Cloning Oneself with AI
Finally, an AI voice clone worth listening to
The best way to understand where the world is going is to experiment in it today. Over the past few months, I’ve been playing with AI to understand its potential and its flaws. Today, I share my most recent experiment in creating an AI voice clone with Eleven Labs. I recently interviewed Victoria Weller, Chief of Staff at Eleven Labs, about the company’s technology. The company’s product didn’t disappoint.
The Deload Podcast interviews CEOs, founders, and builders of frontier technology companies that are transforming how we live. The mission of the podcast expands that of The Deload: To educate you about where the world is going to make smarter growth investments. Now with conversation.
If you haven’t subscribed, join more than 1,500 rational optimists:
One for the Listeners
I never really liked reading. Maybe you can relate. For years, I felt bad about my dislike for reading until I begrudgingly read “Managing Oneself” by Peter Drucker where he poses a critical question:
“Am I a reader or a listener?”
Drucker’s point is that people are predisposed to process information better via either spoken word or written word. You need to know what kind of learner you are to manage yourself well so you can orient your information processing to your strengths.
The concept of being a listener was groundbreaking for me but should have been obvious in hindsight. I’ve always been a better listener than reader. In school, I could listen to a lecture and never do homework or open a textbook and get good grades.
With the freedom of identifying as a listener, I consumed audio content with aggression — an audiobook a week and several hours of podcasts. I didn’t worry about reading. It was great.
Alas there’s more written content than there is audio, and every week there would be a handful of articles I wanted to read but never found the time. I’d have a dozen tabs open on my browser and another dozen from the week before. The curse of being an audio lover. Maybe you can relate to that too.
I tried to solve my listener’s problem.
A decade ago, shortly after I read “Managing Oneself,” I taught myself to code and built an app called Dogear. The idea was to make a tool that would automatically edit an audio file of an author reading his story. I knew I wanted it as a reader, but turns out authors weren’t that interested in spending the time to make mini-audiobooks of their work.
So, I pivoted. My wife and I started a podcast called Dogear where we would read our favorite stories from the week. A bunch of our friends thought it was cool, but it wasn’t scalable. Too much voice work, too much editing. If only we could just have a great AI clone of our voices to do the work for us.
Ten years later, wish granted.
I Endorse this Message
If you’re listening to me read this post, I’m not really reading it. My AI voice clone from Eleven Labs is reading it. If you’re reading this post because you’re a reader, not a listener, then imagine a voice that sounds 96%+ like me as you follow along.
When you hear an AI clone of your voice for the first time, your first reaction is amazement, then it turns to scrutiny. You try to find areas where the voice doesn’t sound quite right. Then you start imagining how you might use your voice clone.
The first thing that struck me is how different the use case for an AI voice clone is vs a chatbot like ChatGPT.
With a chatbot, you can submit any idea to the bot, see what it comes up with, then iterate from there. It’s a collaborative experience that lets humans explore and expand idea kernels into something more robust. I’ve used to do stock research, design luxury handbags, and craft intelligent stock indices. And my Intelligent Indices are performing quite well against the market so far.
With a voice clone, you need to bring creativity to it. You need to give it something to say — a finished script. But whereas a chatbot helps explore creativity, a voice clone attaches creativity directly to you.
Voice is an inseparable part of our identity. Our unique sound literally makes us identifiable to others. Using your voice clone to bring a creative text idea to life is an intentional endorsement of the idea. I could have chosen any number of other lifelike voices from Eleven Labs to read this post, but these are my ideas, and I want to be associated with them.
However, it’s not my voice or any other that gives an idea value. It’s the ideas themselves.
Voice Enhances Value
I love Michael Caine’s reading of “If” by Rudyard Kipling.
Caine didn’t write the poem, it wasn’t his idea, but his familiar British tone adds a flair to the poem that enhances my listening experience. I love his rendition so much that I don’t want to hear anyone else do it.
The important thing for all uses of AI is that it must add value somehow, and voice clones are a tool that can enhance the value of already valuable ideas.
Someone named “wono” on Twitter offered a good thread that sums up AI vs humans and value creation:
Value must be the first consideration in any act of creation. If Michael Caine read the first 2,000 numbers of pi, he wouldn’t make the content more valuable. Value comes from ideas, but ideas can be enhanced with a human touch. That’s where I disagree with wono.
People do care about connection. That’s why influencers are influencers. Yes, they provide value. But the value is enhanced by the connection that fans form over time by consuming the valuable content. Value and connection create a virtuous cycle in the creator economy. Humans root for other humans that they like to succeed. Have you wished for ChatGPT to succeed yet?
The more I’ve explored using AI, like my voice clone here, the more I believe that we’re on the verge of a wave of human creativity never before experienced. AI helps us explore exponentially more ideas, and it will help us create exponentially more content. The net result of human and AI collaboration will raise the bar for excellence and create more value for society as a whole. And now, with AI voice clones, it will also unlock a world of efficient learning for listeners like me and you.
Maybe it’s time to brush off the old Dogear idea as my next AI experiment.
Disclaimer: My views here do not constitute investment advice. They are for educational purposes only. My firm, Deepwater Asset Management, may hold positions in securities I write about. See our full disclaimer.
Intelligent Indices Weekly Update
AI has done better than merely clone legacy stock indices.
Intelligent Indices are a suite of stock indices powered by ChatGPT, Google Bard, and Claude. The three leading AI tools act as an index committee that selects and weights stocks to form Intelligent Indices.
Even after a tough few weeks for markets in general, the Intelligent Indices are still beating their legacy benchmarks. The flagship Intelligent Select, a US large cap equity index, is outperforming the S&P 500 by 10 bps since inception earlier this summer. The Intelligent Tech select is outperforming the Nasdaq 100 by 100 bps. And still most impressive, the Intelligent Select Equal is beating the S&P 500 Equal Weight index (RSP ETF) by 190 bps.
The Intelligent Indices have largely positioned themselves for a pullback in big tech which has led the market rally YTD. If we do experience some reversion to the mean, the Intelligent Indices will perform quite well.
Logically, it seems the market has two paths: either the current rally broadens into other sectors if we get more signs of soft landing, or we get a recession, and the market goes down with higher multiple tech going down more. Of course we could just have tech continue to rally, or the market could go down and tech not go down as much, those just seem less probable than the first two.
I’ll repeat what I keep saying about AI-powered stock indices:
AI is already capable of creating usable stock indices with intelligently engineered prompts, and it’s only going to improve over time.
As AI influences investment processes more, we need intelligent benchmarks as barometers for AI-enhanced investment strategies.
Like many other industries with entrenched dominant players, AI will create the potential for something new and dynamic to disrupt legacy indices like the S&P 500.
It’s hard to imagine a future where there isn’t an AI-powered index scrolling on CNBC all day next to the S&P.
The future of indexing is intelligent.