309: Chip Foundries, Minecraft on NFTs, Walmart's Innovation, Netflix vs Disney+ vs HBO, the Economics of Blade Runner, Ports, DALL-E, and Apple Watch
"I need to set myself a few more pointless goals"
To be wealthy, accumulate all those things that money can’t buy.
—Kevin Kelly
🛀 📺 🎬 🍿 At first, I used to be in the camp that thought streamers should release whole seasons all at once.
Why not, you're not limited by airtime? The weekly cadence was just an artifact of a bygone era, kind of like if albums stayed limited to the playing length of vinyl or CDs forever…
I'm rethinking this. The one-time-dump is really limiting how long shows spend in the zeitgeist, and how much anticipation builds between episodes (which makes people enjoy them more!). Instant gratification can screw with the dopamine cycle.
It also makes it harder to get that slow-build, word-of-mouth organic growth.
Netflix should start releasing episodes more slowly, at least for their premium shows. I’m not sure what's optimal, though. It doesn't necessarily have to be 1 per week. Maybe 2 per week? ¯\_(ツ)_/¯
🏋️♀️ We grow most from overcoming challenges and striving for something.
The thing itself, though, may not be as important as it first seems.
Ideally, you want goals to be great and worthwhile. But not having an inspiring goal in sight shouldn’t mean there’s nothing to aim for and you should just aimlessly wander and waste away the hours.
In between these two extremes are pointless goals that you do “just because you decided to”. And you still benefit from accomplishing it!
The goal itself may not be much, but the work needed to reach the goal can be just as rewarding.
what I call a Big Pointless Goal: an aspiration that lacks grand purpose, yet requires substantial effort to attain.
(An editor of mine also calls these “stupid quests.”)
Many other examples are less extreme. The journalist Kim Cross once attempted 100 wheelies a day for 30 days on her bike. The professional runner Rickey Gates traveled every street in San Francisco. A friend of mine is making a five-foot-long rocking triceratops—think a prehistoric-themed rocking horse—in order to fulfill her childhood dream of riding a dinosaur. [...]
Traditional self-improvement goals, such as cutting out sugar or keeping a gratitude journal, are utilitarian and, most important in this context, not much fun.
Pointless goals, in contrast, are meant to be enjoyed. They trick us into doing the things we love, which can also put us in a flow state, where we’re deeply satisfied, present, and absorbed in the task at hand.
Maybe I need to set myself a few more pointless goals 🤔
💚 🥃 Amazing! In about 24 hours, there were already 50+ people in the Liberty Labs 🧪🔬🧬 🔭 Discord, and new ones keep coming!
🧩 A Word From Our Sponsor: Heyday 🧠
Do you have 100+ browser tabs open right now? 😬
Give your memory a boost with Heyday, the research tool that automatically saves the content you view, and resurfaces it alongside relevant Google search results. 👩💻
It’s like cheat codes for your memory. 😲💡
🧩 Give your memory a boost today 🧠
Business & Investing
🪄 Global foundries where sand is turned into magic
I’m guessing Intel isn’t considered a “foundry” for this because they don’t (yet) really do third-party stuff..?
You can clearly see the power law structure of the market: TSMC is more than twice as big as #2, which is more than twice as big as #3….
🪨⛏ Minecraft rejects NFTs
One of the biggest games in the world — if not the biggest — just stated its policy on NFTs:
Each of these uses of NFTs and other blockchain technologies creates digital ownership based on scarcity and exclusion, which does not align with Minecraft values of creative inclusion and playing together. NFTs are not inclusive of all our community and create a scenario of the haves and the have-nots. The speculative pricing and investment mentality around NFTs takes the focus away from playing the game and encourages profiteering, which we think is inconsistent with the long-term joy and success of our players.
We are also concerned that some third-party NFTs may not be reliable and may end up costing players who buy them. Some third-party NFT implementations are also entirely dependent on blockchain technology and may require an asset manager who might disappear without notice. There have also been instances where NFTs were sold at artificially or fraudulently inflated prices. We recognize that creation inside our game has intrinsic value, and we strive to provide a marketplace where those values can be recognized.
