Bye Bye AI

NEW: Don’t miss another issue!  Get this newsletter delivered in full to your inbox every Monday – subscribe on Beehiiv!

This Week:  AI Disruption of Creators starts now – and 11 other predictions. Plus, Sora first impressions, why I’m down on MGA and the Rizzler and much more.

Happy holidays and thanks for being one of over 30,000 weekly readers of Inside the Creator Economy.  It’s trend season, and over the next few weeks I’ll be sharing some of mine.  The schedule looks like this:

• This week: AI-related predictions that will shape 2025 and beyond (read to the end), along with the regular mix of top stories and Quibis.  
• Dec23: No issue (unless something really big happens)
• Dec 30: Big news, insights and broader predictions for the creator economy, media and tech in 2025
• Jan 6: Back to normal, but with a special section of creepy signals that could develop into top trends in 2026 and beyond.

I’ll be in Dubai from Jan 2 through Jan 15, helping to produce the 1 Billion Followers Summit.  Look for a meetup in Dubai in the days before and hope to see you there!

SORA RELEASED

Last week’s issue wasn’t more than 6 hours old when OPEN AI fulfilled my prediction and released Sora to the world.  It’s a new model, open to Pro and Plus ChatGPT subscribers.  It took four days before I could actually get it via my pro subscription, as they shut down sign ups due to a huge volume of requests.  Once I did get in, it delivered the best results yet of my standard “A Dog Chasing a Ball on the Moon”.   I tried it twice – mostly because the interface is so awful that I couldn’t find my videos.  Sora’s understanding of physics and gravity leaves a lot to be desired., but the videos were coherent and included actual dogs chasing actual balls.  It includes a storyboard mode to help create a consistent visual narrative, which worked well in my tests too (although the moon has another moon in the background, not the earth).  It’s not going to create the next blockbuster movie – but it does signal a dramatic improvement in AI-generated video. What’s it for?  We’ll likely find out in 2025.

The “System Card” portion of the announcement was particularly insightful, as it shares some additional highlights including:

• The model flags prompts containing names of living artists – but apparently not dead ones  (watch dogs, balls and moons in a Van Gogh style). 
• Video gets transformed into patches, just as language turns into tokens on text-based GenAI LLMs.  Insert clown or dalmatian joke here.
• OpenAI did a lot of work to eliminate “violative content”.  But will it be enough?  Pandora’s box is open, wedging it shut will prove elusive.
• Only some users will be able to upload images of people to use for seeding video. I am not one of them, but I’m not overly concerned, as my test of a beast eating amazon cookies was less than impressive.
• Apparently it’s built on top of some Open Source.

Related: Uh-oh.  MKBHD sees his “plant” in a Sora generated product review video.  

SHOULD WE REALLY CELEBRATE THE RIZZLER?

Just got a press pitch from toy vendor MGA and their ad agency Virtue about how they are “shaking up how we think about toys – and holiday marketing”.  The pitch promotes a highly produced music video featuring 8-year-old TikTok star “The Rizzler” and includes a bevy of other young kid actors as well.   MGA celebrates how they are “moving at the speed of culture” with their Rizzler promo.  But as more and more negative stories emerge about parents exploiting their kids for social video fame and fortune, it looks more they moved without thinking.  Maybe it’s just me, but this looks like a highly exploitive attempt to capitalize on an unhealthy spat of forced child labor – which doesn’t fill me with good cheer and warm wishes for anyone involved.   

YOUTUBE IS NOW OFFICIALLY OLD MEDIA

It’s the classic Silicon Valley pricing model:  A super-low launch price, then stick it to your locked-in customers.  YouTube TV originally launched at $35; now with the latest price hike to $83 a month it’s nearly 2.5x more expensive.  YouTube TV now costs about the same as a cable bundle.  What do you call cord-cutters who cut the cord-cutting option?  Serial Slicers?  We’re about to find out.

• Related: The average consumer spends $273 a month on media, according to a new study – or nearly $200,000 over a lifetime of swiping, searching and watching.

