApp group we collected several sets of custom instructions to make ChatGPT more objective and easier to work with. Take a look: https://pastebin.com/u0GrHwMj 

Insightful.

If you don’t mind listening to four white guys speaking for hours about AI risk, this podcast goes deep on the long term implications.

Amazing strategy and development insights from Zoe Scaman.

Not sure I agree with the characterizations but definitely with the sentiment. I’ve noticed how much effort goes into getting to know the “personality” of a new bot, and to update your mental model of how it responds. I expect there to be significant inertia for moving to new models/apps for most people.