Let AI Cure What Ails You

Hello, and happy Sunday! 

This week our team explored the impact of AI on emotions. I wrote about how people are using chatbots for sexual intimacy. Dan found clarity by using the new ChatGPT voice mode to reflect on his own feelings. It’s a new world where our technology is explicitly—pun intended—designed to soothe us. You’ll want to dig into this week's summaries—things are changing fast out there.—Evan Armstrong

Was this newsletter forwarded to you? Sign up to get it in your inbox.

---

Knowledge base

"The Horny Truth About AI Chatbots" by Evan Armstrong/Napkin Math: AI chatbots are the new frontier of digital companionship, with millions of users seeking everything from emotional support to steamy conversations. But before you judge, consider this: 60 percent of Americans report regularly feeling lonely. Read this for a deep dive into the unexpected intersection of AI, loneliness, and human connection.

"Twitter's Future Is a Return to Elon Musk's Past" by Gareth Edwards/The Crazy Ones: You have probably wondered why Elon Musk is so obsessed with turning Twitter into X. Turns out, it's a 25-year-old grudge. In 1999, Musk founded X.com, dreaming of a "global financial nexus." But a brutal boardroom coup (while Musk was on his honeymoon, no less) stripped him of control. Now, he's trying to recreate that dream with Twitter. Read this for a wild ride through tech history that explains the billionaire's current moves and shows how the ghosts of failures past can haunt even the most successful entrepreneurs.

"Here’s a Million-dollar Software Idea" by Evan Armstrong/Napkin Math: Want to make millions? Evan's got a recipe for you: ambient software. It's a new way of thinking about enterprise tech that could disrupt giants like Salesforce. The secret sauce is using large language models that automate everything but the database, making software simpler and more powerful. Read this if you're looking for the next big thing in tech or just want to daydream about your future yacht.

"How Packy McCormick Finds His Next Big Idea" by Dan Shipper/Chain of Thought: Ever wonder how one of tech's most influential voices uses AI to write and invest? In his latest episode of AI & I, Dan sat down with Packy McCormick of Not Boring to find out. From using Claude to edit his newsletter to creating interactive graphics for complex concepts, Packy's AI toolkit is surprisingly simple yet effective. Read this if you want to learn how to leverage AI for better writing, deeper research, and smarter investing. 🎧 Listen to the full conversation on Spotify or Apple Podcasts, or read the full transcript.

"ChatGPT's New Voice Is Eerily Human" by Dan Shipper/Chain of Thought: Imagine having a conversation with an AI that feels as natural as chatting with a friend. That's the reality with ChatGPT's new Advanced Voice Mode. Dan tested it out and found it's not just incremental progress—it's a leap into the future that makes Siri and Alexa look like relics. Read this to learn how AI voice tech is evolving from clunky robot to smooth conversationalist, and how it might just help you become a better version of yourself.

---

Fine tuning

The government wins against Google

The biggest news of the week came out of the United States court system. For many years, Google has paid tens of billions to Apple and Samsung for making its search engine the default on their mobile phones. A judge ruled that Google abused its monopoly power and broke federal antitrust laws.

Why did they win? For most of U.S. history, there has been a cultural aversion to anything representing dominance by one company, especially in instances where it increases costs for consumers. Google has been able to skate by most monopoly accusations because consumers overwhelmingly prefer its free offering over competitors from DuckDuckGo, and happily for Apple’s shareholders, it was paying Apple billions of dollars for the privilege of offering its search engine. Google’s argument was essentially that how could something that cost billions be a monopoly? The government successfully argued that the Sherman Act forbids the extension of monopoly power through the use of contracts. So the act of spending that money was an abuse of monopoly power, even though it didn’t raise prices on consumers. Purchasing market share for Google means the company can use a balance sheet, not innovation, to protect its market position.

