← Back

Multimedia by Default

Originally on mirror

I’ve read many generative AI takes that sound something like: “The cost of creation is going to zero” or “Everyone is a creator now”, including my own. The effect of such phenomenon would be a Cambrian explosion of digital content on the internet; from text to 3D assets to music and everything in between.

If this is true, we can make a few observations about this future state.

The internet will be a lot noisier

This is pretty straightforward, there’s just going to be more shit out there. We’ve already seen this play out on social media apps. As Rex Woodbury pointed out, where Youtube creators have high resource requirements such as professional video production equipment, TikTok creators have very low resource requirements; all they need is a smart phone. This lower barrier to entry has undoubtedly contributed to TikTok’s astonishing creator density (1-in-4 users), compared to Youtube’s 1-in-1,000.

Signal becomes increasingly more valuable

Noise begets signal, signal begets noise. Today, curation is a highly monetizable service on the internet through consumer goods, fashion and merchandise, affiliate partnerships, and advertising.

Personally, I’m keeping an eye on curation through ownership and cryptographic signatures.

Ownership unlocks a new axis for curation on the internet: tools like Guild.xyz permission access based on it, apps like Gallery establish social networks around it and loyalty programs like Starbucks Odyssey mark their users by distributing it. There are very few trends I am more bullish on than digital ownership.

Signatures — similar to Fred Wilson’s thoughts, we may end up valuing original content and provenance at a premium (which is why you should collect this Mirror post). Signatures are breadcrumbs of a user’s activity but unlike cookies they carry intent. Instead of guessing what I’m interested in, I can show you. I expect a data rich network of signatures to play a lucrative role in the future of advertising.

Information is multimedia by default

Generative AI’s killer use case may not be content creation but rather content translation. Text can exist as images, images as audio, audio as video, video as XR experiences or any combination of media_type_A as media_type_B.

In essence, content becomes fluid form and users can select their preferred means of consumption; something we’ve already seen with text-to-audio tools.

Tangentially, this could have significant impact on how we search the internet based on preferred media type. One of my favorite forms of content is debates (as video or audio). I’d love a search engine that ingests a topic question and outputs a (simulated) debate between two subject matter experts; reminiscent of Steve Jobs on Joe Rogan.

Five Good Reads

*Disclosure: Delphi Ventures is an investor in Guild and Gallery

← Back to home