2
0 Comments

ūü§Ė Trends #0057 ‚ÄĒ AI-Generated Content

Why It Matters

AI-generated content is being used for good and evil.

To create conflict, incite violence and deceive.

To entertain, design and increase productivity.

AI is amoral. It amplifies intent.

Problem

Humans are predictably irrational. We are built to replicate. Not to reason.

Solution

AI-generated content takes us from the era of limited to infinite leverage.
Making connections that we miss, learning faster and improving over time.

Players

Images

Voice

Video

  • Rosebud AI¬†- Create virtual characters
  • Alethea¬†- AI-generated avatars
  • Synthesia¬†- Create videos without actors
  • Reface¬†- Create face swap videos

Text

  • GPT-3¬†- An advanced language model
  • IdeasAI¬†- AI-generated startup ideas
  • Shortly¬†- Complete your thoughts
  • Frase.io¬†- Create content briefs and answer questions

Predictions

  • Production¬†costs¬†will drop. Watch David Beckham¬†"speak" nine languages.
  • Actors and artists will perform¬†posthumously. New songs and films can be made with training data. "Immortality" boosts the value of estates as¬†alternative assets.
  • Tools like¬†Synthesia¬†will lead to more¬†million-dollar, one-person businesses. The need to master (or understand) underlying technologies is reduced. Leading to¬†smaller firms.
  • Pirated¬†personalities¬†will be used to¬†endorse¬†products, narrate books and more.

Opportunities

Key Lessons

  • Media is being¬†democratized. If¬†no-code¬†makes you a developer. AI-generated content makes you a producer.
  • Code,¬†low-code¬†and¬†no-code¬†give you limited leverage. They require direction. AI-generated content gives you¬†unlimited leverage.
  • Humans are predictably irrational. For good reason. Evolution doesn't reward¬†rationality. It rewards¬†replication. Irrationality isn't surprising. It's expected.

Haters

"Photoshop has been around for years. So what?"
Doctored images fool us. Higher-fidelity fakes are more convincing.

"People's likenesses will be abused."
That's the path we're on. Create to combat it. What's the blue check version of video? Will we "NFTfy" everything to prove authenticity?

"What about biases in AI?"
Twitter took Tay.ai from chill to racist in one day. AI is a sponge for society's ills.

"Deep fake detection algorithms will protect us."
Deep fakes will improve with detection efforts. GANs model an arms race. Sensity, a deep fake detection platform, is a test layer. Pass and it's ready to release.

Related Reports

Links

  1. A Student Used GPT-3 to Write Fake Blog Posts - An early story showing the power of GPT-3.
  2. A Russian AI 'Graphic Designer' Fooled the World... for a Year - How an AI "designer" charged $1,000+ for logos.
  3. We Recreated Joe Rogan's Voice Using Artificial Intelligence - A look at what's possible with voice cloning. Popular personalities will be pirated.
Trending on Indie Hackers
From 1 to 767 paying users in 3 weeks with no marketing effort 10 comments Launched Tweetflick! 9 comments Status Dumps to be more productive, happier, and avoid burnout as an engineer 6 comments No budget, how you do marketing without spamming? 6 comments Just discovered Firebase Auth for MERN stack and I'm never going back. 3 comments A social media listener 1 comment