AI, Workflows, TikTok Cringe & Pattern Interrupts

Your January Questions Answered

Hello Anon!

Thank you for the great questions via email! Some are answered here and I picked others that will be full posts down the road.

If you pledged already—thank you! If not, please consider it. I have a few topics and guides in mind that are only going to happen behind the paywall.

Questions

How can you use AI to get 95%+ of the way there and then you do the last 5%?

We’re not there yet. Unless you’re talking about very simple things like text on a background.

Adobe and Davinci have some convenient, time saving features right now that speed up the process of cutting things out from a background, filling in things based on the image source, but I have yet to see an AI that will spit out a Photoshop file cut into layers for you, ready to edit.

If I find something like this, I’ll be all over it. For now, it’s enough of a thing to keep up with updates of existing AI like Stable Diffusion. Last night someone put me on to a crazy voice AI. If you were exhausted keeping up with crypto during a bull market, good luck keeping up with AI! Its moving extremely fast with new updates being pushed constantly.

How you use it [AI] to replace components of your existing workflow?

An AI feature that both Resolve and the Adobe products have, as well as web based tools like RunwayML, is the ability to cut out a subject and track it’s motion without having to manually adjust the boundaries of the cut out.

Doing this frame by frame is absolutely horrible and can take literally all day if it’s bad or something is not easily trackable. A tool that just does it without too much fuss? Mana from Heaven.

The other thing that comes to mind, while not a replacement, but a still solving a huge pain point is image and video upscaling. Being able to upscale 1080 footage to 4K is fucking incredible and would have gotten me out of a few serious quagmires back in the day.

Downloaded some free icon but its low resolution and they want you to pay to get the high res version? Use a free AI upscaler. Probably faster than fishing out your credit card and entering payment details.

How/why it can be faster to generate the rendered version and modify it (as opposed to doing it yourself)?

Assuming this is another AI question.

I’m actually not at all talented in terms of visual art skills. I can’t draw or paint to save my life. Without computers, would be pretty screwed. I used a lot of stock footage, download or commission people to draw things for me in photoshop/illustrator that I can then work with, and now AI footage is in the equation.

For moving footage, AI has a pretty specific “look” to it that’s cool, but it’s almost a novelty. The flickering, dreamy, psychedelic look is exciting and different so it commands some attention right now. It also let’s you brag about how you’re an extremely stable and advanced creative genius who uses smart people things like AI in your groundbreaking work.

Two comments here. First, this novelty won’t last forever. Jump on it while it’s fresh and learn to make your computer draw you pictures at all hours of the day.

Second, I’m betting AI will improve to the point where a render looks like actual stock footage. Eventually you’ll use the current AI look for nostalgia purposes. Like how modern vaporwave music is built on a lot of Windows ‘95 era aesthetics.

In the video department, don’t look to AI Animation for speed enhancements. It’s actually slower after you set it up, get all your variables plugged in, then wait for the scripts to run and spit out your frames.

Images on the other hand, totally different ballgame. If you know how to parallax an image or blend it with footage so it looks like it’s moving (remind me to do a post on this) you can move very fast. Really easy to knock out 5-10 variations of something and pick instead of browsing a stock library and hoping you find the right thing.

What are your favorite workflows for jumping between the tools you use?

This might be a little bit of a cop out, but honestly? Folders and good labeling. Helps to iterate projects so if something gets screwed up, you can roll back to a working version.

Adobe Dynamic Link is essential when using tools like Character Animator and Photoshop/Illustrator + After Effects. If I’m working on a project with a lot of PSD assets I’ll tend to keep things in the Adobe ecosystem.

If I’m editing mostly stock footage and not as much in the way of design files, I like to work in Resolve because I have a hardware controller with dedicated function buttons, jog wheels, and other quality of life features. Makes editing and scrubbing through footage much more convenient. Resolve is also nice because unlike the Adobe suite, your animation and post tools are all in one place for the most part, you just have to change tabs.

Neither of these tools are my preference for working with audio. I can do some basic things, but for more intense Audio work, I’ll flip over into a DAW like Reaper or Logic, monitor my video output while I do whatever needs to happen, then export audio files back out which I’ll then drag back into my video software.

How Do I Implement The 12 Animation Principles Efficiently?

I had to look this up if I’m being honest. I have zero formal training in visual anything or any type of technology. As I’ve said before, I went to school for classical music (not recommended for a bunch of reasons) and took exactly one very dated music technology course.

Rather than trying to go through each of these 12 principles I’d just work on everything together by making some basic projects that fold them all in. The bouncing ball tutorial I linked in the the After Effect post covers most of them and the integration of everything together is really where the magic happens anyway.

It will be slow going at first (glacial) but this is more efficient than trying to learn isolated concepts in a vacuum.

General theory of how one makes things engaging without making them cringy?

The best for last. This one made me actually laugh out loud.

“I've watched quite a few tik toks where things were thrown in on top very obviously for engagement (e.g. "chocolate covered broccoli" from educational game design). How does one do it tastefully?”

Good luck with that! Seriously. Yes, it can be an incredible traffic source, but if you do the TikTok thing, just accept that you’re now in the cringe business. Normcore fluoride stare people go there to be entertained. The learning consists of things like how to make jello shots in an ice cube tray and which cigarettes on the ground are still good to smoke. Do not be upset. Best to accept that this is the content they want. Lean into the cringe.

If you want to be arty and make pretty content, stick to Instagram. If you want people with any kind of attention span at all, make YouTube videos.

On TikTok, a quick scroll is an absolutely hysterical (in the panicked sense) schizo mix of content. Thirst traps, “life hacks”, beginner design tutorials, homeless/3rd world people on live begging for money, the no talking engagement thing, bloopers, affiliates, scam ecom dropshipping course shills...you get the idea.

Humans are attracted to motion

As far as principles of engagement on any platform you’re always looking for some kind of pattern interrupt or hook to get them to stop what they’re doing and tap. Then keep them there until some kind of CTA.

The pattern interrupt can be all kinds of stuff. A radical statement, a dangerous situation, a person’s face with an extreme expression (see YouTube thumbnails), dramatically better production quality than the stuff around it, a swimming pool full of M&Ms, whatever.

The story part, where you have to hold attention for a few seconds, needs to have a compelling message. Visually, I find it’s best to keep things moving, literally. You should never be too far away from a moving graphic or transition to another shot. Moving graphics don’t even always need to be nuts, can be something as simple as a text transition.

A common newbie mistake is to just park it on a static shot for too long and expect people to wait it out. This isn’t a David Lynch film. Everyone has ADD-like tendencies on the internet. Instead of picturing people watching your video on a phone, pretend its a literal slot machine. That’s the level of attention you’re dealing with in the short form video world.

People have an inherent instinct to pay attention to motion because this in the past this either signaled danger, social information, or an opportunity to acquire calories.

Ever dangle a string in front of a cat? They crouch, pounce and they can’t help it. No choice. Certain visual cues are hard wired to make them do that.

People are no different. Next time you’re in a room for a few people with a TV on, watch how they act. Upon entering the room and during conversation, they can’t help but glance over at what’s happening because they keep seeing some kind of motion over there in the periphery.

Put on any cable news channel and take note of all the little graphics on screen. If you’re watching someone read headlines at a desk, there’s usually some kind of graphics animation happening at all times. This is deliberate to help keep or pull in ambient attention.

If you’re trying to siphon attention from social media with short form visual content, you want to make sure to keep the string moving with your editing style and make the cat play chase.

Thanks for reading. Don’t forget to subscribe and pledge a subscription amount you’d feel comfortable paying for this newsletter.