As such, to ensure that Minecraft players have a safe and inclusive experience, blockchain technologies are not permitted to be integrated inside our Minecraft client and server applications nor may they be utilized to create NFTs associated with any in-game content, including worlds, skins, persona items, or other mods.
The history of Walmart (Podcast) 🛒🛍
Friends-of-the-show Ben and David (💚 🥃) kicked off season 11 of Acquired with a great one:
It’s easy to take Walmart for granted today, but in their early days, they were quite the innovators *and* synthesizers of lots of scattered innovations that Sam Walton found in all kinds of places. WMT 0.00%↑
They were early to using computers for inventory and logistics, even having their own private satellite network to get real-time data from stores and broadcast meetings to the far corners of their empire. Even the very concept of having customers walk around the store and pick up the merchandise themselves or of discounting was once innovative (I’m not saying WM invented them, but Walton was early in that era).
(it was a pleasant surprise to see that the episode is sponsored by friends-of-the-show NZS! Hey guys 👋)
📺 Netflix vs Disney+ vs HBOmax
Disney+ growth is pretty insane (though I guess they priced it to move, especially at first). I wonder about some alternate timeline where either Netflix or Disney had bought WarnerMedia and gotten HBO 🤔 NFLX 0.00%↑ WBD 0.00%↑
Via Chartr
The Economics of Blade Runner 🤔 🐑
Interesting musings by James Pethokoukis:
Of course, Blade Runner is now regarded as a film classic, sci-fi or otherwise, for its stunning visuals and deep meditation on a future that includes sentient, artificial life. And many fans and film critics find its sequel (also a box office laggard), the Denis Villeneuve-directed Blade Runner 2049 from 2017, to equal the original, both visually and intellectually. Count me as one such fan — and repeat viewer — of the two films.
But not once have I ever watched either film and thought, “This is what the future might actually be like.” The world-building, while cinematically and dramatically compelling, doesn’t add up — even assuming all of its futuristic technologies are possible. Indeed, that assumption undermines the verisimilitude of the world-building.
This is a common dilemma in fiction. Utopias can be really boring 🥱
Iain M. Banks found a way around this in his ‘Culture’ series by having the utopian civilization dispatch ‘special circumstances’ agents to less advanced civilizations, and most of the conflict happens there.
A universe of technological acceleration
Here’s what the Blade Runner-verse asks me to believe: The post-1960s Great Stagnation of tech progress — at least as it transfers into measurable business productivity growth — ends. (Or maybe never happens in that reality.) Humanity finally achieves many of the technological leaps anticipated by 1960s futurists and technologists: artificial general intelligence, sentient AI, bioengineered android bodies far more capable than human ones, off-world colonies across the Solar System, and flying cars propelled at least partially by anti-gravity technology (which also, presumably, helps enable space colonization). [...]
Instead of one of the most famous cinematic depictions of a future dystopia, the world of Blade Runner should be one of dramatically greater wealth, health (there’s a historical correlation between tech progress and higher life expectancy), resources, and problem-solving capabilities.
Most of the sci-fi that we consume has to be dystopian in some way to be dramatic, but we shouldn’t make the mistake of believing that this means that any progress necessarily leads to dystopia.
As Jason Crawford often says, progress means solving certain problems and creating others, and then solving those problems, and so on.
There’s no laddering-up process for humanity that is smooth all the way, without challenges, but the alternatives of stasis or decline aren’t exactly appealing either (if you take off the romanticism-tinted glasses 👓).
⚓️ World’s biggest ports, growth from 2005 to 2021 📦📦📦
Ryan Petersen (founder of Flexport): “Trade growth since 2005 as shown by the scale of the largest ports.”
Note that the port of Los Angeles fell off the top 10. Rotterdam is the only non-Asian port to hang on, though at the very bottom…
Boeing 🛩 📈📉
Via Chartr (dual Chartr edition today! Okay) BA 0.00%↑
Science & Technology
CXL: Next-gen interconnect fabric to pool huge amounts of memory in data-centers 🖇
I don’t know that much about the technical details of this one, but friends-of-the-show Doug and Dylan wrote interesting pieces about it.
I want to highlight how this works at a very high level, and if you want to dig deeper, you can read the piece (click their names).