---

SPONSOR:  Whalar NA President @Jo Cronk joined Campaign US Editor Luz Corona recently on Campaign US’s Chemistry Podcast to explore how Creators are reshaping the industry. Hear insights on Whalar’s mission to liberate the creative voice, co-creation’s transformative power, and the future of creativity. Tune in and listen here:  https://www.campaignlive.com/us/campaign-chemistry-podcast 

---

QUIBIS

YOUTUBE

• New Dubbing Tool Launches:  YouTube just made it easier for creators to reach audiences globally with a new auto-dubbing feature.  See my AI / creator economy predictions below for more on this trend.  Only 8 languages supported today – but clearly more to come!
• YT LOVES AI:   9 Other ways YouTube used AI to become more “helpful”.

META

• Trial Balloon: Instagram’s new Trial Reels feature lets pro creators try out a video for 24 hours – without sending it to their followers.  Will be interesting to see how creators end up using it – and how the algorithm scores them.

TIKTOK

• Disaster Planning:  Creators are now preparing for a TikTok shutdown in January.  And beauty brands are too.  TLDR – Instagram will win big. (h/t @brad Berens).
• It’s Up to SCOTUS:  An appeals court declined to temporarily block the ban, meaning it’s up the Supreme Court.  If they accept the case it’ll push everything out until June.  If they decline, then the ban goes into effect Jan 19^th.  Time to bookmark my “TikTok Countdown Clock”.
• What?  Them?  There?  TikTok for Developers channel launches on YouTube.

OTHER CREATOR ECONOMY

• Your Audience Team is Dead:   Turn them all into creators, says an interesting prediction from Neiman Labs .  All of Neiman’s thoughtful 2025 prediction articles are worth perusing.
• Twitch 2024:  Missed this last week, Twitch released a 2024 recap in chat form.  Key stats – over 21M live streamers, nearly 10M new to the platform – and over 20 billion hours watched.  Wow.
• Patreon Recap:  Patreon releases its greatest hits of 2024, with 60M free memberships and more. 
• Ghosted: Snap’s Year in Review.  Beauty, sports, fashion, music, movies and snacks.
• Beast vs. NFL – Who’s Bigger?  By at least one measure, Mr Beast draws 20% more sporting-related viewers than the NFL’s top 2024 game broadcast.  
• Beast Games Woes?  US National Labor Relations Board says some realityTV contestants should be employees.  Might trickle down to Beast Games, other creator reality shows. Beast Games drops Thursday, btw.
• Carlos Watson Guity:  Charlatans, prophets and con artists proliferate during technology and cultural disruptions.  Who will be the Ozy Media of the AI era? 
• Events Are Hard:  Creator Ludwig shuts down the experiential side of his Offbrand agency after a year.

RESEARCH

• Social Platform Usage Among Teens Flat:  Pew’s annual teen survey shows YouTube still on top, while TikTok, Instagram and Snapchat remain widely used.  But with three years of data, it’s clear that social video usage is flat, with only X meaningfully declining and Whatsapp gaining ground.  Market and time-share for the other platforms hardly budged.
• TikTok Shop and Dry January:  No, they aren’t related, except in @Dan Frommer’s latest Consumer Trends 2025 study.  Inside he shares how half of TikTok’s US users are also TikTok shop customers.    Frommer also details how consumers don’t trust TikTok’s privacy policies – and many don’t trust the products for sale either. He also dives into the popularity of weird food trends on TikTok and how it causes GenZ to wait in line to chowdown.  
• Creators Are Everywhere: New report on the creator economy from CreativeClass dives into creators across 20 countries, with insight into India, the US, Brazil, Indonesia and more.  Continues to validate that 1 in 10 of the US population considers themselves creators – but India has 3x the creator population of the US.  Nigeria shows up as a hotspot as well (HT @ankur mehta).

