Exploration #100

Hate Longscrolling? Jump Around!

• If You Click Only One…

• Things to Think About…

• Things to Know About…

• And finally…

Welcome to the 100th issue! This week the focus is on AI as we look at the newest version of Claude, a new text-to-image generator that can reliably create in-image text, the state of AI filmmaking, and more. But first…

An AI Face for Public Media?

Last week I got a cold email in a secondary inbox that I check every few days. Most cold emails barely register on my radar, but this one ended up providing food for thought. The general pitch was for creating a custom avatar for the organization.

It actually makes a lot of sense. Public media organizations think about brand voice in terms of language used, and brand voice in terms of the actual sound of voices that represent us. And we think about how we are visually represented by talent, staff and leadership both on our air and in our communities.

So if you could create an avatar, what would the personification of your public media organization be?

The cover art for this issue of the newsletter is actually a quick stab at answering that question at a national scale. I actually asked the new generative art tool Ideogram to create for me an image of a person that what it ‘imagined’ a typical PBS viewer to be. And here’s one of Ideogram’s takes on a typical NPR listener:

Not every generative image hit the mark (Ideogram still has problems with fingers and, in the case of the NPR listener attempts, tote bags), but given that I guided the creation of these ‘people’ in less than 30 minutes, imagine what you could do with the fully fleshed-out personas that many of our marketing teams have already drafted. And then, with a tool like Alibaba’s EMO project, a standard ‘support public media’ script, and a voice synthesizer like one of the ElevenLabs tools, don’t you have the beginnings of virtual representation?

You can create an avatar for your station or organization right now. It’s probably not quite ready for pitching the Bee Gees special just yet, but in a year it could potentially be handling customer support after normal business hours. What would your organization’s avatar (or avatars) look at and sound like?

Thoughts and Thanks on Hitting ‘100’

Before I turn you loose on this week’s articles, and head off to finish packing for South by Southwest, I just want to say thanks to you.

While this is only the 9th exploration published to the beehiiv platform, it is actually the 100th exploration that I’ve sent out since I started this newsletter on Outlook just over two years ago. There were only a couple dozen of you on that first email and the tone back then was extremely informal. So the newsletter has grown up (a little).

I want to give special thanks to Amber Samdahl, David Huppert, Mikey Centrella, Debbie Hamlet, Jessica Mitchell-McCollough, and Liz Maestri for very specific guidance and support over the last couple of years.

Similarly, I want to thank the Nebraska Public Media Labs team. This newsletter is the product of an ecosystem of experimentation and innovation that they help maintain on a daily basis. I am inspired by their excitement for new ideas, their courage to take professional risks, and their willingness to be informed by - but not chained to - the station’s past successes.

I also want to thank my GM, Mark Leonard, and my colleagues on the Leadership Team at Nebraska Public Media for giving me the space and time to create something like this for the public media community. And, last, but by no means least, Eric, Tim, Tim, Tonya and Maria at NETA for helping this effort level up with the creation of the Public Media Innovators Peer Learning Community.

I was lucky to spend the first 10 years of my public media career as a television programmer. That discipline has a clearly defined and supportive community within the larger world of public media. Your support of this newsletter and now the PLC has only reaffirmed my belief that there is a similar group of innovators/mavericks/restless souls within public media that should be clearly defined and are worthy of a supportive community.

Thank you for inviting me into your inbox each week.

Okay, on to the links…

If You Click Only One…

10 AI terms everyone should know (Susanna Ray - Microsoft) - Here's a nice little interactive graphic that succinctly defines: AI, machine learning, large language models, generative AI, hallucinations, responsible AI, multimodal models, prompts, copilots (which wouldn't be a thing if this weren't coming from Microsoft) and plugins. Send this to anyone (team members, your boss, your board chair, etc.) with whom you want to have a common lexicon for discussing policy and procedures around using AI in your organization.

Things to Think About…

You Can Make a Movie with AI (in 60 Minutes) (Dave Clark & Dan Shipper - Every) - A video interview (there's a podcast version as well) that shows you Clark's creative process using Midjourney (imagery), Runway (animation) and ElevenLabs (audio). One thing that should sink in as you watch (or scrub) through is the fact that this still requires creative vision and artistic skill.

A moodboard for AI (Julien Posture - On Looking) - Beyond temperament, I can't claim much in the way of artistry. The images that accompany these newsletters are more often born of a 'hunt-and-peck' creative process than they are generated through a crystalized vision. But I do appreciate the discipline that goes into good art and design, and because I'm often trying to represent AI in the cover images here I found this piece a good nudge to think more critically about those representations.
—Jump straight to the moodboard here.

The Future of Censorship Is AI-Generated (Jacob Mchangama & Jules White - Time) - This free speech argument is ultimately an anti-algorithm argument, calling for human involvement in the determination of what speech should be free and what should not.

