345: Twitter A.D., Google, Zuck, BASF, Snowflake NRR, TSMC, Muscle Mass & Cognition, SpaceX & NASA
"It’s rare to catch lightning in a bottle"
Success traps can be harder to get out of, because at least with failure traps, there's incentives to change things. With success traps, the incentives are to keep doing the same thing. —Greg McKeown (paraphrased)
🤔 👨🔬 Everybody’s in the business of giving Mark Zuckerberg advice these days. I’m not sure what I’d say to him if he asked me (not that there’s a risk of that happening!).
Probably that it’s harder to innovate and move fast inside big companies with unlimited resources and few natural forcing functions, so maybe his VR/AR ambitions would have a higher chance of success if spun-off as a separate, 100% focused, leaner, more startup-like entity?
Maybe let the “R” out of the nest. When it can fly by itself and it gets to the “D” phase, you can always bring to bear the BigCo resources to take that research and develop it to scale.
It’s rare to catch lightning in a bottle. Even rarer to do it multiple times. You’re never promised a next one, and there’s a danger of letting past victories slip away if you don’t stay focused and in top fightin’ shape… 🥊 🏋️♂️
🌲🐿💭🚶🏻♂️🌲 While walking in the woods, it happened a few times that I walked up very close to squirrels or chipmunks that hadn’t seen me coming. They had huge jump-scares, freaked out and bolted for a tree or a hole in the ground.
It made me wonder, was I *that* stealthy, or can animals daydream the way humans do? (maybe just mammals with brain architectures similar to ours?)
Do their minds sometimes wander and they go inward and kind of lose track of what’s going on around them the way that we do?
Seems like a trait that would be evolutionarily pressured against unless there are also other benefits. Or maybe it’s a side-effect of some other beneficial adaptation (that’s what happens with many traits that are hard to explain directly).
I couldn’t find clear research on this (there are papers on dreaming, but few on daydreaming). ¯\_(ツ)_/¯
🎧 I enjoyed Ryan Holiday and Chris Williamson’s discussion about discipline.
There’s nothing new, but it contains good reminders — most of life’s important things are fairly simple and obvious but what’s hard is keeping them top-of-mind and applying them consistently over time, because it’s only over long periods that you get compounding and reap the fruits of expanding the surface area of opportunity around you.
💚 🥃 🦁 Thanks to C.H. (👋) for becoming a paid supporter yesterday!
⭐️ We’re at 17/100 of our goal! ⭐️
Liberty’s Highlights is reader-supported. To support my work, consider becoming a paid supporter. 🦁 감사합니다!
🧩 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 within your existing workflows. 👩💻
It’s like cheat codes for your memory. 😲💡
🧩 Give your memory a boost today 🧠
🏦 💰 Liberty Capital 💳 💴
🐦 Suggestions to make Twitter better (DMs & Algo Edition)
Nobody knows what will happen to Twitter with Musk at the wheel, but it’s definitely a new era.
Old twitter (the B.E. era?) often sucked because it never improved, but you could count on it being there because it never changed.
When you start changing things radically, they can get better, but there's always a risk it goes terribly wrong and you drop the salt shaker in the sauce... (the dispersion of possible outcomes becomes really wide in + and - directions).
Here are a couple of suggestions to improve the flavor:
Bring the direct message (DM) system at par with every other messaging platform out there (iMessage/Whatsapp/Signal/Messenger/etc). No need to reinvent the wheel, just copy all the table-stakes features that everybody else has (threaded replies, encryption, editing/deleting replies, using more emojis as reactions, etc)
But the big one that Musk may be interested in:
Allow some customization of your timeline algorithm. I don’t mean editing the code on Github, just some levers you could pull to make the experience closer to what you want. There’s no one-size-fits-all.
I still prefer to use the reverse-chronological timeline for reasons I explained in edition #204 (time flies!), but for the majority of people who are on the algo timeline, it may be beneficial to have a setting pane where you can say: “I want/don’t want XYZ”. Not just on/off switches, but rather sliders to increase or decrease the strength of certain signals.
f.ex. I want to see more/less tweets from people I don’t follow but that have been favorited by people I follow. I want a heavier/lower weight to be given to big popular “high-engagement” accounts. I want a bigger/lower weight on controversial tweets. I want a heavier weight on accounts I interact with often vs those I don’t. I want a heavier weight given to tweets from accounts I follow that tweet rarely (otherwise easier to miss), etc. Heavier/lower on retweets, on images and videos, on verified accounts, etc.
