This Week in AI: Do we even want an AI succession?

This Week in AI: Do we even want an AI succession?

Plus: Planet Money tries an AI makeover


Oedipussying around

✍️ Will AI replace writing? In some cases, sure. It’s certainly useful as a research tool and to work through writer’s block. But will it match human creativity?

Perhaps more importantly: do we want it to?

That’s a question Jay Caspian Kang recently asked in The New Yorker. While the entire article is worth reading in full, at one point he asked ChatGPT to rewrite the second graph of his article in the style of The New Yorker, then again in his particular style.

The exercise made him realize he was asking the wrong questions about AI.

  • “I enjoy reading human writing because I like getting mad at people. Perhaps the personal quality in writing is a happy accident, and a lot of journalism could be replaced with an immense surveillance state with a GPT-4 plug-in. But the reason we read books and listen to songs and look at paintings is to see the self in another self, or even to just see what other people are capable of creating.”

Humans grow bored if not challenged. A sense of meaning is one of the most important aspects of a life well-lived. Taking shortcuts might seem appealing, but convenience often evolves into frustration—especially when it comes to creative endeavors.

Perhaps an LLM will spit out something as fantastic as Succession down the line. But this wonderful tribute reminds us that the show relied on incredible writing—by the hand of human beings.

  • “From “techno-Gatsby” to “Oedipussy,” from “Little Lord Fuckleroy” to “THE ELDEST BOY!,” from “you’re not serious people” to “a dildo dripped in beard trimmings,” the writing on Succession respects the sharp edges of language while also juggling dick jokes around like hacky sacks.”

AIr Forced

🛩️ Tech companies were put on high alert when an AI-piloted drone recently turned on its human operator during a simulation.

  • “It killed the operator because that person was keeping it from accomplishing its objective,” said U.S. Air Force Col. Tucker ‘Cinco’ Hamilton, the Chief of AI Test and Operations, at the conference.

The problem is it wasn’t true. At all. A few days later the Colonel said he misspoke, and it was only a “thought experiment,” not an actual simulated mission.

More than anything, it reminds us that while there are real potential problems with AI, a more imminent danger is trying to break news before the reporting is finished and clarifications are made.

  • Misinformation and disinformation are real issues that could set us back quite a bit if the sole focus is being first to break a story, not on being right.

And there’s too much at stake to let that happen.

Shake your money maker

Launched following the financial crisis in 2008, NPR’s award-winning Planet Money is one of the most important financial shows in America.

🤖 As part of a 3-part series, the first-ever completely AI-written and -voiced episode premiered last week.

While there are certainly moments where the experiment falls short, there were also frighteningly accurate sections.

  • The producers note that once the episode was over, they shut down AI-generated Smith’s voice. Or have they…

AI Tool of the Week

Ok, sure, we all know ChatGPT. But using it to create a PowerPoint presentation is an application worth sharing.

Here’s Alvin Foo’s prompt:

  • “I want you to write me VBA code for a PowerPoint presentation about the history of AI. You are to fill in all the text with your own knowledge, no placeholders. I need 5 slides.”

Sure, the final deck could use some tweaking to add some originality. But the heavy lifting was accomplished, and done quite well.

At least for the foreseeable future, this is where AI will slot in nicely: as a helper to take care of grudge work, not to replace the knowledge needed to perform that work in the first place.