Don't think of our AI future as humans vs. machines. Instead, consider these possibilities (Thomas Koulopoulos - FoxNews) - God help me, I agree with an OpEd piece on FoxNews. "The conversation around AI today is all too often framed in terms of replacement rather than augmentation and amplification. This perspective is a relic of industrial-era thinking, which doesn't apply to the nuanced ways AI can complement human capabilities. AI, particularly in forms like Generative AI, is not just about automating tasks but enhancing human creativity and efficiency."

AI is Taking Water from the Desert ($) (Karen Hao - The Atlantic) - Last week I shared an article from The Verge that asked, how much electricity does AI consume? But as we think about the externalities of emerging tech (especially those involving AI and cloud computing) we also need to consider the impact on water usage. This piece from The Atlantic is an eye-opener.

Things to Know About…

The use of Artificial Intelligence (BBC) - The Beeb has updated its guiding principles on the use of gAI (originally released last October), making them formal editorial guidance.
—Related: An update on the BBC’s plans for Generative AI (Gen AI) and how we plan to use AI tools responsibly by the BBC’s Rhodri Talfan Davies

Claude 3 Is The Most Human AI Yet (Evan Armstrong - Napkin Math) - A couple of weeks back I commented that Claude was drifting further and further back into the humdrum middle of the chatbot pack. So it's nice to see Claude back in the game with Gemini and ChatGPT. Claude now comes in three sizes: Haiku (mobile), Sonnet (the free version you’ve been using), and Opus (which should have been “Epic,” IMHO, but kudos all the same for trying a different naming convention). Opus is the paid version, which comes in at the current standard of $20/month.
—And then, from Benj Edwards at Ars Technica: The AI wars heat up with Claude 3, claimed to have “near-human” abilities
—My bottom line: If you have been a Claude fan then this is good news. But if you are a ChatGPT/Copilot or Google Gemini fan, there isn’t any reason to switch at this point. And keep in mind, we still have Le Chat and Groq out there as potentially viable options as well. Chatbots, it seems, are quickly becoming commoditized.

Ideogram is suddenly the AI image generator you need to know about (Joseph Foley - Creative Bloq) - Once Midjourney mastered photorealistic humans (and others followed suit), accurate text generation became the next Holy Grail for generative art tools. As you can see from the images in this week's exploration, Ideogram earns its dark horse status. It isn't perfect, but if you are needing a break from DALL-E (and I was), this is a solid option.
—Jump straight to setting up a login and trying Ideogram here.

Google to relaunch Gemini AI picture generator in a ‘few weeks’ following mounting criticism of inaccurate images (Hayden Field - CNBC) - Just a quick update to the Google Gemini controversy from last week.
—And, oh, BTW, as Megan Morrone reports in Axios, it's not just Google: Meta AI creates ahistorical images, like Google Gemini

Treating a chatbot nicely might boost its performance — here’s why (Kyle Wiggers - TechCrunch) - From the start of using ChatGPT back in 2022, I made a choice to interact with chatbots as I would with people. It just seemed like a good habit to keep in practice. Turns out, that may have positive effect after all.

Adobe previews new genAI tools for custom audio (Andy Stout - RedShark News) - For those Creative Cloud users in the audience, Adobe is finally starting to roll out tools to service more than just its visual editors and creators.
—Here's the announcement from the Adobe Research Team: Adobe previews new cutting-edge generative AI tools for crafting and editing custom audio 

Google DeepMind’s new generative model makes Super Mario–like games from scratch. (Will Douglas Heaven - MIT Technology Review) - While this model, named Genie, isn't ready to crank out high quality platformers like Mario (it currently functions at 1fps, not 30fps or 60fps, which is needed for video games, it definitely feels like a glimpse of the future.
—You can read the Genie announcement from Google at Genie: Generative Interactive Environments

Google Is Paying Publishers to Test an Unreleased Gen AI Platform (Mark Stenberg - Ad Week) - Hopefully some of this resource will flow toward public media at some point. Key quote from Google: “The experimental tool is being responsibly designed to help small, local publishers produce high quality journalism using factual content from public data sources—like a local government’s public information office or health authority. These tools are not intended to, and cannot, replace the essential role journalists have in reporting, creating and fact-checking their articles.”

I analyzed 5M freelancing jobs to see what jobs are being replaced by AI (Henley Wing - Bloomberry) - From the early days of generative AI, forecasts have abounded talking about the impending destruction of white collar (especially creative) jobs. That prediction lacks nuance and historical understanding. While we will certainly look back through the telescopic lens of history and see the destruction of types of jobs by gAI, in real time we'll see the evolution of jobs to encourage and enhance Ai collaboration. People will still need to opt-in to the evolution, but that opportunity will be there for the taking. Wing provides some interesting data analysis to show how gAI will both enhance and disrupt.
—In a similar vein is this piece from Elizabeth Bennett in the BBC: AI could make the four-day workweek inevitable

And finally…

Laurie Anderson on making an AI chatbot of Lou Reed: ‘I’m totally, 100%, sadly addicted’ (Walter Marsh - The Guardian) - And finally, this is not the walk on the wild side (sorry, not sorry) that you might expect, but instead a more personal reflection on the patterns that make us human.

Have a creative, productive week!