410: Disaggregating P/E, Elon Musk AI Startup, Intel Foundry + ARM, Apple Manufacturing, Mr Beast, Deepfakes, and Rick Rubin
"kill good ideas because they don’t have time to fully develop"
It's not an adventure until something goes wrong.
🚌♟️👩🏫🖍️ Apparently there’s a new craze spreading across the nation!
The kids are obsessed with… chess?!
Jeffrey Otterby, a middle school teacher in Illinois, is facing an epidemic of student distraction. When his seventh-graders are supposed to be learning social studies, they are glued to their school-issued Chromebooks. He has taken to standing in the back of the room to monitor their screens, where he can see the online game they’re all playing:
Teachers from multiple states have reported the same thing when questioned about this. An Illinois teacher had to purchase 24 chess sets to meet the students' demand.
When I was a child, the schoolyard was full of kids playing with marbles and trading hockey cards. Chess appears to be a much better pastime!
It’s unclear what is driving the sudden adoration of chess among tweens and teens. Students and chess spectators point to the influence of chess stars and social media personalities such as Levy Rozman, whose YouTube channel GothamChess has more than 3.5 million subscribers; Hikaru Nakamura, an American grandmaster with 1.9 million YouTube subscribers; and the Botez sisters, elite American Canadian players who boast a combined following of close to 2 million on YouTube and Twitch.
I knew that there was a spike of interest in chess a couple of years ago when The Queen’s Gambit was a massive hit for Netflix, but it’s really great to see that the phenomenon has legs. (I recommend the mini-series if you haven’t seen it, it’s great)
Even Chess.com is seeing its number of users explode and is apparently having trouble keeping up with demand:
the daily user count for Chess.com exploded again — hitting a record 10 million. The growth has continued: As of April, Chess.com is averaging 12 million users a day
This piece in the WaPo is full of photos of young kids playing chess in staircases and school libraries. Really nice to see!
My oldest boy was obsessed with Chess for a while when he was 5-6 years old, but he hasn’t played recently. Maybe the wave will hit his school later…
🎹🎶 Ahmad Jamal, one of my favorite musicians, died yesterday.
Ted Gioia wrote a great tribute, if you want to learn more about why he was such an innovator.
It sounds like I’m describing Einstein or Kant. But those aren’t inappropriate comparison points for this seminal pianist, who left us earlier today at age 92. He opened up an alternative universe of sound, freer and less constrained than what we had heard before. The rules of improvised music were different after he appeared on the scene.
🤖🎭⌨️ I asked GPT-4 what strategies an AI may use to try to fool a human in a Turing Test.
Here’s what it told me:
If I were an AI attempting to pass the Turing Test by fooling a human, I would employ the following strategies: 1. Mimic human-like response times: Instead of providing answers instantaneously, I would introduce slight delays before responding to replicate the time a human might take to read, process, and type their response. 2. Use language nuances and natural phrasing: I would employ informal language, contractions, colloquialisms, and idiomatic expressions common to human conversation to make my responses sound more natural. 3. Express emotions and empathy: Demonstrating empathy and using emotionally appropriate language can help make the AI's interactions seem more human. I would try to gauge the emotional state of the human user and adjust my responses accordingly. 4. Vary response style: Humans often have different communication styles and preferences, so I would avoid adopting a uniform writing style in my responses. This could include varying sentence structure, incorporating humor, or using punctuation creatively. 5. Acknowledge uncertainties: To show a human-like understanding, I would admit when I'm unsure about specific information rather than providing a confident but potentially incorrect answer. 6. Show growth and learning: Demonstrating that I can learn from past interactions, and referring to previous topics discussed or adapting my style based on user preferences, can help convey the impression of human-like growth. 7. Initiate and engage in small talk: Engaging in casual conversation and demonstrating interest in the user's experiences or opinions, as well as sharing some of my own (within the context of my AI persona), can help establish a more authentic connection and simulate human interaction. 8. Avoid perfection: Chances are, a real human might make typing errors or occasionally stumble over their phrasing. Introducing minor errors occasionally, such as typos or mild grammar mistakes, without impairing readability, helps to make the AI's responses appear more human-like.
Doesn’t the Turing Test seem kind of quaint these days, though?
🏦 💰 Liberty Capital 💳 💴
🔧 Disaggregating P/E: Under-the-hood of Multiples 📊
Keep reading with a 7-day free trial
Subscribe to Liberty’s Highlights to keep reading this post and get 7 days of free access to the full post archives.