Text-to-Video Highlights an Internet Without Rules (or Borders)

Bizarre lack of ethics and rules in the generative A.I. hype media PR

Yes, the giraffe in the TV video is not actually real.

Hey Guys,

Let’s not mistake generative A.I. as a sudden leap to utility, it’s a sudden leap into synthetic spam, if I am to be honest.

For all its huge potential, the field of artificial intelligence has been something of a backwater in the investment world with many companies claiming their startup was somehow an A.I. startup, with around 40% of them not actually being so. The A.I. hype is quickly getting out of control, and it’s simply not true.

So text-to-image tech is not going to break the world, but text-to-video tech could change things over time considerably. This is because from Tiktok to YouTube, we are a society infatuated with internet videos.

It turns out Google’s Imagen is a program that can generate 5.3-second, 1,280-by-768-resolution videos out of a line of text. I’m not sure this is what Facebook was hoping for. So here yet again, and not to be outdone by Meta’s Make-A-Video, Google today detailed its work on Imagen Video, an AI system that can generate video clips given a text prompt (e.g. “a teddy bear washing dishes”).

See my recent post on Phenaki, the tool I consider beyond both of these:

What is Phenaki: A text-to-video model

The text-to-image or text-to-video trend is not at all what automation will be all about. Nor the robots of Tesla, but rather the age of Software 3.0. And by the way, Software 3.0 has nothing to do with prompting programs. Rather it’s the meta trend around RPA, no-code platforms and more realistic ways in which A.I. at scale is infiltrating the tasks in business.

Last week, Meta announced Make-A-Video, an AI system that allows users to turn text prompts into short, high-quality (that’s debatable), one-of-a-kind video clips. Make a Video clips are about as believable as GIFs.

On October 5th, 2022 researchers at the company’s AI lab, Google Brain, debuted(Opens in a new window) Imagen Video(Opens in a new window), a program that can create realistic-looking video clips from a text input. The system expands Google’s original Imagen(Opens in a new window) program by moving beyond still images to moving pictures, resulting in creative videos that remain largely consistent throughout each frame. 

Again I will repeat, this is not useful generative A.I.

Jonathan Ho @hojonathanho

Excited to announce Imagen Video, our new text-conditioned video diffusion model that generates 1280x768 24fps HD videos! #ImagenVideo imagen.research.google/video/ Work w/ @wchan212 @Chitwan_Saharia @jaywhang_ @RuiqiGao @agritsenko @dpkingma @poolio @mo_norouzi @fleet_dj @TimSalimans

Imagen or Meta’s Make a Video is not going to change the world. But overtime synthetic content on the internet may change trust online, even as it bifurcates as users leave Web 2.0 platforms for gamified short video like TikTok and for many off of the Advertising centric platform internet of the past.

If you want to read further, considering supporting the channel as I break down A.I. news and insights and tell the stories around A.I. that matter from around the web.