If I had that kind of control, I’d use the algo timeline at least some of the time, maybe even all the time 🤔
From Twitter’s point of view, if it makes the platform better for users, they’re more likely to stick around, use it more, etc, so there’s definitely a monetary benefit to these types of improvements. You could even make it part of Twitter Blue, power users would pay for *that* (unlike the current version). TWTR 0.00
👥👥👥👥👥👥👥 Google’s Headcount 👥👥👥👥👥👥👥👥
I know Google has a lot of employees and is hiring really fast, but this line by my friend MBI (💎🐕) really drove the point home:
[The] number of headcount added by Google in the last [3.5 years] = Meta's total employees as of 2Q'22
And it’s not like Meta is running the leanest, tightest ship out there… GOOG 0.00
✂️🇪🇺 ‘BASF to downsize ‘permanently’ in Europe’ 💶
The world’s biggest chemical company says “high energy costs make the region increasingly uncompetitive."
“The European chemical market has been growing only weakly for about a decade [and] the significant increase in natural gas and power prices over the course of this year is putting pressure on chemical value chains,” chief executive Martin Brudermüller said on Wednesday.
BASF, which produces products from basic petrochemicals to fertilisers and glues, spent €2.2bn more on natural gas at its European sites in the first nine months of 2022, compared with the same period last year. [...]
Germany remains BASF’s most important market for revenues, accounting for 18 per cent of its sales in the year to date, compared with 14 per cent from China. (Source)
The folly of relying on Russian gas while closing down nuclear power plants will cause waves of suffering that will echo for many years…
❄️ Doing the math on Snowflake’s Net Revenue Retention ☃️
Friend-of-the-show and Extra-Deluxe supporter Byrne Hobart (💚💚💚💚💚 🥃) wrote some good stuff about Snowflake:
if you take that net dollar retention and compound it out on a quarterly basis, a customer who did $100k in revenue as of the second quarter of 2018 would be producing $933k in revenue as of last quarter
Kind of incredible.
Of course, that NRR won’t last forever, but because Snowflake isn’t SaaS but rather consumption-based, it may last a while if the secular trends that they’re riding are as durable as they seem, and if they keep executing well and don’t punch themselves in the face somehow.
If they kept marketing and R&D spend at current levels, and never landed another customer, 158% dollar retention—the worst they’ve ever reported—would get them to breakeven by the end of 2024
Byrne has an interesting metaphor for what the company is trying to do:
One way to view the long-term upside case is that, in a very abstract sense, Snowflake is Enterprise Google. Not in the sense that it has the same kinds of products or sells them in the same way, but in the sense that it’s accomplishing the same kind of mission and has the same sorts of self-reinforcing loops. Google wants to organize the world’s information, but some information wants to stay hidden, because its value diminishes if everyone has it. […]
Snowflake wants to make it possible to analyze the world’s information, and to extend that possibility to whoever owns the information in question or can afford it. And while Google has created a truly valuable public good as well as a phenomenal business, it’s likely that the majority of the monetizable data in the world is the property of one big company or another. And as Snowflake makes it easier to realize that value, both by analyzing the data and selling it, they’ll become a big company, too.
🇹🇼 TSMC Fabs & Semiconductor Manufacturing Market Share
I was surprised to see that in 1990, Europe had more semiconductor manufacturing than the US. How things change. Of course, the absolute numbers of wafers at the time were tiny compared to now (wild guess, but just one of TSMC’s current fabs probably makes more chips than all of Europe at the time).
The map on the right shows the location of TSMC fabs in Taiwan. There are benefits to having so many fabs close to one another, rather than spread around the world:
Despite its Arizona investment, TSMC is trying to sit things out as the concentration of its fabs and suppliers in a tight cluster in Taiwan has enhanced its efficiency. “There are a lot of benefits to the way they are running things — especially the close connection between R&D and high-volume manufacturing where you can send an engineer to a fab just an hour away,” Hanbury says. “The cost savings and benefits of expertise are part of the TSMC model.”