---

AI PREDICTIONS for 2025

BYE BYE GEN AI 

This is the last year I will be making generative AI predictions.  Not because AI will diminish, but like most foundational technologies (electricity, internet, aluminum production) it will fade into the background as it infiltrates just about everything. Sure, we will still highlight odd, transformative, mis-placed and sensational AI developments.  But otherwise, GenAI breakthroughs will become a hidden part of general tech advancements.  Sora, for example, is a breakthrough – but video generation, editing, packaging and production will be the highlight in the future – and not so much the AI underneath.

• Related:  Regular readers know I’m a big believer in Gartner’s Hype Cycle.  I’ve seen so many technologies ride up the slope of inflated expectations and then crash into the trough of despair.  Generative AI will hit the bottom in 2025.  You’ll hear lots of doom and gloom from creators, “experts” and journalists.   But look towards 2026, when Ai starts to climb the slope of enlightenment and real breakthrough uses cases emerge.

GOING LOCAL

Video, text and image AI models will increasingly move from cloud-based to locally run models.  Meta will lead the way.  You can now run powerful models on high-end Macbooks and Windows PCs with 64 Gigabytes of memory, and over the next year even more power will be locally available at home, in the studio and on the go.   Dell and other PC manufacturers will lean into selling fully-configured computers tuned to run AI models with enough memory, Nvidia hardware and storage – and preloaded with text, image and video LLMs. Look for demos and discussions at some of the events I’m producing next year, including Creator Lab at NAB, Open Sauce and beyond. 

• Related:  How to configure and run a GPT4-level model on your Macbook – albeit a fairly beefy one.  

VIDEO LICENSING FOR AI GETS MEANINGFUL

So just how much is your content worth in the age of AI?  In 2025, I predict that many creators – especially of the OG variant – will be approached by companies offering real money to license not just their finished videos but all their raw footage as well to train their video AI models. Those creators with terabytes of b-roll, failed takes and other raw video will be especially in demand.  These licensing agreements will materially affect many creators.  But without careful thought, some creators will get royally screwed.  How?  First by granting a forever license to their videos – which will deliver just a one-time payment.  A better model would be converting a video corpus into “patches” (the video version of tokens) and then licensing an abstraction to Ai companies on a single-use basis.  That may lead to recurring revenue, but if the AI bubble pops then creators could lose out on this licensing opportunity altogether – at least for a few years.  The second risk?  Non brand safe or other unfortunate videos that might be hiding in your library, featuring off-color jokes, illegal activity (like driving 3x over the speed limit in a residential area) or other things you might be ashamed of. Cleaning before selling will be essential. How do you value your back catalog of content for AI?  It’s the wild west out there, so be careful.  But don’t miss the gravy train either.

• Related: New “research license” model from Shutterstock could provide a framework for other media companies and creators too.
• Counterpoint:  Former OpenAI co-founder Ilya Sutskever says that AI models have basically come to the end of the internet, and very little high-quality training content remains.  Perhaps the content licensing market will fade more quickly than I predict above.

AI BACKLASH LEADS TO A LIVE STREAMING and VLOGGING RENAISSANCE

Platforms will be flooded with AI slop.  It will be hard to tell AI generated content from real content, even though the platforms will tune their algorithms to attempt to differentiate between the two.  Viewers will increasingly seek out content from real people as the AI backlash accelerates.  Telltale markers of human generated content, including misssspellings, awkward pauses, mispronunciations and more will signal that a real human wrote, produced, edited that content.  But increasingly we will seek out real people on live streams – warts and all – to satisfy our human need to connect.  Kai Cenat led the way here in 2024, as he became the most subbed on Twitch and has fundamentally changed the platform in the process.  Live ecommerce will also increase but faces saturation headwinds (more on that in my general predictions in the next issue).  We will also see an increase in some of the oldest YouTube formats – vlogging, hauls, let’s play, and other slice-of-life content, as AI slop leads us back to what we first loved from flesh and blood creators.

