175: Ho Nam, WCM, Cloudflare 11th Generation, Health Thoughts, Selling a Process for 2 & 20, Intel & Europe, Company Benchmarking, Norway's EVs, and Richard Turner
"you haven’t somehow escaped the game."
If you only read the books that everyone else is reading, you can only think what everyone else is thinking.
— Haruki Murakami
🤪 Brian Eno on his use of randomness:
“For me it’s a way of searching for a musical space that I wouldn’t [find] using just my taste.
I mean, one’s taste tends to propel you into the same areas again and again. The interesting thing about randomness is that sometimes, you’re taken to an area where you didn’t expect to go. And sometimes, it turns out to be a really interesting place.
Randomness is a tool, [it’s not about the randomness itself].”
—Brian Eno to Rick Rubin on Broken Record podcast
Back in the intro of edition #146 I wrote a little about this:
I’m not big into weirdness for weirdness’ sake [...] but I think you have to inject a certain amount of craziness, weirdness, and exploration of strange stuff into your life…
Why is it so important?
Because it’s how you find a lot of the good stuff!
Look at it from the other side (see how I avoided “invert it”? that’s a peek at how writing is behind-the-scenes right there): What’s the opposite of this strange/weird/crazy stuff?
Normal stuff.
Don’t get me wrong, there’s plenty of good stuff that is perfectly normal, but if you stick to that area, don’t expect your results to be very different from other normal people (because if your head is filled by mostly the same things, your thoughts are more likely to be similar too). [...]
You can pretty easily find the “normal good” stuff (it’s very visible — they have a big marketing budget and plenty of fans that’ll tell you about it), and combine it with the “strange good” stuff so that “one” + “the other” = more good stuff for you.
You have to find a way to inject some randomness/exploration/strangeness into your life somehow (maybe this newsletter serves some of that function for you — happy to be of service! *George Patton Salute*)
🥼🩺 I’ve been thinking about health stuff lately, partly because I posted about a 4-day fast that I recently did, and lots of people had questions about that.
I also had a chat with Louisa Nicola, someone directly involved in the field, and talked with a few others about the impact of sleep and alcohol and such...
I don’t know whether it’s a random fluctuation or if this topic is having ‘a moment’ in my circles, but I thought I’d share some of my thoughts on how I think about all that health stuff at the meta level.
Some of the main elements in my framework are: First figure out if it’s safe, not making a decision is still making a decision, and the perfect can be the enemy of the good.
As soon as you get into health, it’s hard not to notice just how much noise there is, and how much conflicting information is out there, a lot of it from people who are trying to sell you something.
There’s also a strong reaction from many people along the lines of: “You’re not a doctor, you’re not a psychiatrist, you’re not a psychologist, you’re not a dieticians, you’re not a nutritionist, etc… You can’t have an opinion on this, whatever you believe has to basically be worthless because you didn’t study this in school, etc”.
Well, skepticism isn’t bad per se. I mean, you can say the same thing about finance/investing, right? I also have never studied or worked in finance, so does it preclude me from having an opinion? Is it not possible for me to learn a lot of it by myself, without a formal education, Good Will Hunting style?
But it’s also true for finance that most people who think they know something about it know a lot less than they think, and should probably stick to index funds…
So the health stuff skepticism has a good hit rate, but it doesn’t mean that everybody is so clueless that they’re basically more likely than not to hurt themselves, and that there isn’t a big difference between a non-expert just making stuff up and a non-expert who has been listening to experts for years relaying what he’s learned from them.
I have a strong belief that whatever you do, that’s a decision. If you decide *not* to decide something, not to learn about it, you haven’t somehow escaped the game.
You’ve made a very real decision that will impact the trajectory of your life, possibly in a negative direction vs if you had decided that you can actually learn some things about an intimidating or complex field.
Same if you decide to only do things that are 100% certain and perfectly clear: That can keep you from doing lots of very good things just because you require certainty and want to avoid a mistake at all costs, leading to an outcome that is worse on the net than if you had been less rigid about it.
I know that nobody is going to take my life in charge and tell me what I should do. I don’t have a personal Peter Attia on payroll to help me with all that stuff day-to-day, it’s up to me to figure out as much as I can, and I think that as long as I use some caution, I should be able to do much better than ignorance, even if I don’t get perfection (who does?).
First step is to spend a lot of time building your BS detector and finding which sources can be more highly trusted. Make sure you understand something enough to know how and/or why it does what it does, and aren’t just doing something because someone said you should (same as for finance).
