Issue #31. The gloomy edition

My two new writing courses are finished, my extra weekly interview show is running smoothly, and nobody in my family has been in the hospital for a while. Life seems to be under control again (knock on wood), so let’s do some sidequesting.

On today’s quest:

— 19% of language professionals have seen AI-related job loss
— Even more illustrators and translators report job loss
— Have you really tried AI yet?
— Facebook’s AI is supposed to be great
— The data set is everything
— Quick hits

Scoop: Language professionals report seeing job loss to AI

A new Intelligent Editing survey of ~500 language professionals found that 19% say they have seen layoffs or hiring slowdowns due to AI. Another 20% said they “maybe” had seen such things, probably acknowledging that sometimes it’s hard to tell.

Many respondents said they want nothing to do with AI, but among those who dream of a world where AI can eliminate tedium, the top desires are help finding consistency mistakes, fixing formatting, and formatting citations.

See the full public results of the Intelligent Editing survey on the company blog.

The important point to sell to employers and clients is that while AI has been proven to improve the work of poor writers, it hasn’t done so well improving on already great writing. This is where your expertise still matters.

Illustrators and translators report job loss to AI

“About 26% of illustrators and 36% of translators reported losing work due to AI. Alarmingly, income devaluation is a concern for 37% of illustrators and 43% of translators. The survey also highlights a pervasive worry about future earnings, with over three-quarters of translators and illustrators fearing negative impacts on their income.” — HowToBe…

You haven’t used AI until you’ve had an existential crisis

Ethan Mollick of Wharton gave his management students an assignment to automate parts of their jobs and told them to expect to feel insecure about their abilities once they understood the capabilities of AI, the Wall Street Journal reported (gift link). A telling quotation was "You haven't used AI until you've had an existential crisis."

This fits with my ongoing advice that to really understand what AI can do, you need to use the best models and really dig into it. A cursory test with most of the free versions won’t you an accurate picture.

Facebook’s new AI is supposed to be great

In his email newsletter, Ethan Mollick also says nobody should be using the free version of ChatGPT anymore because Facebook’s free tool is much better, supposedly almost as good as the paid version of ChatGPT. You can try it at Meta.ai.

Note: It doesn’t look like you can upload documents or images for it to work with yet.

The dataset is everything

I recently saw interesting post a developer at OpenAI made back in July 2023 saying the only difference between models is the dataset — that when trained on the same dataset, the developer believed all the models would end up the same: “When you refer to “Lambda”, “ChatGPT”, “Bard”, or “Claude” then, it’s not the model weights that you are referring to. It’s the dataset.” An interesting point for people who are concerned that the models have been trained with stolen data.

Quick Hits

✏️ Business schools are now encouraging students to use AI as they race to prepare them for a new job market. — AOL

✏️ Amazon continues to be flooded with AI-generated books. This time, it’s loading ads for such books on ad-supported Kindle lock screens. — How to be…

✏️ Watch Google’s robots play soccer. — Science

✏️ When and how to cite AI — Grammar Girl (podcast, transcript, YouTube)

What is AI sidequest?

Using AI isn’t my main job, and it probably isn’t yours either. I’m Mignon Fogarty, and Grammar Girl is my main gig, but I haven’t seen a technology this transformative since the development of the internet, and I want to learn about it. I bet you do too.

So here we are! Sidequesting together.

If you like the newsletter, please share it with a friend.

^{Written by a human.}