🧪🔬 Liberty Labs 🧬 🔭
💪🧠 The link between Muscle Mass & Cognitive Function 🏋️♂️
Peter Attia writes about a review of multiple research papers on the topic:
Skeletal muscle is more than just a workhorse tissue required for movement. Contracting muscle also functions as an endocrine organ, releasing a wide variety of peptides and proteins – collectively known as “myokines” – which act as signaling molecules to other parts of the body.
Certain myokines, such as cathepsin B, can cross the blood brain barrier and exert direct effects on neurogenesis pathways. Some myokines are known to have anti-inflammatory effects both systemically and in the brain specifically. Muscle-generated BDNF – another myokine – promotes expression of genes involved in neuronal mitochondrial biogenesis, which in turn is critical for maintaining a healthy pool of mitochondria and avoiding excessive oxidative stress or dysfunctional energy metabolism, processes which subsequently impact muscle protein synthesis and breakdown.
The downstream effects of many myokines include triggering release of more myokines, creating a positive feedback loop and amplifying their effects. On the other hand, physical inactivity is thought to inhibit myokine release.
So it’s not just about having a healthy heart, avoiding falls and injuries, blood flow, etc. There’s a more direct effect on the brain:
the authors proposed a new hypothesis: muscle loss and physical inactivity causes a decrease in myokine secretion, which then triggers a “negative spiral” through an interorgan crosstalk and ultimately leads to cognitive decline.
According to the investigators, exercise therefore doesn’t merely provide you with an extra health boost, without which you would maintain your baseline level of health. Physical inactivity actually impairs the processes described above, resulting in a decline in overall health as well as cognitive function.
If that’s not motivation to go run around and lift heavy things, I don’t know what is!
🚀 SpaceX passes Boeing to become NASA’s #2 supplier 🛰
NASA obligated $2.04 billion to SpaceX in fiscal year 2022, which ended last month, according to new federal procurement data.
For the first time, the amount paid by the space agency to SpaceX exceeds that paid to Boeing, which has long been the leading hardware provider to NASA. Boeing received $1.72 billion during the most recent fiscal year
Who’s number 1?
The California Institute of Technology, which manages the Jet Propulsion Laboratory field center for NASA, remains the agency's No. 1 contractor, with $2.68 billion in funding. The academic institution is responsible for operating the California-based NASA field center and distributing funding for myriad robotic spacecraft missions such as Mars Perseverance and the Europa Clipper.
Starlink for vehicles 🚙📡🛰🛰🛰🛰🛰🛰🛰
Speaking of SpaceX, Starlink announced that they now have a SKU for RVs, with an antenna that can be used on moving vehicles.
The new Flat High Performance Starlink allows users to enjoy high-speed, low-latency internet while in-motion.
With a wide field of view and enhanced GPS capabilities, the Flat High Performance Starlink can connect to more satellites, allowing for consistent connectivity on the go.
The hardware is designed for a permanent installation on your vehicle and is resilient in harsh environments.
It makes me wonder if this is a stealth “tank mode”. Maybe not literally for tanks, but for other military vehicles in Ukraine… 🤔
🎨 🎭 Liberty Studio 👩🎨 🎥
Fun with AI-generated Avatars 🐦🎨🤖
I know I’m probably the only one who finds this dad humor funny, but for the past couple of weeks I’ve been changing my radio-telescope Twitter avatar a couple of times a day and replacing it with variants of the original that I made with Stable Diffusion.
It made me smile to imagine that at least a couple people would do a confused double-take and go “uh?! Was this always like that? Am I going crazy or did it change?”.
So far, the only person who messaged me about it is friend-of-the-show Conor Mac (🏴🦅) — he’s got that eagle eye!
Thank you for following Snowflake, I, too, find the company interesting.
What's you take on Slootman’s pay package? It includes more than 13.7 million options with a strike price of $8.88. At current prices, it's worth around $2 billion. Isn't this kind of a lot? I know he is a very capable executive and has a great reputation, but still...
I searched for the original picture of the radio telescopes and tried to figure if the new ones are just alternates. Also, searched in all the games I played as a kid (Red Alert, Starcraft etc..), trying to identify them. Damn AI 🤖 Haven't though of that.