AI CUSTOMIZED ADS PROLIFERATE

The deep tendrils of AI will increasingly disrupt the advertising and marketing world (see Omniture + IPG).  Although we won’t see marketing messages completely customized for each individual viewer, more and more ads, sponsorships and other promotional messages will be tailored for different demographics, languages and cultures.  From a single ad concept AI will generate hundreds – and then thousands of variants, replacing the characters, messaging, language, tone and esthetics in each iteration.  These tailored ads will then be served up based on individual profiles, location, ethnicity and previously observed preferences.  Amazon currently leads the way here – but we will see tremendous experimentation next year.  In success, this could siphon dollars away from creator-based campaigns. Will it work?  Check back next December.

INFINITE CREATIVITY AT SCALE

2025 will be the beginning of a massive AI disruption of the creator economy – but not the first one.  Technology cycles have changed media ever since Gutenberg invented the printing press around 1440 and broke the church’s monopoly.  The internet disrupted distribution, allowing anyone anywhere to reach everyone everywhere with gatekeepers getting in the way – and led to the demise of traditional media and the rise of the creator class.   AI will disrupt that creator class by welcoming in a billion more creators via translation, dubbing, editing, clipping, audio and other tools.  And because we’ve run out of time – we spend 40% of our day already consuming media  – those new creators will steal viewers away from today’s creators.  But creators themselves will also be disrupted by changes in parasocial relationship-driven experiences as we begin to shift some of our attention to fully AI realized familiars.  

• Related: My friend – and insightful expert – @Phil Ranta predicts “AI Creators (will) fail on major platforms” – and suggests creators should use AI to speed up existing workflows in 2025, ignoring 2^nd and 3^rd order step changes.  My prediction?  Those who ignore AI in 2025 will be disrupted by it in 2026.  AI’s potential to create content, amplify and replicate parasocial relationships and enable formats, experiences and creations that we can’t imagine today will start siphoning off significant attention from fans, viewers, enthusiasts and others.  Ignore the 2^nd and 3^rd functional changes AI will bring at your peril.  

TODAY’S AI TRENDS THAT WILL ACCELERATE IN 2025

We’ve already seen AI change the creator world.  Here are just a few that I expect to accelerate in 2025, and that will continue to disrupt the creator economy as we move from the internet-fueled disruption of distribution to the AI-fueled disruption of creation.

• Translation and Dubbing:  The tools will get better, and the platforms will evolve.  Ignore the rest of the world at your peril.  
• Clipping: Ai-enhanced clipping has birthed a killer new video format.  Expect market leader Opus to continue to shine here, but new companies will allow creators and media companies to infinitely remix their back catalog and video libraries as well.
• Editing:  Descript and other editing companies will continue to make it easier to go from raw footage to finished product.  More Sora-like capabilities will infuse these products as well.
• Packaging and Analytics:  Spotter, VidIQ and other companies will continue to innovate, but the platforms will also build their own tools.  Expect a shake-up and shake-out as the creator market expands.  I predict a push-back on premium-priced services, but more customers overall as more creators emerge.
• Back Office: AI will continue to help with the unwanted tasks of running a business – from legal, to operational, finance and HR.  But it’s still no replacement for seasoned executives.
• Production: The craft of storytelling will continue to be augmented by AI video generation, storyboarding, FX, 3D modelling and more.  Will AI replace art?  That will be one of the big debates of 2025. Will AI replace some of the human production teams.  Yes.

It’s all leading to AI providing infinite labor for content creation and business formation, which will initially let creators focus on storytelling while delegating repetitive and onerous tasks to AI.  But over time this will lead to hyper-personalized content tailored to fan preferences – and in many cases replacing today’s creative storytellers with new types of creators and fully formed automatons.

UP NEXT – The other trends affecting the creator economy in 2025 and beyond – on December 30th.

AI Disclosure: 100% written by me – no human or AI ghostwriters were involved (except for the cover art!).

Like this free newsletter?  Buy me a coffee and say thanks!  Or let’s do a meetup in your town.

I’ve built and sold multiple creator economy startups to top media companies – including Discovery and Paramount. Subscribe here on LinkedIn to get this newsletter every Monday.

Let me know what you think – email me at jim@louderback.com. Thanks for reading and see you around the internet.