My Top Ten AI Predictions for 2024
Everyone's an expert now
True confessions, I let AI tools generate the list. I probably could have written this off the top of my head, but I posed the same question to ChatGPT, Claude, and Perplexity, and Bard. I looked for patterns, rejected the ridiculous ideas, and combined it with what I already know. Still, I’m willing to stand behind it and I’ll revisit in a year to see how I did. Read the whole post on Medium
Required Reading: The Best Link I’ve Found So Far
This is my sixth month and 55th installment of this newsletter; where I’ve tried to make sense of and figure out what’s next in the intersection of Design and AI. This article from the Designer Fund gave me more answers, clarity, and optimism, than anything else I’ve read. The future isn’t perfect, but if you’re trying to navigate those choppy waters and figure out where you fit in, I think you should check it out
IBM’s AI Academy
Thanks to my old friend Boaz Ashkenazy and his AI newsletter, because that’s where I learned about IBM’s AI Academy (for a general overview) and IBM’s Skillsbuild (professional development). Yet another tech behemoth offering free AI training because they know (hope?) what their future employment needs will be. Add this to previous mentions of free AI training from Google, Microsoft, LinkedIn, and Amazon.
Artificial Intelligence (AI) is rapidly becoming integral to various industries, underscoring the necessity of ethical AI education. This is vital to ensure AI is used responsibly and beneficially. IBM's CEO, Arvind Krishna, has highlighted the importance of AI training, with the company setting an ambitious goal to educate two million people by 2026. This initiative addresses the global AI skills gap, preparing individuals for AI's growing presence in different sectors.
Google’s Music FX
I do not possess any musical talent and I couldn’t explain anything about how music works or how it’s made or when it’s good. However, I know have a generative tool that will allow me to make just about any kind of soundtrack for my video projects. I first used there v1 audio generator last summer to produce some clever 20 second samples. This is a giant leap forward and it can produce 30, 50, or 70 second passages and they can be set to loop. The entire interface is gorgeous, intuitive, and fun to use. Google has really put some thought into this one, and it wont generate vocals or allow you to reference any existing artists. Try it out | Read about it | Listen to my sample
Google’s Imagen and Why it Matters
So this might just be another example of Google over promising and under delivering; and I did say in Monday’s newsletter that they could no longer be trusted after the Gemini stunt. However, there a many innovative divisions within Google’s AI sphere and some move more boldly than others. While I don’t have access and I haven’t seen any images yet, this report from TechCrunch says their Imagen model can generate images with text and logos, as well as a layered document. While this is currently the holy grail for all the image generators, it could mean more bad news for entry level graphic designers. Check it out
Imagen 2 can generate … emblems, lettermarks and abstract logos … [and] has the ability to overlay these logos onto products, clothing, business cards and other surfaces.
Midjourney Prompt Libraries
Several of my students have asked, “Where are all the good Midjourney prompts?” While I can’t swear by all of these and I still think it’s best to experiment and build your own repertoire through trial and error, here’s a whole bunch of resources in one article. Check it out
Visual Electric is Off To a Great Start!
While the rest of us are waiting for our web-based image generators (see Midjourney, below). Visual Electric is off to a roaring start. I’m really interested in how to build a usable interface for generative AI and we don’t have a lot of mental models for this. Adobe Firefly is good, but the results are mixed. Leonardo.ai is great considering it’s fixing almost all of the problems related to Stable Diffusion, but a little overwhelming. Ideogram is fun to play with, but still a toy. Enter Visual Electric, which I heard about today, and I’m very impressed with. Close to Midjourney in quality, but web-based and a delight to use. It’s more of a brainstorming tool than a creator of finished products, but I’m going to commit $20 for a month and have a full report soon. See my 3 minute quick tour, below.
I’ve Only Generated 4000 Images in Midjourney!
Midjourney’s new web based interface is available, but only to people who have generated 10,000 images. Tim, from Theoretically Media, has a full review (The video seems restricted, but it’s worth a trip to YouTube). Once again, I’m interested in the UI of the controls and from what I can see, they’ve done a pretty good job. Hopefully, they’ll relax that 10k limit soon, but if you’ve reached it, you can try it out at https://alpha.midjourney.com/. The next tier of invites happens for people who’ve generated 5000 images, so I better get to work.
That’s a wrap! 10 days until Christmas. I hope you’re having a great holiday season and you’re getting a lot of value out of this newsletter. Are people constantly commenting about how informed you are? Let them know that they can be just as smart as you and forward this newsletter.