Dear Friends,
It’s starting to feel real. The autumnal equinox is behind us. The days get shorter. Q4 starts next week.
And what a fun quarter it promises to be. San Francisco will be overwhelmed with good music this weekend. While my body will be dancing to the beats at Portola Music Festival1, my heart will be with Valerie June, Beth Orton, M. Ward, Steve Earle, and Rickie Lee Jones at Hardly Strictly Bluegrass. This weekend also brings a long-awaited reunion with some of my oldest, closest friends. And all under 70-degree, sunny, blue skies.
After its infamously cold and foggy August, San Francisco sure knows how to shine in September and October. 🌞
Social comparison … with my AI clone
Iris and I flew back from Stockholm on Sunday2, which gave me some time to catch up on reading and podcasts. Here are some of the stories that caught my eyes and ears:
Over at the Verge, David Pierce is doing some great reporting about the reality of today’s AI (not mere speculation about its future). I enjoyed this 25-minute episode where Pierce creates various AI clones of his voice and explains how Amazon and Apple are training new AI voice tools on their catalog of audiobooks. Makes me think that voice actors are facing the same future as scribes in the 15th century. Truly worth a listen.
AI won’t just replace voice actors, but also interpreters. Spotify has partnered with OpenAI to clone podcasters’ voices and translate them into other languages. They’ve already released an AI-translated version of Lex Fridman’s conversation with Yuval Noah Harari, and the quality is incredible. I can’t wait for this to be made available to everyone. It will be a trip to hear my AI clone speaking Spanish compared to my actual voice. What if I prefer how my AI clone speaks? And how far away are we until I can travel to China and have my AI clone speak perfect Mandarin in real-time through my AirPods or the speaker on my phone?
Let’s assume that I live in Mexico for the next decade, and that over that time I slowly stop speaking Spanish because I come to prefer the Spanish of my AI clone? Depressing, I know.3 How and whether we speak a foreign language is just one of many questions about how we might change our behaviors based on AI simulations. In a similar vein, I appreciated
’s interview with about her new book, Flawless: a look at how Korea’s futuristic beauty industry is coming for us all.4 Beauty filters on TikTok and Instagram show us idealized versions compared to our reflection in the mirror, leaving us insecure and tempted to alter our appearance. If more of our friends get Botox, face fillers, and plastic surgery, then we will feel pressured to keep up. Are we destined to feel less and less comfortable with how we look?
Back in June, I wrote a longish piece about “How Apple's Vision Pro & AI Will Forever Change Friendship,” but I didn’t expect things to move along this quickly. Amazon had an event last week pitching the newest version of Alexa as a chatty friend. I always assumed that Alexa was for setting timers, but Amazon says its logs show that most users try to have back-and-forth conversations. Creepy. We also learned that Amazon will invest up to $4 billion in OpenAI’s rival Anthropic to accelerate its AI development.
The next day, OpenAI announced that ChatGPT can now “hear, see, and speak” using the same underlying technology that powers Spotify’s podcast translation voice cloning.
As if that wasn’t enough, Microsoft is rolling out its AI Copilot this week to take over your repetitive tasks, and Google released Duet AI to do the same for Google Docs and Gmail.
But the big news came from today’s Meta Connect conference, where the parent company of Facebook, WhatsApp, and Instagram announced that they’ll be putting 28 AI characters in all of their apps, including characters based on Dwyane Wade, Kendall Jenner, MrBeast, Snoop Dogg, and Paris Hilton. The Wall Street Journal reported that celebrity-based AI characters are the basis of Meta’s plan to take back young users away from TikTok.5 Creepy, creepy, creepy. The craziest news of all is that while Meta starts out with these 28 AI characters, next year they are launching AI Studio to let anyone create their own.6
This is all moving faster than I expected, or can keep up with.
I’m excited about a future where I can “speak” and understand any language in the world. I’d love to take a photo of a cabin and have ChatGPT give me step-by-step instructions about how to build it. And I’d love for an AI copilot to show me how to waste less time on Excel.
But I hate that our AI-optimized future will make us feel less secure about ourselves and more distant from our friends.
Will we figure out a way to take advantage of the useful tools without becoming robots ourselves?
Sim Simma
Here’s a track by 20-year-old Little Simz, who will be playing at this weekend Portola’s Music Festival. I think her rhymes are hilarious.
Stay human, my friends. And have a great week!
David
PS: “I’m giving thanks for being human every morning.” From Oakland’s own, Michael Franti:
Especially excited for Little Dragon, Bonobo, Jon Hopkins and Little Simz
In next week’s newsletter, musings on “Peak Sweden” and the comforts of degrowth
But then again, I swore I’d never wear AirPods in public, and never thought that I’d stop paying attention to how to get around town because I now mindlessly follow Google Maps.
Or maybe the book is about the trap of optimization — when you can measure yourself against an ideal (say, the distance between your cheek bones and eyebrows) and then optimize to get as close as possible.
TikTok lured away young people from Instagram friends by giving them funny, photogenic influencers. And now Facebook hopes to steal them back by giving them AI friends, including a sassy AI persona called “Bob the robot” who will “use its ‘sharp wit and biting sarcasm’ to try to convince the kids to stick around on Meta’s increasingly aged social media platforms.”
If you want to be creeped out further (especially if you’re a parent), check out Alex Heath’s conversation with Mark Zuckerberg — who sounds incredibly relaxed and confident these days.
OMG that 2009 G5 video! It's the pre instagram reel-recap, but 7 minutes long on You Tube ;)
Alex has changed a wee bit, but his dance moves haven't!
I can't even start to think about my kids talking to AI friends online. I'm still overthinking the idea of a "basic" GPS watch for the oldest... and I'm dreading navigating the teen social media stuff.