Datacenters are an incredibly expensive affair. Microsoft stated that up to 50% of their server costs is from DRAM alone [...]
A one size fits all model does not work, which is why you see cloud providers with dozens if not hundreds of different instance types. These are trying to optimize offerings of hardware to differing workloads. Even then, many users end up paying for things they really don’t need.
Our results show that 50% of all VMs never touch 50% of their rented memory.
Instance choices are not perfect, nor is the matching of those instances to hardware. In comes the issue of platform level memory stranding. Servers are configured for pessimistic instance type scenario.
At Azure, we find that a major contributor to DRAM inefficiency is platform-level memory stranding. Memory stranding occurs when a server’s cores are fully rented to virtual machines (VMs), but unrented memory remains. With the cores exhausted, the remaining memory is unrentable on its own, and is thus stranded. Surprisingly, we find that up to 25% of DRAM may become stranded at any given moment.
How do you solve this?
if we were to add larger pools of elastic memory (DRAM, to be precise) to large pools of CPUs (we have that today), we could solve the interconnect problem at the datacenter level. This would create a synchronous pool of memory that is always on the same page next to the CPU, not partitioned into hundreds of pieces of stranded and out-of-sync memory.
The implications of this are huge. I think the best way to think about this is the Apple M1 as a model of what heterogeneous compute improvement can offer at the micro-level. The datacenter is effectively one giant Mandelbrot fractal of the same problem on a larger scale. The same progress the Apple M1 chip has on the microscale would be available for an entire datacenter! The cost and performance improvements would be equivalent to a generation of performance scaling
CXL is the protocol by which chips can communicate with these external pools of memory in a performant, scalable way.
It’s basically a practical implementation of Jensen Huang’s vision of the data-center as the new unit of compute; if you start thinking of it like that, you can try to optimize your resources at the DC level rather than at the server/rack level.
⌚️🫀 Apple Watch saves woman’s life from otherwise asymptomatic tumor
More and more of those stories are coming out thanks to sensors that do things like fall detection, heart EKG, etc:
Kim swears by it now, but she didn't believe her Apple Watch for the first couple of nights in late May when it woke her up warning that her heart was in atrial fibrillation.
"The third night the numbers went a little too high for comfort," Kim said. "Then I said you know what, go to the emergency room if they tell you it's nothing to worry about then toss the watch."
But doctors in Maine soon confirmed that her heart was beating erratically for a simple and scary reason. She had a myxoma, a rare, fast-growing tumor that was choking off her heart's blood supply and would have eventually caused a stroke.
"They said 'Well how did you know you had A-fib?' and I said 'Because my watch told me,'" Kim said.
She had no other symptoms beside the A-fib but quickly made arrangements for a trip to Mass General, where she underwent surgery on June 27. Doctors removed a four-centimeter tumor that almost certainly would've killed her. (Source)
It'll be interesting to watch this evolve over time as more and more sensors can be embedded in devices. I'd love some kind of Dexcom-lite blood glucose monitoring, CO2 sensors as proxies for ambient air quality, etc… AAPL 0.00%↑
The Arts & History
🖼 🎨🤖 OpenAI changes licensing of images from DALL-E 2
Very cool:
Starting today, users get full usage rights to commercialize the images they create with DALL·E, including the right to reprint, sell, and merchandise. This includes images they generated during the research preview.
Users have told us that they are planning to use DALL·E images for commercial projectsDALL, like illustrations for children’s books, art for newsletters, concept art and characters for games, moodboards for design consulting, and storyboards for movies.
Less cool is that to be able to open it up to a wider-release beta, they’re limiting the number of images that each user can generate based on a credit system… Oh well, I guess compute doesn’t grow on trees.
Every DALL·E user will receive 50 free credits during their first month of use and 15 free credits every subsequent month. Each credit can be used for one original DALL·E prompt generation — returning four images — or an edit or variation prompt, which returns three images. [...]
In this first phase of the beta, users can buy additional DALL·E credits in 115-credit increments (460 images) for $15 on top of their free monthly credits
Two new episodes a week would be my ideal cadence. 4 to 6 weeks in total. Not gone in a day but doesn't take 3 months either. With so much content these days I find the all at once dump daunting.