AI Insider No. 9

Stop doom scrolling Supreme Court decisions and read this. Surprise! This entire edition of the newsletter is written by me just for you. No Bing, no Bard, no ChatGPT bylines. Don’t get too used to this, tho. I just happened to be in the mood to yammer this week. But, as most of you know, I’m retired. Newsletter writing is a full-time job. I’m not trying to go there. So, enjoy this occasional solo bylined missive! The artwork is AI generated as usual because I am no artiste.

AI is All About Me

By Michelle Johnson

As we’re being told practically every day, there’s a lot going on in AI. Much of it falls in the category of funding. But, honestly, my eyes glaze over when folks start talking venture capital and mergers and acquisitions. Blah, blah, blah. What I’m thinking is: What’s in it for me?

For instance, a text-to-video company called Runway raised $141 million this week.
I’m not going to yammer about that. Instead I’ll tell you about how easy it was to create this morphing video below using Runway. It was just a matter of figuring out which one of Runway’s tools I needed to use (Frame Interpolation), uploading my images (created in Midjourney, dragging and dropping to order them and hitting “generate.” I pushed a slider around to adjust the timing, then boom. Done. In just a few minutes I was downloading the final result. I hope the cash infusion allows Runway to really take off. (Pun intended.)

Oh, Bard. You’re Such a Baby

This week the Supremes put a nail in the coffin of several things near and dear to my heart. So I tried to keep up. When I asked Bard to give me the scoop on the affirmative action decision he messed up bigtime, not unlike the court.

First, he incorrectly told me that the court hadn’t ruled yet. When I pushed back and said, “Oh, yes, they have. Check again.” Bard apologized and then proceeded to tell me that Harvard and UNC had won and Clarence Thomas had written the majority opinion about said win. Wrong and wrong.

Try again, Bard. The third time was the charm and a good reminder that these tools are in beta. They are toddlers learning to walk. You need to keep an eye on them lest they go tumbling down a set of stairs or something.

And speaking of Bard, if you follow me on Facebook you may have seen me report on getting up close and personal with Bard this week. I posted a few excerpts. If you’re interested, here’s the full text of our delightful chat.

ChatGPT Mobile Now Connected to the Interwebs

If you’re a paying ChatGPT Plus user (like me), you likely know that you can search the web via ChatGPT 4 on your desktop or laptop via a plugin. This week OpenAI announced that the iOS ChatGPT app has hooked up with Bing, meaning it’s now internet connected. Why is that a big whoop? How many times have you asked ChatGPT for something and it responds: I only know stuff that happened up to 2021. Well, now it can search the web for more recent and hopefully relevant info on your phone, too, presuming that you’ve downloaded the ChatGPT app from the App Store.

Deep Thought

I’ve learned that there’s a name for people like me who have too much time on our hands so we’re messing around with AI like crazy: AI Enthusiast. Yes, that sounds about right. I have zero background in AI. And I don’t have to do any of this for work. I’m just enthusiastic about this technology despite all of the doomsday predictions. If I’m going down, I’m going down having some fun!

Just the other day I was listening to a live chat with David Holz, a founder of my favorite AI image generator, Midjourney. Someone asked a question that’s stuck with me these past few days because I think it partially explains my addiction to this format.

The question: Was the plan to develop Midjourney to appeal to designers and other “creatives” … or not? Holz was emphatic that they did not set out specifically to reach that market. If anything, what’s happened is even more interesting. They have created a magical tool for people who couldn’t design their way out of a paper bag..

Sure, professionals have incorporated generative image AI into their workflow with spectacular results. But us mere mortals who are far from visual creatives are also blowing up stuff. Just for the hell of it because the tool doesn’t get in the way.

There’s a reason that Midjourney’s prompt begins: /imagine. And DreamStudio’s prompt begins: /dream. Kudos to all of these visionaries for opening the portal to this magical place.

Reality check: Yes, I know these tools were trained on other people’s work. Mostly they are smart enough to create something that’s derivative. Still, I wish there was a way to get everyone paid.

Fun Stuff to Try

I’ve been really impressed that Adobe came up with text to image maker Firefly, then added generative fill, which lets you erase a section of an image and replace it with something else. (OK, my journalism friends. Breathe. Just breathe. This is not meant for journalistic use. Yes, I know it will be abused, but Adobe is working on stopping this.) Anyway, it’s not surprising that lots of startups are jumping into this space. Here’s a new one called Playground. It’s easier to show you what it does rather than explain it.
Watch the video and sign up here: playgroundai.com

Let’s face it. Therapy is expensive. Save a buck and try a chatbot instead: LotusTherapist
(Note: This is not medical advice! It’s a fun thing to try out and learn about how chatbots work. That’s it.)

Aht Gallery (Yes, I spelled this right. This is Boston, dammit!)

That’s it for the latest edition of AI insider. I hope you enjoyed reading it and learned something new. If you have any feedback or suggestions for future topics, please let me know by replying to this email (if you’re a subscriber). Thank you and see you next time! 😊