Let’s look at why the risk/reward ratio of fasting is very asymmetric as an example: humans have been fasting forever, millions do it around the world today, and *everybody* did it a short-time ago in evolutionary time, in the era before grocery stores and fridges, and before agriculture, which isn’t that old in the grand scheme of things. So evidence of safety is overwhelming. On the rewards side, benefits for weight control, blood-glucose control, and diabetes treatment/prevention are well documented, and we’re starting to get more and more information about the benefits of ketosis and autophagy, which your body will ramp into after a few days without food. It’s hard to imagine that evolution would be so wasteful as to only build important metabolic processes to be active in a fed state, and not also have plenty of important processes activated in the fasted state that our ancestors spent so much time in.
💪Back in edition #144, I gave myself the challenge of getting back to being able to do 10 full pull-ups (from a starting point of 4 at the time, with some difficulty on the last one).
I hope some of you have been following along, with whatever objective you felt was achievable-but-a-stretch, and have made good progress. Don’t stop!
It’s been a while since I gave an update on this, and frankly, I’ve been slacking a bit.
I did fairly well at first, going up to 6, but I’ve been plateauing since and am still in the 6-7 range.
I need to step up my game, get back on the horse, never give up, never surrender, <insert more motivational clichés here>, etc.
Now having written the above, I have no choice but to go do some pull-ups. If you have a pull up bar nearby, take a break and join me, it’s good for you! 🫀
💚 🥃 Let’s make a deal. Today I’ll spare you the pitch about becoming a paid supporter, and in return, you’ll considering becoming a paid supporter (it’s fast and painless):
Investing & Business
Benchmarking is a Tool, Gotta Use it Right
Friend-of-the-show Shomik Ghosh has a nice forceful tirade against, um, not benchmarking per se, but against the unthinking use of benchmarking by some, without taking into account that not-all-else-is-equal (ceteris paribus is a nice legal fiction, but this is the real world):
Every stage and state of a business is different. Yes it is important to understand how other comparable companies are performing and what best in class metrics look like. However, one business can have the same scale of revenue and even be in the same category like CRM software but could have drastically different TAM and market share trajectories at the same level of metrics. So for example Salesforce vs Veeva are completely different. If they both had $300M ARR, Salesforce at that time would be targeting a much broader market with more competitors while Veeva’s TAM would be smaller but likely they have more market share captured in the life sciences vertical.
Amen. It’s too easy to draw superficial comparisons between things and forget that there are very real, and very important differences that change everything.
It doesn’t just take place when looking at financials, it happens in every field, including at the product level:
Oh, some esoteric Android handset that sells 1 million units a year is using this fancy new screen technology, why isn’t Apple doing the same? Are they dumb, don’t they know they’re falling behind?
Oh, what do you mean it’s impossible to produce that type of screen at the volumes and color calibration that Apple requires at this point and the capacity to do so won’t exist on Earth for a while? Oops.
The most common example of Benchmarking Fallacy I run across is in growth stage board meetings. Companies with $20M ARR+, everyone starts to focus on gross margins and net revenue retention. The common refrain well be “we ran benchmarks across your stage and your gross margins are coming out 10% lower. We need to work to improve that asap as your gross margins should be higher. How can we work on improving them.” [...]
What if the best thing for the business is actually to take gross margins even lower? Let’s say adding more customer support and professional services now on a lower revenue base will onboard customers better and retain them for longer in the future. This will generate more Free Cash Flow in the long term even if the short term margins look worse. But for the health of the business, this may actually be the right tradeoff, assuming no near term funding needs. [...]
What management teams and investors need to focus on more is the stage of the business, state of the market, customer base, and what is right for the business at that point in time that will eventually lead to enabling the business to have best in class metrics according to the benchmarks in the future.
Exactly. Keep your eye on the ball, and make sure it’s the right ball.
Selling “Process” for 2 & 20
Interesting thread by Charioteer Investing (below are just a few tweets taken from the middle of the thread, click the link for the whole thing):
Despite this highly detailed playbook, John's performance didn't actually seem that great. He would be at a shop for about 2 years, blow up, and then end up at another shop. He repeated this several times and is still repeating this pattern to this day.
I always wondered how he could keep getting jobs as a PM when he kept bombing. Turns out, investment committees/hiring teams looooooved "The Playbook". His actual results mattered less than the fact that he had a super detailed plan for how to invest.
