This Week Hi, I’m Jim and I helped build VidCon, OG-MCN Revision3, PC Magazine and TechTV. Happy fourth of July for my US readers, and since most of us here are taking the next few days (or all week) off, I’ve got a more compressed set of stories at the top and then a slightly longer piece where I provide a new way to think of Generative AI that might help us use it better – I call it “The Possibility Engine”. Would love your comments either in email directly (jim@louderback.com) or below.
- In the VidCon craziness last week I missed YouTube’s release of its 2023 culture and trend report. Lots of good factoids inside and worth a read.
- One of my favorite creator expert Jon Youshaei just released a series of video nano-tips for LinkedIn creators. Great stuff.
- YouTube misled advertisers by running true-views ads on terrible sites. Agencies are asking for refunds. You would think after 25 years we’d have figured this out. My take here.
- Congrats to Runway for raising $141m to develop AI tools for creators. I’ve got thoughts.
- TikTok wants to sell its own products, launching Trendy Beat digital store in UK, US and beyond.
- A wonky legal look at GenAI, Media and Copyright. Fear not, as Adobe is indemnifying customers who are worried about creating potentially infringing images with Firefly.
- Shein took a bunch of gullible influencers around a Chinese factory. It didn’t end well and now its IPO plans are imperiled. Taylor Lorenz adds context on her new YouTube channel.
- Shopify launches a direct pipeline between brands and creators that promises to make it easier to earn commissions. Mostly beauty and lifestyle products though.
- Unicorn Down: Fake it ‘till you make it is probably not the best business model.
- This might be the worst apology video ever.
- Wondering how the Cannes set creates? Check out the fun yet cringey video from banker and Musical.ly star Terry Kawaja.
- Guess the craze is over – TikTok shuts down its BeReal clone.
- Is any attention good attention on TikTok? McDonalds milks the “Grimace Killed Me” meme.
- Hey LinkedIn – I hope you’re using the output of your AI writing tool to detect AI-generated posts. Otherwise, this is a bad idea.
- Interesting story about how Hotmart is helping LatAm creators make money (ht The Information).
- If you’re not reading Avi Gandhi’s Creator Logic newsletter, you should – great profile this week of The Credit Brothers.
From Precision to Prediction to Possibility – How Generative AI is Changing Computing (and Creators)
For some it’s a new way to easily create content. Others perceive sentience – a virtual girlfriend, evil dictator, or funny sidekick. Still others see impending doom.
But most times I think we get generative AI wrong. It’s a brand-new way for computers to interact with us, using statistics in a new way. It’s what I think of as a “Possibility Engine”. Here’s what I mean:
Setting quantum computing aside, a computer is a precise tool, a “Precision Engine” if you will. At its core, every computer is simply an organized array of switches. Much like a light switch, each one is either on or off. Add 2 to 7 and you will aways get 9 – because each switch (aka bit) is only on or off, only zero or one.
We even had to develop a special piece of computer code to add in randomness, because computers are just so darn precise.
When you pair statistical theory with computers, you get a “Prediction Engine”. The ability to analyze immense amounts of data has led to research that – if done right – lets us predict the possibility of something within a 1 to 2% margin of error.
That does mean that once or twice every 100 times the model is wrong, but we tend to forget that.
So over time our exact computing engines (think calculators) became prediction engines as well.
Even when the models fail (Trump 2016), they don’t REALLY fail. Edge cases can still lead to highly unlikely outcomes – did you ever see someone score four doubles in a row rolling dice? Unlikely but it happens.
But generative AI has turned our computing machines from “Probability Engines” to “Possibility Engines”. No longer are they always right (2+7), or almost always right (Trump Loses). Now they are often right, but often wrong as well.
That Tokyo hotel could be the steal of the century – as ChatGPT recently told me – or it could be a flea-ridden firetrap. Perhaps that legal brief is spot-on, or maybe it’s full of made-up citations.
If you expect truth from Generative AI you will be disappointed, and perhaps $5,000 poorer too.
Instead, think of it as an endless source of possibilities, which you can then forge into your own reality.
For example, I regularly ask ChatGPT to develop 10 click-bait headlines for my newsletters. Instead of just picking one and running with it, I’ll use them for inspiration, taking a word here, a phrase there, and a conceptual framework from another.
I do the same thing with Midjourney. I start with a simple prompt to see what comes up. Then I’ll either refine what pops out, or change and update my initial prompt. Some of my favorite cover images for this newsletter have come from Midjourney surprising me with an unexpected creative result that I then refined and enhanced.
When planning an upcoming vacation to Japan I considered input from ChatGPT, Bing and Bard, but used their recommendations as a filtered opportunity for investigation – not reliable facts.
When I talk to creators, they tend to say the same thing. One uses Dall-E for character design, the other for storyboarding. But it’s not a time-saver, more of an inspiration engineer, a creative kick in the pants. At a #VidCon panel I asked the audience of 800 creators what AI they use today. Most of the 11 creator-tested tools mentioned aid creativity without replacing it.
Sure, in five or fifty years generalized AI might emerge that cures cancer and/or lays waste to the planet. But until then approach generative AI with a healthy dose of both Socrates and Descartes. Question everything, but like Descartes, pull the foundations of your reality out of the miasma. Wonderful and awful possibilities abound – but absolute facts are few and far between.
Related:
- But nobody mentioned Vimeo’s new tools – has anyone used them?
- The Verge on how AI is forcing a metamorphosis of the web – but it might be too soon.
- Casey Newton with a more pessimistic take on the signal and noise of AI.
Thanks for reading and see you around the internet. Send me a note with your feedback, or post in the comments! Feel free to share this with anyone you think might be interested, and if someone forwarded this to you, you can sign up and subscribe on LinkedIn for free here!
If you’re interested in working together as a sponsor to reach the leaders in the creator economy, check out Inside the Creator’s sponsorship packages and/or email me at jim@louderback.com
And don’t forget to listen to The Creator Feed – the weekly podcast Renee Teeley and I produce – get it on Apple Podcasts, Spotify or Stitcher!