Don’t expect things to change all that quickly. As expected, Google is going to appeal this ruling, and there may be years of courtroom battles ahead. The two parties previously agreed to delay the remedies stage, where they’ll joust over what changes are needed to get Google in compliance. All of the options have downstream consequences. If Google is forbidden from making revenue-sharing agreements with browser providers, Mozilla’s Firefox will likely lose 86 percent of its revenue, effectively killing that organization. If the government mandates that Apple inserts a pop-up allowing customers to choose a search engine, Apple will lose out on the $20 billion a year in revenue when customers pick Google anyway. Apple, through sheer process of elimination, may find that the most long-term profitable choice is to create its own search engine, taking billions of dollars in R&D spend. There are no obvious answers, and no easy paths. Even if this is the right and a net good decision for incentivizing innovation, much of the internet rests on the basis of these types of agreements. Expect a shakeout.

OpenAI’s can solve the student cheating problem (sort of). Speaking of complicated incentives, OpenAI has been sitting on a watermarking tool that identifies ChatGPT work with 99 percent accuracy. The tool works by slightly altering the pattern of token generation so that the output is identifiable. I’ve talked with over a dozen university instructors in the past year, and they all agree that ChatGPT has been a plague in the classroom. Many students use these tools to cheat and shortcut thinking. As such, you would think it is obvious that OpenAI should release its watermark. However, OpenAI conducted a survey that found that releasing the tool would sour more than 33 percent of its users, who would switch to another provider who isn’t going to narc on them. While OpenAI, the nonprofit, would release it, OpenAI, the startup valued at over $70 billion, would do anything possible to stop it. Even if the company does release it, the broader its adoption, the more likely students are to find ways to cheat the system. Again, no easy answers, but doubly challenging when you’re trying to be a nonprofit and a capitalist enterprise simultaneously.

Riches in niches. In happy media news, anime streaming service Crunchyroll hit 15 million subscribers while rolling out exclusive merch and events. In a world where doom and gloom has been the norm in media, Crunchyroll is a bright spot because it stayed focused for nearly 20 years on dominating anime. Impressive company.—EA

---

Data mining 

The CEO as a prophet of doom. Business success is often a question of timing more than destination. For example, it’s been obvious for a decades that streaming was the future of television, but how to get there, where to consolidate, and how to sell has been less obvious. Now, there has officially been a bifurcation between “CEOs who were right” and “CEOs who were too slow.” This week both Paramount and Warner Brothers Discovery wrote down a combined $15 billion on their linear TV networks, tacitly admitting they mistimed the decay rate of traditional pay television.

Source: Financial Times visualization.

Both of these companies have ended up burdened by their own TV businesses. By contrast, an emphasis on physical experiences (Disney), niche audience focus (Crunchyroll), or internet dominance (Netflix) have fared significantly better. This is despite Warner Bros. and Paramount having large streaming platforms and deep libraries. Each of these companies has relatively similar ingredients, but failed on questions of timing and execution. There’s a lesson in that.—EA

---

Alignment 

What to read when you only have 15 minutes. I'm a sucker for bite-sized brilliance—those nuggets of insight that hit you like a ton of bricks in the time it takes to scroll through your Instagram feed. If you're looking for a commuting companion that'll jolt you awake faster than your soy milk latte, explore the world of Japanese death poems. I know, "death poems" sound about as cheery as, well, death, but hear me out. These are the mic-drop moments of Zen monks, their final thoughts before swirling toward the abyss. It's like eavesdropping on history's most insightful deathbed confessions, the kind that make you grateful for your morning bagel and have you considered the meaning of life? before your first meeting. My current obsession? This gem: "For over 60 years, I often cried katsu, to no avail. And now, while dying, once more to cry katsu, won't change a thing." Boom. Subway wisdom. No matter how much we shout at the universe, it shrugs back.—Ashwin Sharma

---

Hallucination

What if the Swiss Army came out with a device to fix your computer problems, too?

That’s all for this week! Be sure to follow Every on X at @every and on LinkedIn.