I'm often reminded of this phenomenon when i hear managers talk about their "process". The biggest illusion in the investment industry is the "meritocracy of returns" aka the belief that performance is what matters and if you simply put up awesome numbers you're bound to win
I'm not the first to say this but getting a job in investing is way more about signalling than skills. Investing is typically a highly undifferentiated product, most managers simply end up owning the same 30-50 stocks as everyone else. So how do you charge 2/20 for that?
The answer is that you convince the client that you are worthy of their capital because you have a "process" that is the "secret sauce" and is better than everyone else's.
I’m an outsider to managing other people’s money, having never studied or worked in finance, so I can’t confirm this from experience.
But it certainly rings true based on a lot of what I’ve heard from people in that world…
h/t Rishi Gosalia (💎 🐕)
‘Intel to Invest Up to $95 Billion in European Chip-Making’ Over a Decade
Intel Chief Executive Officer Pat Gelsinger on Tuesday said the company was planning two chip factories at a new site in Europe and could potentially expand it further, with the increases raising the total investment over about a decade to the equivalent of as much as €80 billion. [...]
Intel said it plans to commit manufacturing capacity at a factory in Ireland to the auto-chip sector. And it is establishing a chip-design team to help others adapt designs so they can use Intel’s manufacturing capabilities. [...]
Mr. Gelsinger on Tuesday predicted that the market for car chips would more than double by the end of the decade. Semiconductors, he said, would account for more than 20% of the material costs for new premium-segment cars, up from 4% in 2019, as new driver-assistance capabilities, flashy touch screens and other features that require more processing power become more widespread. (Source)
I’m sure that the “up to” is highly dependent on how much they get in incentives and subsidies, but that seems to be the game to play now if anyone is to even try to keep up with TSMC. INTC 0.00%↑ TSM 0.00%↑
Interview: Me in Neuro Atheletics
Neuroscientist Louisa Nicola asked me a couple questions about the importance of having a clear mind and sleep, and my answers were published on Tuesday in the Neuro Athletics newsletter.
Interview: Mike Tian of WCM Investment Management
I’ve enjoyed all the interviews I’ve heard with people from WCM, which has to be a pretty good signal for the quality of that firm.
The latest one by Ted Seides is with Mike Tian.
This is how the show-notes describe the episode:
Our conversation begins with Mike’s definition of moats and assessment of moat trajectories, using Visa as an example. We then turn to his application of moat trajectory analysis in China, including the value of heavy industries, analysis of culture, consumer preferences, risk, and some investment examples along the way. We close with Mike’s thoughts on the moat trajectory of WCM itself.
What I enjoy most is Mike’s thoughtful way of thinking about things.
h/t Rishi Gosalia, Redux (💎 🐕)
Interview: Ho Nam of Altos
You’ve probably seen it, because, well, I sent you an email of it yesterday! But just in case Gmail stupidly sent it to some hidden-away folder, check out my interview with Ho Nam from Altos Ventures.
(tip: you can reply to any of my newsletter emails with “hi” and it’ll tell your email filters that we’re connected, and don’t worry, I’ll know what it means, it won’t bother me at all — in fact, I get a smile every time I see one of those “hi” emails come in 😗👍)
Ever noticed how errors are always in favor of the bank?
Science & Technology
Cloudflare’s 11th generation Edge Server
Interesting blog post by a Cloudflare engineer describing their latest edge server hardware:
We aim to introduce a new server platform to our edge network every 12 to 18 months or so, to ensure that we keep up with the latest industry technologies and developments [...] We consider this length of cycle optimal: short enough to stay nimble and take advantage of the latest technologies, but long enough to offset the time taken by our hardware engineers to test and validate the entire platform. When we are shipping servers to over 200 cities around the world with a variety of regulatory standards, it’s essential to get things right the first time.
The metric they optimize most for is “requests per watt”, a proxy for efficiency, both on the power and cost side.
Because of this focus, they’re now developing two branches of servers in parallel, the main x86 brain, and the new ARM-based branch, which is build around Ampere Altra chips.
On the Intel vs AMD front:
We evaluated Intel’s latest generation of “Ice Lake” Xeon processors. Although Intel’s chips were able to compete with AMD in terms of raw performance, the power consumption was several hundred watts higher per server - that’s enormous. This meant that Intel’s Performance per Watt was unattractive.
We previously described how we had deployed AMD EPYC 7642’s processors in our generation 10 server. This has 48 cores and is based on AMD’s 2nd generation EPYC architecture, code named Rome. [...]
based on the samples we received from our vendors and our subsequent testing, hardware from AMD and Ampere made the shortlist for our generation 11 server. On this occasion, we decided that Intel did not meet our requirements.
