Welcome to the AI For Humans newsletter!

There’s an X post that sent shock waves through the AI community this weekend and, while it’s pretty nerdy, it made me think about a lot of stuff.

And, at 12.1 million views and counting, it seems others felt the same way.

Cheng Lou is a software engineer who’s worked at Facebook, Apple and now at Midjourney and spent his career thinking about how code and interfaces collide.

Now, he’s turned his attention towards how text is displayed across the web.

His new release Pretext uses AI models that have been trained on all the weird quirks of how different browsers draw text, covering every font and language, including complex ones like Chinese, Japanese, Korean, and right-to-left scripts like Arabic.

In technical jargon, this gives it pixel-perfect accuracy without needing any CSS at all.

You can render text smoothly and directly on Canvas or WebGL, perfect for dynamic content that generates (and changes) layouts on the fly.

For us normal humans… it means cool new ways to make static text more interesting.

But why has something this nerdy taken off? And why are designers pooh-poohing it?

More importantly, what does this mean for you?

Please support AI For Humans by learning about our sponsors below:

ChatGPT is insanely powerful.

But most people waste 90% of its potential by using it like Google.

These 1,000+ proven ChatGPT prompts fix that and help you work 10X faster.

Sign up for Superhuman AI and get:

Claim your free prompts

The demos that Cheng shared are undeniably cool to see (especially for those not drenched in the history of text on the web) and below I’ll share the incredible things people have already made.

But one of the more interesting aspects of Pretext is the blowback it’s gotten from those who’ve worked in this space for a bit.

Even Cheng himself acknowledges that it’s gotten a little too much attention for what it is and this post made me chuckle.

But the hype and the blowback get to the heart of both the biggest problem and most exciting thing about the current rise of AI coding and really AI at large…

AI is making what was institutional knowledge into general knowledge faster than we’ve ever seen before.

On one hand, if you’re outside of the design engineering world, you’ll see something like this and say ‘Whoa! I can do that now? Maybe that side project I have can look much cooler!!’

However, from inside, you might say: ‘This isn’t as big a deal as it seems because <insert person> did this ten years ago and sure, it’s cool but I mean it hasn’t changed my day-to-day life’.

And, weirdly, us humans can feel one way about AI hitting their specialty vs something they don’t know.

I have a friend who’s a writer who HATES the idea of AI writing anything and believes it will never be able to capture what he does.

On the other hand, he loves AI image and art tools and often uses them to bring for things from his imagination.

Like I discussed last week, my ultimate take is that this becomes a numbers issue.

Sure, there will still be specialized knowledge workers but as tools like this spread to the masses we’ll see things that never would’ve come from the inside.

Love this newsletter? Forward it to one curious friend. They can join in one click.

As I’ve said before, I suspect what makes humans special isn’t the way our brains function but our ability to generate novel ideas.

And, of course, our ability to build off of other humans’ ideas in unique ways.

When a tool like this comes out, the most exciting thing to me is to see that collective ability kick into action across all the different use cases that each individual human comes up with.

You certainly don’t have to be a design engineer to see the incredible creativity on display in the following:

And these examples are from mere days after the tool was released into the world at large… who knows what we’ll see in the coming weeks.

Oh and, yes, the ‘going viral’ part does really matter.

Sure it’s what gets hackles up because it feels like it’s about people trying to hijack attention and post for clicks.

But it also brings much larger audience and almost forces you, the human, to sit up and notice. And maybe actually try it for yourself.

This isn’t the first of these viral tool moments and, if this one doesn’t tickle you or make you think about what’s possible in your world, that’s totally fine.

The important part is that there will be many, many more of these moments and one of them might get you off your butt and make something cool.

When we talk about AI speeding up, it’s not just the models getting better.

It’s also speeding up how we humans can think about our own creativity and get what pops into our brains out into the world.

See y’all next week for a new episode!

-Gavin

Last Friday’s AI For Humans: OpenAI’s new Spud model needs to be good👇

Our Youtube buddy Tim (aka Theoretically Media) dropped a very cool Seedance 2.0 short film featuring his signature AI actor FlameGirl.

The film itself is very good and you can watch it here.

But even better is Tim’s full breakdown video on how he made this and a lot of the learnings along the way. The video is amazing and make sure you follow Tim for more.

Speaking of Seedance 2.0 prompts, Techhalla shared a full thread of them that you should try as well.