The past is the present biting into the future.
–Edwin Land
🎬🧠🌱 I've been consciously trying to seed some future nostalgia lately.
I've been watching a lot of older films with my kids. Some are movies I first saw when I was around their age, and I can still remember what it felt like to see them then.
Seeing a great film as a kid isn't quite the same as discovering it as an adult. There’s a limited window to make watching these films together part of their childhood. I’m trying to use that window well!
I used to think of films and TV as just entertainment, something you do when you’re not doing something that matters more. But over the years I’ve come to realize that a really good film or a really good TV show shared with the people you love is one of the peak human experiences. 🏔️
So why not make time for it instead of just letting it happen randomly? Family Movie Night™ is a tradition in this house!
When older adults look back over their lives, a disproportionate number of their memories come from childhood and adolescence. Psychologists call this the reminiscence bump (related to the U-shaped curve of memory I mentioned in Logtime).
This pattern shows up with music, too. Songs we hear as teenagers have a weird ability to stay welded to us for decades. So I’ve been playing my kids music from all kinds of genres and eras. At first, I was doing most of the choosing. Now they’re requesting songs, exploring on their own, and showing me music I didn’t know about!
I love that.
It also makes me think more consciously about films, music, books, trips, little traditions, etc. What are the things that might become keys to whole rooms of memory 30 years from now?
Not that I want to turn childhood into some kind of memory-optimization program 😅
I have no idea what will stick. I’m just trying to get more shots on goal: sharing things I genuinely love, watching them together, talking about them afterward, revisiting favorites, building up shared references and stupid inside jokes.
And then maybe, decades from now, one of those old movies comes on, and for a moment they’re back here with us. 👨👩👦👦
🤖🗓️📝👩❤️👨 Later in this Edition, I write about the new bots/agents that are starting to become popular as digital coworkers and personal assistants. I’ve tried Grok Bot and Instinct, and I’m waiting for Meta Muse to be available in Canada to give it a try.
While testing them, I had this idea: Agents for couples!
I’d love to be in a group chat with my wife and an agent for schedules, calendars, and household coordination stuff. (No agents in normal conversations! Just for this.)
A kind of household chief-of-staff for logistics.
It would be great to have an agent there taking notes, adding things to the calendar, scanning our inboxes for emails from the kids’ school, and reminding us automatically about things we need to know when we need to know them.
👦👧🗓️🎗️ September is Childhood Cancer Awareness Month.
If you want to do something concrete about it, St. Jude is hard to beat. It’s a children’s hospital and research institution with an unusual model: families never receive a bill from St. Jude for treatment, travel, housing, or food, so parents can focus on their child.
But your donation also helps kids who will never set foot in St. Jude. They share their research and treatment advances with doctors and researchers around the world.
When St. Jude opened in 1962, only about 4% of children with acute lymphoblastic leukemia (the most common childhood cancer) survived.
Today, St. Jude patients with acute lymphoblastic leukemia have a 94% survival rate. Their work has also helped push overall childhood cancer survival in the U.S. from about 20% to more than 80%.
Science FTW! 👩🔬🧪⚗️🧫
That’s what I like about the model: you’re helping families at St. Jude directly, while also funding research that can help sick kids anywhere. 💚
💚 🥃 The paid subscription includes the full feed, every podcast, Zoom Q&As with me, and access to the private Discord clubhouse, where a core group of us hang out and discuss the interesting stuff we find:
🏦 💰 Liberty Capital 💳 💴
💰 Nvidia’s Portfolio Isn’t Like Yours (and I don’t just mean the size 😅)
Nvidia now has an equity investment portfolio worth roughly $99 billion, including stakes in companies that buy its products (that’s why you hear so much about the 🔄 nature of it).
Our first reflex may be to judge these the way we judge our stocks: Did Nvidia pay a good price based on whatever valuation metrics? What return can it expect going forward?
But Nvidia has more ways to make an investment pay off than regular, passive investors do.
Helping a customer fund its next data center or lower its cost of financing can generate extra GPU/networking/etc profits, on top of whatever Nvidia earns from the appreciation of the shares. Backing a supplier can help it expand production (leading to more sales, or a bigger allocation than competitors), while backing a software company can help bring more workloads to Nvidia’s ecosystem.
That means an investment with a mediocre standalone return could still be attractive once those extra profits are included. 💵 💸
Here are some of its announced investments and commitments:
OpenAI: a $30 billion investment announced in February.
Anthropic: a commitment to invest up to $10 billion.
SpaceX: a stake reportedly worth roughly $21 billion as of June 30. That’s the value of the holding, rather than the amount Nvidia originally invested.
CoreWeave: another $2 billion invested in January, adding to its existing stake.
Lumentum and Coherent: $2 billion each in announced investments in these optical-component suppliers (Lumentum, Coherent).
Intel: a $5 billion investment announced in September 2025, with plans to jointly develop PC and data-center products.
Synopsys: $2 billion invested in the chip-design and engineering-software company.
Nebius: a $2 billion investment announced in March to help expand its AI cloud.
Marvell: $2 billion invested alongside a partnership on custom chips and optical networking.
Nokia: a $1 billion investment announced in October 2025, alongside a partnership on AI networking.
Corning: an equity arrangement of up to $3.2 billion, including an option to buy shares, alongside its optical-fiber manufacturing partnership.
SB Energy: a $1.5 billion investment commitment in the developer behind OpenAI’s Ohio data-center project.
Of course, Nvidia takes on some of the customer’s risk as part of the deal. There’s no such thing as a free lunch. 🥪
The catch is that those sales have to be additional: If the customer would have raised money elsewhere and bought the same Nvidia chips anyway, Nvidia can’t count those profits as a benefit of its investment.
🔌 400,000 GPUs: A Power Plant Problem ⚡
OpenAI trained Astra on more than 100,000 GPUs at its Stargate site in Texas. Jensen Huang is already pointing to 400,000 GPUs “coming online next”.
We don’t yet know whether all of those will work together on one training run. But what would a frontier lab pay to get that electricity six months sooner?