The 11th gen x86 branch is testing 48, 56 and 64 core samples chips from AMD’s 3rd generation EPYC architecture, code named Milan:
Gen 10 used 256 gigabytes of DDR4-2933 RAM, and gen 11 is going to use 384GB of DDR4-3200.
The technical term for this is “a crapload of RAM”.
They also looked at 512GB, but that didn’t give enough of a performance boost to justify the cost.
On the storage front, they went from 3x Samsung PM983 x 960GB on the 10th gen to 2x Samsung PM9A3 x 1.92TB on the 11th.
The real question, though, is: Does it run Crysis?
Norway’s Passenger Car Transition Away from Oil
From almost nothing 10 years ago to almost everything today. Kudos.
I’d love to see this happen elsewhere, in much bigger markets.
I realize that Norway has very special incentives, but it’s a bit like solar power: At first it seems like a niche thing that’ll only work in the sunniest places that have the highest electricity costs, but then soon enough it makes sense in vast swathes of the planet.
h/t Paul Graham for posting the graph and making fun of me for not noticing the source.
The Arts & History
👨🎨 Damian Lechoszest 🎨
Stumbled upon an artist that I like recently, oil painter Damian Lechoszest (that’s his official website, but he’s also on Instagram and Tiktok).
Look at this timelapse video of a some of his paintings:
It’s a great reminder of just how many layers of paints are in a painting, and how much of it is iteration.
You don’t start from the blank canvas and “print” the final result on it, you first do a crude thing, and you improve that crude thing, and you then improve that and so on until you get the final result.
I also love this one, the way he paints light and shadows, very cool:
‘Dealt’ Documentary on Richard Turner
This is a follow-up to edition #174 where I posted the trailer of this short film from 2017.
I had a chance to watch it a couple days ago with my wife, and we both really enjoyed it. He’s quite a character, both inspirational and tragic at the same time, a nuanced and extreme person.
I think there’s a lot to learn from his life, and it both makes me want to make sure I don’t make the same mistakes he made, and step up my game the way he has.
There’s a scene when he’s in a school teaching some kids, and he meets a young girl and talks to her a bit — that’s probably my favorite part of the film, it’s very touching, and makes me wonder how that girl is doing now. Everything about her just sparkled with intelligence.
Anyway, good film. Having seen it now, I can recommend it.
You can choose a ready guide, In some celestial voice, If you choose not to decide, You still have made a choice 🇨🇦🤘
A reader sent me this. From my inbox to your comments: An excerpt from David Stark’s “The Sense of Dissonance”:
Each evening during their hunting season, the Naskapi Indians of the Labrador peninsula determined where they would look for game on the next day’s hunt by holding a caribou shoulder bone over the fire. Examining the smoke deposits on the caribou bone, a shaman would read out, for the hunting party, the points of orientation of the next day’s search. In this way, the Naskapi introduced a randomizing element to confound a short- term rationality that would have concluded that the one best way to find game would be to look again tomorrow where they had found game today. By following the divergent daily maps of smoke on the caribou bone, they avoided locking in to early successes that, while taking them to game in the short run, would have depleted in the long run the caribou stock in that quadrant and reduced the likelihood of successful hunting. By breaking the link between future courses and past successes, the tradition of shoulder- bone reading was an antidote to path dependence in the hunt.
I am not arguing that we should organize search with a roll of the dice. Nonetheless, the lesson from Labrador does nicely express one group’s attempt to deal with the counterpart, in that region’s ecology, to the nonergonomic QWERTY keyboard. Indeed, studies in evolutionary economics and organizational analysis do suggest that organizations that learn too quickly sacrifice efficiency. Allen and McGlade, for example, use the behavior of Nova Scotia fishermen to illustrate the possible trade-offs of exploiting old certainties and exploring new possibilities. Their model of these fishing fleets divides the fishermen into two classes: the rationalist “Cartesians,” who drop their nets only where the fish are known to be biting, and the risk-taking “Stochasts,” who seek new schools of fish. In simulations where all the skippers are Stochasts, the fleet is relatively un- productive, because knowledge of where the fish are biting is unutilized; but a purely Cartesian fleet locks onto the “most likely” spot and quickly fishes it out. More efficient are the models that, like the actual behavior of the Nova Scotia fishing fleets, mix Cartesian exploiters and Stochastic explorers.