AI round-up: Week of October 7, 2024

“You can no longer trust that anything you see, or hear, or read was not created by AI.

Humans, walking and talking bags of water and trace chemicals that we are, have managed to convince well-organized sand to pretend to think like us.”

-Ethan Mollick, Co-Intelligence

That seems like a rather harsh oversimplification.

Then I read his latest blog, which featured an audio component. It was the podcast output from Google Notebook.

So … yeah.

the BIG five

1. We’re going to have to let you go. You’re doing just too good of a job.
Huh?

Imagine this. You build, train and perfect your digital twin. They’re doing such a great job your company thinks it can do your job.

You think, ‘Nice try … but if I leave, my twin is coming with me.’

“Not so fast,” the company says. “Where was that twin trained?” (Hmm … my work computer.) “When was it trained?” (Dang. On company time!) “Who paid for you to train it?” (Ok, ok … I see where this is going.)

Far fetched? Nah. Nothing really is anymore. Recently we had our MAICON roundtable regroup with some clients and friends. This question came up as we talked about digital twins. Who owns the twin? One response was, ‘I do; I trained it.’ Another response was, ‘Our legal department might have something to say about that’. It was a fascinating conversation.

And then, one day later, I saw this article: Zoom will let AI avatars talk to your team for you. So does Zoom own your twin?!?

Side note: guess what someone asked in the comments?

digital twin question

2. In case you didn’t get enough of Google NotebookLM…
You can check out the podcast I made. The source is my MAICON summary. The result is mindblowing. These aren’t real people!!!

3. Breaking down the AI productivity suites
If you missed it, ChatGPT introduced Canvas, its entry into the producitivity suite discussion. Canvas joins Claude Artifacts and Google’s NotebookLM. And one thing is painfully clear – these AI companies SUCK at naming products. If only they had a powerful tool to help them be more creative.

Anyway, Shelly Palmer does a great job breaking down the benefits of each option. As you’ll see, the answer to which one may be all of the above.

4. Hey, at least we asked.
Udemy, a large e-learning platform, has opened—and closed—its opt-out window for anyone who didn’t want AI learning on their content. That’s 250,000 online learning courses, by the way.

The company is catching heat because the opt-out option was only available for two weeks. But hey, at least they asked, which is more than some companies do. (Looking at you, LinkedIn!!!)

Have we hit that point where we should assume most of our content is feeding the beast?

5. 95% you say?
I'm just curious at what point the narrative switches from AI won’t take jobs.

Learn a little

Adobe Photoshop and Premiere Elements have been updated to feature new AI.

Do you remember that scene in Clueless where she uses her computer to find the perfect outfit? That’s what these features remind me of. Clueless, and Paul Rudd, have aged well.

Did you hear…

…Chromebooks are getting a new button dedicated to Google’s AI? (The Verge)

…you can ‘say no to your AI overlords’? (The New York Times) I guess some people don’t like AI answering all of their questions or trying to write all of their stuff.

…Google is working on its own reasoning AI model, similar to OpenAI’s o1 model. (Bloomberg) The question is, will this also be named after a fruit?

…CoPilot now has a voice and can see and hear what you do while browsing? (Digital Trends)

…the ‘godfather of AI’ won a Nobel Prize in physics (Bloomberg)

Must read/must discuss:

We’re not so good at detecting AI content from real content. But hey, we’re only human.

Well, that’s not gonna cut it when it comes to cyber-attacks. Axios does a great job breaking down just how serious the threat of this gap is. And maybe more importantly, how vulnerable we are without any type of regulation around AI-generated content.

It’s not the rosiest of articles, especially around election time, but something I think we should all be aware of and try to do our best in spotting this type of content. Question everything (sadly).

-Ben

As a reminder, this is a round-up of the biggest stories, often hitting multiple newsletters I receive/review. The sources are many … which I’m happy to read on your behalf. Let me know if there’s one you’d like me to track or have questions about a topic you’re not seeing here.