Media is dead...long live AI Media

Author

Matthew J. Sánchez
January 30, 2023

It’s not personal, it’s business. Media is old news, in comes AI Media. AI is channeling Boss Baby this year, gracing us with several corporations across industries embracing the technology to essentially rewrite news, film, music, and more.

What do we have for you today?

AI in the music studio: we’re introduced to Google’s MusicLM making songs and melodies. The future of video: we meet Text-to-4D and TrueSync tech. AI goes corporate (but not in like a douchy way. Right? Right??): BuzzFeed stock skyrockets with its AI announcement — how else are companies thinking of AI?

BuzzFeed quizzes be like…

Generative AI is making its way to the enterprise. We learned recently BuzzFeed is among the cluster of corporations embracing #TeamAI. A couple days ago they announced they’ll be relying on OpenAI’s GPT-3.5 (the technology behind ChatGPT) API to create new editorial output ranging from content, quizzes, personalized writing, and more.

This comes on the heels of the media company’s 12% workforce reduction in December. Following the announcement on Thursday, $BZFD witnessed a 200% stock increase. Could this be the start of a new trend in the “Future of Work”: AI media?

BuzzFeed isn’t the only company moving forward with AI enterprise plans as we all know by now (Microsoft, Google, CNET…although this one not so perfect). Others have shared that the real prize use case for Gen AI in the enterprise world will be “knowledge management.” In other words, AI to disrupt the current way corporate knowledge is stored, accessed, and interacted with.

In my opinion, I think BuzzFeed’s decision is good; now their AI talent can focus on the lower priority stuff…like, yeah, what type of pasta noodle *am I* based on my love language? My guess: campanelle. What about you?

“To me, that’s Cinema” - Bong Joon Ho 🎬

I watched the new Puss In Boots movie over the weekend (highly recommend!). Also a reminder that Runway is hosting the first annual AI Film Festival in NYC in February! But let me not get distracted.

The future of video is coming! Actually it’s already here. This week’s been loaded with video AI advances it’s cool to imagine how this help the entertainment & media industries. First we see text-to-4D happening to create dynamic scenes thanks to MAV3D. Not to mention if you’ve been following along, Evan Halleck, a Director and Editor on the VFX team for the critically acclaimed film Everything Everywhere All at Once, used Runway to add magic to the film, “creating different versions of Michelle Yeoh in different universes.”

Speaking of which. Then there is the debut of 3D performance tracking, dubbed TrueSync, by Flawless AI. Their generative AI is used to alter filmed dialogue, enabling a world where AI reshoots and translations are a breeze.

But not everyone’s happy about the new wave of AI infiltrating video production and editing. For instance, Voice Actor Zane Schact on Twitter calls out NVIDIA’s real-time Eye-Contact AI technology as a disastrous glimpse into how AI programs are “[removing] what little humanity you can still feel in art.”

AI spitting bars and making riffs

Google announced MusicLM over the weekend. The language model is designed and trained to generate music from text. MusicLM performs better than earlier systems and can be conditioned on “both text and a melody in that it can transform whistled and hummed melodies according to the style described in a text caption.”

You can feed it practically anything. Give it audio of a melody, it can whip up new music inspired by that melody. Ask it to make music from a painting, from a story, or some other descriptive prompt and watch it go (here’s a thread). The dataset MusicLM was trained on is public now, called MusicCaps, and features +5.5k high quality music examples already annotated. Simon Willison was nice enough to build out a page and support for those serious about exploring the data or training their own MusicLM.

My future as an undercover, pop/hip-hop artist on the side might be possible now. Someone tell Drake, Tame Impala, and Bad Bunny to warm up the studio. 

Gen AI Deals that make your eyes (and mouth) water 💰

What caught our eye? 👀Supernormal claims to differentiate itself from other transcription tech by “extracting key details” focusing on actions, tasks, and decisions from meetings. Supernormal’s technology aims to be transformative to users who find themselves in meeting-heavy occupations. After all, “who (re)watches an hour-long meeting when the key notes are what are important?” says Co-Founder Colin Treseler.

New developments to spam your #random Slack channel 💬

Things to learn when you need a raise

    Want access to the Generative AI investors database? Share and subscribe to AI In the Middle (below) to receive it! ⬇️

    Before you go

    Who wants to hear some dad jokes? ChatGPT take it away. baddum tsss 🥁

    And that’s a wrap for January! Thanks for the love and sharing our newsletter. Have any tips, companies, technology, or things like this you want to share? I’m a Twitter DM away. 

    — Matthew J. Sánchez (@matthewjsanchez)