Issue 80

On today’s quest:

— How to use AI audio in Google Docs
— Just weird: LLMs’ musical taste
— Word watch: Oneshotted

How to use AI audio in Google Docs

My number one tip for proofreading is to listen to your work being read aloud, and that just became a lot easier in Google Docs.

Right now, I keep a $5 a month Eleven Labs account for just this purpose. I copy and paste my writing into the service, and it reads the work to me in an expressive, natural-sounding voice.

Google Docs uses one default voice, whereas Eleven Labs gives you many choices (which isn’t a big deal), but I don’t like how Google Docs reads the entire page, not letting you select which part you want to hear. Despite that drawback, it sure is handy having the audio right where I’m doing my writing. I hope they add more control over the process in the future.

Just weird: LLMs’ musical taste

Long-time readers know my favorite kind of story is a “weird AI” story, and we got a new one this week!

Tyler Cosgrove tested the musical “taste” of different LLMs by having them pick musicians they preferred in head-to-head matches from a list of 5,000 most-played artists. This was the prompt:

The ultimate winners list for each reasoning model was made up almost entirely of names that began with symbols or numbers. Here’s an example:

Non-reasoning models produced lists with much more “normal” names:

So far, I’ve seen two theories about why reasoning models would favor numbers and symbols:

1. It could relate to the models being overtrained on sorting tasks.

2. It could relate to the models having extra training on mathematical reasoning.

(Tyler said he was partially inspired by a post about the way different models made geographic maps by Henry, and it’s a fun one too.)

Built for Managers, Not Engineers

AI isn’t just for developers. The AI Report gives business leaders daily, practical insights you can apply to ops, sales, marketing, and strategy.

No tech jargon. No wasted time. Just actionable tools to help you lead smarter.

Start where it counts.

Work smarter daily

Word watch: Oneshotted

“Oneshotted” has been adapted from gaming and drug use to AI.

Gaming. In gaming, “oneshotting” something is killing it in one shot in a first-person shooter game.

Drug use. In drug use, people get “oneshotted” when they have a life-changing reaction to one dose of a drug, like when someone’s brain is permanently altered by a strong hallucinogen.

I’ve seen two uses related to AI:

1. A oneshot is the result you get from just one prompt. For example, someone might post an AI-generated image and brag or qualify that “This is a oneshot!” or they might show you a vibe coded app and say, “I oneshotted this.”

2. Getting oneshotted is also used to describe AI delusions, which seems to be playing on the drug-use meaning even though it’s illogical because AI delusions typically happen after much more than just one exchange.

Urban Dictionary also has a few uses related to different kinds of short fiction.

Quick Hits

Using AI

I Tested How Well AI Tools Work for Journalism. Short summaries were much more accurate than long summaries. AI research tools were not useful for scientific literature reviews. — Columbia Journalism Review

Psychology

I share all my deepest thoughts and feelings with ChatGPT – but our friendship is doomed — The Guardian

Legal

Anthropic AI Class Action: Important Information for Authors — The Authors Guild

Writing

ChatGPT Shaming Is Making Our Writing So Much Worse. In the age of A.I. paranoia, people are cutting em dashes, skipping metaphors, and leaving in typos to prove their human. — Slate

Signs of AI Writing — Wikipedia

Medicine

Bill Gates funds $1mn AI Alzheimer’s prize. New competition is latest effort to deploy artificial intelligence to find cures for serious illnesses — FT

Education

How I am thinking about GenAI in the classroom. Student use of GenAI is not the problem, it's a symptom of bigger problems. — Science For Everyone

One writing teacher’s approach to addressing AI (PPT) — Anna Mills

AI is moving so fast, a PhD might be obsolete before you graduate — Windows Central

Cybersecurity

How AI-enhanced hackers are stealing billions. It is a boom time for cybersecurity firms. — The Economist

Bots now make up the majority of all internet traffic — The Independent

Other

95% of generative AI implementations in enterprise 'have no measurable impact on P&L', says MIT — flawed integration cited as why AI projects underperform — Tom’s Hardware

A.I. May Just Be Kind of Ordinary — New York Times

DeepSeek V3.1 just dropped — and it might be the most powerful open AI yet — VentureBeat

Spanish Tourists Say ChatGPT Ruined Their Vacation — And It Could Happen To You, Too — HuffPost

Draftsmith

A special thank-you to Draftsmith — an editing assistant in Word that supports the way you work. Draftsmith has supported the newsletter by signing on as a premium sponsor for a whole month of Mondays. 🙂 Get 30% off with code SIDEQUEST30.

What is AI Sidequest?

Are you interested in the intersection of AI with language, writing, and culture? With maybe a little consumer business thrown in? Then you’re in the right place!

I’m Mignon Fogarty: I’ve been writing about language for almost 20 years and was the chair of media entrepreneurship in the School of Journalism at the University of Nevada, Reno. I became interested in AI back in 2022 when articles about large language models started flooding my Google alerts. AI Sidequest is where I write about stories I find interesting. I hope you find them interesting too.

If you loved the newsletter, share your favorite part on social media and tag me so I can engage! [LinkedIn — Facebook — Mastodon]

Written by a human (except the examples below, obviously)