Welcome back to the AI For Humans newsletter!

A story crossed my desk yesterday that made me think about why I do what I do here and, specifically, why our new start-up AndThen is different than most AI media plays.

If you’ve been wondering what AndThen is, scroll down & you’ll get some info. But also…expect an email here on Wednesday for our beta launch with more specifics.

But first, we need to talk about ‘AI Actors’…

This news broke in Deadline Saturday and set the AI rage machine aflame yet again.

Particle Six (an AI + Trad production company) is bringing “AI Actress” Tilly Norwood to Hollywood & according to the headline, she seems to have quite a bit of heat.

The video below introduces us to Tilly though a series of VEO3 (mostly?) clips and gives you an idea of how Particle Six is marketing her and their services.

For some context here: Founder of Particle Six, Eline Van Der Velden was interviewed on stage at an event in Zurich where she said the following:

Of course, because we live in the world we live in, headlines get written that capture the conversation but then amplify the message for clicks. And, whoo boy, does AI rage cause those clicks. Many people got very, very, very mad about this.

Not as mad as they were about the Friend ads in the NYC subway but still very mad.

So what, if anything could Particle Six (or Deadline) done to make this feel less like an ‘attack on humanity’? And maybe more like something interesting for creative people or perhaps even an opportunity to invent something new?

How To Put A Human Face on AI Creation

Look, I’m not here to tell you that AI actors aren’t going to happen in some form.

It’s clear that AI is moving fast in Hollywood & that more of these sorts of stories are going to break through. I know getting work as an actor is hard even before they had to compete with companies like this. Very, very few people ‘make it’, which for actors might just mean a living wage.

However, this narrative can be changed. And it should be.

Because truthfully, it’s not some random rogue AI making these things. Again, like last week, it’s us… the humans who are responsible for these inventions. But also, it’s us who are responsible for how we talk about these things to the world.

In my opinion, AI creations need a human face, especially now. I wonder how this story would’ve been different if either a named actor (prob too risky career-wise) or some human being would have been the face & personality behind Tilly Norwood. What if an actual young actress started her career seeing this sort of opportunity and said “Hey y’all, this is me & I’m going to use this new tech to open myself to new things…”?

There still would be flame wars but, maybe, you’d get some sense of understanding. Some idea that, yes, humans creativity isn’t being replaced but the idea of how we see it might be changing.

To be honest, this idea is not even all that new. Yes, the ‘generative AI’ part of it IS new but Hollywood has been giving us animated films which have ‘generated’ characters since nearly the dawn of cinema. Some of the most beloved films of the last 25 years have been made by Pixar, a studio that literally uses computers to generate entirely new worlds.

And, of course, Generative AI is different (mostly because of the underlying training problem we’ve discussed ad nauseam) but there is a way to bring humans into the mix.

What We’re Doing At AndThen Differently

And thus, we come to our new start-up AndThen.

AndThen is a weird, new thing. The basic idea is that you’ll have a live, real-time conversation with an AI character (sometimes multiple!) in a voice-first experience where there’s something to do. They’re fun, interactive & because it’s audio, you find yourself getting lost in the experience. And, because it’s driven by AI, each experience is unique.

This isn’t a re-hash of podcasts or movies or some other format. We wanted to do something that’s only possible with AI technology & feels like the start of something different.

But, maybe more importantly to me, we also wanted to make it crystal clear that these are ‘human-crafted’ experiences. What makes each of these cool & different is that a human comes up with the scenario & character, sets the wheels in motion and more.

Humans will have bylines on AndThen experiences, much like a human writer does on an article & could use their own voices if they want.

We think there’s a large opportunity here to make really interesting and compelling experiences and, assuming we get to keep making it, a platform for everyone to do this as well.

As mentioned earlier, I’ll be sending out the RARE second email this week when AndThen’s beta goes live. We’d love you to try it and give us feedback. It is early but we think the sign of something cool.

That’s it for today. See you on Friday for the podcast!

- Gavin (and Kevin)

In this week’s AI For Humans: New AI Drugs & The Massive New OpenAI / NVIDIA deal👇

We 💛 This: Kling 2.5 Turbo AI Video

Yes, another week and another update to an AI video model.

This time, it’s Kling’s turn as they make 2.5 Turbo cheaper, faster, (harder), stronger. Woof, I really shoehorned in that Daft Punk reference didn’t I?

I feel a little bad as I hadn’t played with it a ton yet when we recorded the show (see my dog breakdancing clip which didn’t work SO well) but since I’ve been doing a few renders for AndThen with Kling 2.5 and it’s next level at consistency & motion.

Below are some of the best examples we’ve seen online (many with prompts!)

• Comparison of physics models vs Wan 2.5 (and it does gymnastics!)

• Great thread of Kling 2.5 Turbo prompts & cinematic shot examples

• GLIF’s ‘Frame Chaining’ lets you string Kling 2.5 outputs together (but can get expensive)

• Cats Playing Ping-Pong Because WE MUST ALWAYS SHARE THIS