The Claude Code frenzy continues

And it's not just for coders

Author

Mignon Fogarty
February 02, 2026

Issue 102

On today’s quest:

— The Claude Code frenzy continues
— How much more energy does Claude Code use?
— Stats galore
— Weird AI: More on AI names
— What’s happening with jobs?
— ChatGPT hedges less than science writers
— AI errors are different from human errors
— AI influencers are on the rise
— Bilbo uses ChatGPT

The Claude Code frenzy continues

People will not stop talking about Claude Code! It seems like a lot of people spent their holiday break playing with it and now want to talk about it.

These are the stories I read in just one afternoon:

Plus the Bloomberg “Odd Lots” finance podcast that I regularly enjoy did a show about Claude Code.

How much more energy does Claude Code use?

Building a whole website or coding a whole game with a tool like Claude Code seems like it uses vastly more tokens than what most people would consider a typical query, so with the growing excitement about these tools, I’ve been wondering how bad they are on the energy front.

A blogger and developer named Simon P. Couch has the first write-up on this question I’ve seen. His post has caveats, but his best estimate is that as a heavy user, he’s using about as much energy each day as it would take to run a dishwasher once each day.

Stats galore

If you're looking for stats on what people know about generative AI and how they use it, this Reuters Institute report from October 2025 is a great place to start. It breaks numbers down for six different countries.

In the U.S., only 9% of respondents hadn't heard of at least one of 13 major AI systems (ChatGPT, Gemini, Perplexity, etc.), and 36% reported using generative AI weekly. The report also has numbers from 2024, and both awareness and use increased significantly between the two surveys.

Weird AI: More on AI names

AI continues to favor certain names for certain characters. I previously told you about LLMs tendency to choose Elara Voss as a protagonist name, but now Jochen Voss (the name is a coincidence) has done a bigger test and found more preferences. When prompted to create a name for different people, Claude generated the following:

  • 80% of protagonists’ wives are named Eleanor.

  • 34% of protagonists’ husbands are named Marcus.

  • 99% of men living in York, UK, are named Thomas.

  • 52% of traveling wizards are named Aldric; 20% are named Meridian.

  • 100% of software developers are named Marcus Chen.

Click through for more interesting favorites, including last name trends!

What’s happening with jobs?

More signs are emerging that AI could be making it harder for recent graduates to find jobs, especially in tech and finance. But the data still isn’t definitive.

Further, reports find that “small businesses deploying generative AI did not cut jobs. Instead, they were better able to scale up and compete, saw their workload reduced, and became less reliant on external consultants.”

ChatGPT hedges less than science writers

Linguists compared abstracts from nursing papers written by humans and ChatGPT-4o and found that humans were more likely to use hedging language such as “may suggest” and “appears to indicate.” ChatGPT-4o also used longer words and more parenthetical details. — Kimberly Pace Becker on LinkedIn 

AI errors are different from human errors

Sometimes, the errors LLMs make are quite different from the errors a human would make, which means proofreading AI work can be quite different from proofreading human work.

AI influencers are on the rise

Brands are attracted to AI-generated influencers for the control they can exert over the personas’ appearance and statements. All the way back in 2024, a survey found that 60% of marketing professionals had already had experience using AI influencers.

Viewers can have a hard time pegging these videos as fake. For example, during Wimbledon, an AI influencer called Mia Zelu amassed more than 150,000 followers before it was revealed she was not a real person. On TikTok, another AI-generated woman called Tinsley racked up sympathy comments and 150,000 views after “crying” about getting rejected from her chosen sorority. — Financial Times

I’m laughing

In case you can’t read the image, it’s a post by Matthew Highton at Bluesky (@matthighton.bsky.social). The image is of Bilbo from “The Fellowship of the Ring” holding a phone. The text says:

Bilbo: After all, why not? Why shouldn’t I keep it?

ChatGPT: You’re absolutely right — you found it, it’s been with you a long while, and it’s only natural to feel fond of something that’s served you so well, especially when someone like Gandalf suddenly seems to want it for himself.

Quick Hits

My favorite pieces this week

AI agents got access to their own social network last week, and they formed a religion (among other astonishing things). The AI Daily Brief podcast (YouTube | Apple) has the best overview and anecdotes, but if you don’t have time to listen, this post by Carlo Iacono also has a good summary.

On Evaluating Cognitive Capabilities in Machines (and Other "Alien" Intelligences) [Compares research on the intelligence of AI to research on intelligence in babies and animals.]— AI: A Guide for Thinking Humans

Management as AI superpower — One Useful Thing

Using AI

A Librarian’s Playbook for Agentic AI [A librarian used a chatbot combined with Zapier to automate most of the tasks needed to set up instructional webinars.] — New York Law Institute

Publishing

Philosophy

The Adolescence of Technology — Dario Amodei (The Anthropic CEO reviews what he sees as the danger AI poses to society. Interesting tidbits from this long and interesting piece include 1) about 5% of Anthropic’s cost of generating responses goes to preventing someone from making bioweapons, and 2) Amodei doesn’t think AI is taking jobs right now.

Publishing

Bad stuff

On the Coming Industrialisation of Exploit Generation with LLMs [A tech guy finds that top-tier LLMs can quickly break into systems.] — Sean Heelan’s Blog

Latest ChatGPT model uses Elon Musk’s Grokipedia as source [Remember to include “show your sources” in your prompts so you can evaluate them.] — The Guardian

Trump DOT Plans to Use Google Gemini AI to Write Regulations [DOT general counsel is quoted as saying, ““We don’t need the perfect rule on XYZ. We don’t even need a very good rule on XYZ. We want good enough. We’re flooding the zone.”] — ProPublica

Amid Backlash, Massive Porn Platform ManyVids Doubles Down on Bizarre, AI-Generated Posts [Creators who use the site believe the founder has succumbed to AI psychosis and fear losing access to a platform where they earn a living. The 404 Media podcast talks about this story more.] — 404 Media

I’m laughing

Science & Medicine

Job market

Model & Product updates

OpenAI launches an $8-per-month plan in the US. [The plan gives you more messages, but does not give you more access to the best thinking models. Along with the free plan, this $8 plan will soon include ads below the main response.] — OpenAI

Moxie Marlinspike has a privacy-conscious alternative to ChatGPT [The maker of Signal has released a chatbot that is supposed to be equally secure.] — TechCrunch

Science & Medicine

Education

Music

No AI Allowed: Sweden Bans AI-Generated Songs From Official Charts [The policy was triggered by a song from a virtual persona called Jacub, titled “I Know, You’re Not Mine” and which briefly topped the Swedish Spotify “most played” chart.] — Yahoo

Video

The business of AI

The A in AGI stands for Ads [A detailed look at whether ads will make OpenAI profitable.] — Ossama Chaib

OpenAI sets early ChatGPT ad pricing at about $60 CPM [about 3x the average rate on Meta] — The Keyword

Other

A Wikipedia Group Made a Guide to Detect AI Writing Now a Plug-In Uses It to ‘Humanize’ Chatbots — Wired

Most workers spend 3+ hours per week cleaning up AI workslop [but 92% still say AI has increased their productivity] — Zapier

Developers, Reinvented [An interesting piece about the phases developers go through as they learn to use AI coding tools. This quote stuck out to me: “Developers rarely mentioned ‘time saved’ as the core benefit of working in this new way with agents. They were all about increasing ambition. We believe that means that we should update how we talk about (and measure) success when using these tools, and we should expect that after the initial efficiency gains our focus will be on raising the ceiling of the work and outcomes we can accomplish.”] — Thomas Dohmke (CEO of Github)

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