📫 Subscribe | 🗣️ Hire us to speak | 🤝 Sponsor this newsletter | 🎓 Events

Welcome to the Friday edition of our newsletter. We spend Fridays going deeper into tools and trends related to generative AI (and Tuesdays sharing news updates). This week, we both react to the Bucksbaum Lectureship in Business.

Wednesday night, we had the opportunity to attend Drake University’s Bucksbaum Lectureship in Business, featuring AI journalist Karen Hao and The Atlantic CEO Nicholas Thompson. The conversation offered two different perspectives on a technology we spend a lot of time teaching people to use.

Hao described today’s AI companies as empires, drawing parallels to historical empires as they accumulate land, consume enormous amounts of electricity and expand their influence. Their leaders are making decisions with consequences for all of us, but we don’t elect them.

Thompson offered a more optimistic perspective, sharing how he is using AI to help The Atlantic grow its subscriber base and support hiring more journalists (not to mention using it to find the perfect running shoe).

The conversation gave us plenty to think about, both as people who use AI and as people who help others make sense of it. Here’s what stood out to each of us, what challenged our thinking and what we’ll bring into our work.

Your client doesn't care how their site was built. They care that it launches on time, runs on their own domain, and keeps working long after you hand it over.

With Lovable, you describe what your client needs and get a working version you can show them in the first meeting. Then you take it all the way to a live site on your client's domain, built to be found on Google. Lovable's project monitoring feature spots bugs and suggests fixes, so your name stays on work you're proud of.

Start today and take $5 off your first month.

Build for clients, $5 off

_{^{Discount of $5 applies to first month of any paid plan. You need to sign up via this webpage in order to claim the discount. The discount applies to new and returning users without an active paid plan. Terms apply.}}

Chris Snider: Thompson explained how The Atlantic uses AI to identify articles likely to turn readers into subscribers and then promote those articles. That hit home because it’s exactly how my family recently became a subscriber. It also hit home because I am a journalism professor, and anything we can do to help drive subscribers and save journalism is near and dear to my heart.

Chris Porter: I was struck by how Hao and Thompson can have a fairly strong level of disagreement, particularly with respect to use of commercial LLMs like ChatGPT and Claude, all the while sharing a strong commitment to raising awareness of the many legal and ethical challenges associated with this technology (for Hao in her New York Times bestselling book Empire of AI and for Thompson in his work managing The Atlantic, which has recently published many critical pieces on current AI developments).

Chris Snider: Hao challenged how I think about open-source AI. Like Thompson, I’ve been more hesitant to trust Chinese models than U.S. models. But she made a compelling case that running an open-source model on your own device can give you more control over your data than handing it to an AI company. It made me rethink whether I’ve been putting too much emphasis on where a model comes from and not enough on where it runs and who has access to what I share.

Chris Porter: Prior to Wednesday night’s discussion, I hadn’t incorporated Hao’s account of the major AI companies as empires into my own thinking. However, as I was sitting in the audience listening to Hao make a case for why some of the behavior of these companies looks like a metaphorical land grab (and in some cases, even a literal one), this depiction resonated well with me as explaining why mathematicians have been so upset with OpenAI for dropping over 700 AI-generated math publications solving nearly 400 open mathematical problems (discussed in last Friday’s newsletter): with that release, OpenAI planted their territorial flag squarely in the middle of world of mathematical research, releasing papers with “OpenAI” listed as an author, completely bucking the standards of the mathematical community for developing, releasing, and communicating new mathematical developments.

Chris Snider: Thompson shared the idea of having a “purity line,” a boundary you won’t cross in your AI use. For him, that’s writing. I think we all need to consider where that line belongs as these tools become capable of doing more of our jobs. Just because AI can do something doesn’t mean we want to hand it over. Some work matters because doing it ourselves helps us learn, express ourselves or find meaning in our jobs. I’m adding this idea to my presentation at AMA Iowa’s Experience event next week, asking participants to think about their own line and why it matters.

Chris Porter: Hao made the point that the term “AI” functions much like the term “transportation,” in the sense that “transportation” doesn’t refer to one specific type of device but a fairly broad collection of devices that includes bicycles and rockets. Just as forms of transportation come in different shapes and sizes, propelled by different means and capable of conveying varying sizes of groups of passengers, AI systems range from the small to the large in domains such as size, scope of application, and resources required to train and deploy. I found this analogy to be helpful enough to make reference to it in a class discussion just yesterday on how different conceptions of human intelligence might potentially map onto different approaches to building AI systems.

Did you miss Wednesday night’s discussion between Karen Hao and Nicholas Thompson? Check out the recording here.